id stringlengths 9 13 | content unknown | decision stringclasses 13
values | reviews listlengths 3 12 | metareview listlengths 1 3 | sentence_texts listlengths 15 712 | opinion_groups listlengths 1 114 | conflicts_validation listlengths 1 155 | rebuttal_validation listlengths 1 155 | opinions listlengths 1 155 | PDF_path stringlengths 36 43 | PDF_version stringclasses 4
values | MD_path stringlengths 34 41 | conference stringclasses 13
values |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
rygp3iRcF7 | {
"title": "Area Attention",
"abstract": "Existing attention mechanisms, are mostly item-based in that a model is trained to attend to individual items in a collection (the memory) where each item has a predefined, fixed granularity, e.g., a character or a word. Intuitively, an area in the memory consisting of mult... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
... | [
[
{
"role": "AC",
"data": {
"metareview": "although the idea is a straightforward extension of the usual (flat) attention mechanism (which is positive), it does show some improvement in a series of experiments done in this submission. the reviewers however found the experimental results to b... | [
"I prefer the idea of using some statistics (such as variances) of multiple items for attention.",
"This direction may lead to better attention units for future works.",
"I do not fully understand the argument, \"Attention mechanisms are designed to focus on a single item in the entire memory\".",
"In my unde... | [
[
5
],
[
2
],
[
16
],
[
35
],
[
0
],
[
25
],
[
26
],
[
34
],
[
4
],
[
18,
21
],
[
31
],
[
3,
9,
13,
15,
28
],
[
17
],
[
19
],
[
20,
32
],
[
22
],
[
23
],
[
... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"cor... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"cor... | [
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
0
]
}
],
"category": [
"ORIG-MTH"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1
]
}
],
"category": [
"SIGN-BRD"
]... | benchmark/PDF/ICLR2019_rygp3iRcF7.pdf | openreview | benchmark/MD/ICLR2019_rygp3iRcF7.md | ICLR 2019 |
Bye5OiR5F7 | {
"title": "Wasserstein proximal of GANs",
"abstract": "We introduce a new method for training GANs by applying the Wasserstein-2 metric proximal on the generators. \nThe approach is based on the gradient operator induced by optimal transport, which connects the geometry of sample space and parameter space in impli... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4
]
},
"scores": {
"Solid": null,
"Presentation": null,
"No... | [
[
{
"role": "AC",
"data": {
"metareview": "Both R3 and R1 argue for rejection, while R2 argues for a weak accept. Given that we have to reject borderline paper, the AC concludes with \"revise and resubmit\"."
}
}
]
] | [
"The paper intends to utilize natural gradient induced by Wasserstein-2 distance to train the generator in GAN. Starting from the dynamical formulation of optimal transport, the authors propose the Wasserstein proximal operator as a regularization, which is simple in form and fast to compute. The proximal operator ... | [
[
21
],
[
22
],
[
22
],
[
4
],
[
10
],
[
16
],
[
17
],
[
23
],
[
11
],
[
19
],
[
0,
8,
20
],
[
3
],
[
14
],
[
15
],
[
1
],
[
2
],
[
13
],
[
7
],
[
5
],
[
... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct"
] | [
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"incorrec... | [
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
0
]
}
],
"category": [
"ORIG-MTH"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1
]
},
{
"role": "Author",
"dat... | benchmark/PDF/ICLR2019_Bye5OiR5F7.pdf | openreview | benchmark/MD/ICLR2019_Bye5OiR5F7.md | ICLR 2019 |
HJx38iC5KX | {
"title": "Domain Generalization via Invariant Representation under Domain-Class Dependency",
"abstract": "Learning domain-invariant representation is a dominant approach for domain generalization, where we need to build a classifier that is robust toward domain shifts induced by change of users, acoustic or light... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4
]
},
"scores": {
"Solid": null,
"Presentation": null,
"No... | [
[
{
"role": "AC",
"data": {
"metareview": "This paper proposes a new solution to the problem of domain generalization where the label distribution may differ across domains. The authors argue that prior work which ignores this observation suffers from an accuracy-vs-invariance trade-off whil... | [
"The paper proposed a problem that most prior methods overlooked the underlying dependency of classes on domains, namely p (y|d) \\= p(y). Figure 1 is used to illustrate this issue.",
"If the conditional probability of source domain and target domain is not equal (i.e., p(y|x_S) \\= p(y|x_T) ), the optimal inv... | [
[
8
],
[
0
],
[
1
],
[
2,
10
],
[
3
],
[
5
],
[
6
],
[
9
],
[
4
],
[
7
]
] | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct"
] | [
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct"
] | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0
]
}
],
"category": [
"ORIG-PROB"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1,
3,
4
]
},
{
"role... | benchmark/PDF/ICLR2019_HJx38iC5KX.pdf | openreview | benchmark/MD/ICLR2019_HJx38iC5KX.md | ICLR 2019 |
H1ecDoR5Y7 | {
"title": "Local Stability and Performance of Simple Gradient Penalty $\\mu$-Wasserstein GAN",
"abstract": "Wasserstein GAN(WGAN) is a model that minimizes the Wasserstein distance between a data distribution and sample distribution. Recent studies have proposed stabilizing the training process for the WGAN and im... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8
]
},
"scores": {
... | [
[
{
"role": "AC",
"data": {
"metareview": "All three reviewers expressed concerns about the assumptions made for the local stability analysis. The AC thus recommends \"revise and resubmit\"."
}
}
]
] | [
"In the paper, WGAN with a squared zero centered gradient penalty term w.r.t. to a general measure is studied. Under strong assumptions, local stability of a time-continuous gradient ascent/descent dynamical system near an equilibrium point are proven for the new GP term. Experiments show comparable results to the ... | [
[
5
],
[
6
],
[
8
],
[
3
],
[
13
],
[
14,
15
],
[
18
],
[
1,
7
],
[
9
],
[
10
],
[
11
],
[
17
],
[
2
],
[
0
],
[
4
],
[
12
],
[
16
]
] | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct"
] | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct"
] | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0,
1
]
}
],
"category": [
"SIGN-SOT"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
2
]
},
{
"role": "Author",
... | benchmark/PDF/ICLR2019_H1ecDoR5Y7.pdf | openreview | benchmark/MD/ICLR2019_H1ecDoR5Y7.md | ICLR 2019 |
B1xOYoA5tQ | {
"title": "Multi-way Encoding for Robustness to Adversarial Attacks",
"abstract": "Deep models are state-of-the-art for many computer vision tasks including image classification and object detection. However, it has been shown that deep models are vulnerable to adversarial examples. We highlight how one-hot encodi... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9
]
},
"scores... | [
[
{
"role": "AC",
"data": {
"metareview": "This paper proposes a method for improving robustness to black-box adversarial attacks by replacing the cross-entropy layer with an output vector encoding scheme. The paper is well-written, and the approach appears to be novel. However, Reviewer 4 r... | [
"This work proposes an alternative loss function to train models robust to adversarial attacks.",
"Specifically, instead of the common sparse, N-way softmax-crossentropy loss, they propose to minimize the MSE to the target column of a random, dense orthogonal matrix. I believe the high-level idea behind this work... | [
[
1
],
[
21
],
[
16
],
[
0,
2
],
[
19
],
[
7
],
[
26
],
[
33
],
[
35
],
[
4
],
[
6
],
[
12
],
[
14
],
[
17
],
[
18
],
[
22,
30
],
[
24
],
[
25
],
[
29
],
[
... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"cor... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"c... | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0,
1
]
}
],
"category": [
"ORIG-MTH"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
2
]
}
],
"category": [
"CLAR... | benchmark/PDF/ICLR2019_B1xOYoA5tQ.pdf | openreview | benchmark/MD/ICLR2019_B1xOYoA5tQ.md | ICLR 2019 |
HJe3TsR5K7 | {
"title": "Learning Joint Wasserstein Auto-Encoders for Joint Distribution Matching",
"abstract": "We study the joint distribution matching problem which aims at learning bidirectional mappings to match the joint distribution of two domains. This problem occurs in unsupervised image-to-image translation and video-... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8
]
},
"scores": {
... | [
[
{
"role": "AC",
"data": {
"metareview": "This paper proposes a new image to image translation technique, presenting a theoretical extension of Wasserstein GANs to the bidirectional mapping case. \n\nAlthough the work presents promise, the extent of miscommunication and errors of the origin... | [
"The whole model can be simplified by this: using auto-encoders for X and Y's reconstruction, then use Triple GAN loss to match the joint distribution of (X, Y).",
"However, the deterministic model with GAN loss looks problematic to me.\nquestions:",
"1.\nAlthough the authors showed strong evidence in their exp... | [
[
18
],
[
19
],
[
20
],
[
21
],
[
22
],
[
13,
16
],
[
17
],
[
8
],
[
0
],
[
2
],
[
3
],
[
4
],
[
10
],
[
5
],
[
6
],
[
9
],
[
15
],
[
1
],
[
12
],
[
14
],... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct"
] | [
"correct",
"correct",
"incorrect",
"incorrect",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct"... | [
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
0
]
},
{
"role": "Author",
"data": [
10,
11,
12,
13,
14,
15,
25,
26,
27
]
}
],
"c... | benchmark/PDF/ICLR2019_HJe3TsR5K7.pdf | openreview | benchmark/MD/ICLR2019_HJe3TsR5K7.md | ICLR 2019 |
r1ledo0ctX | {
"title": "Consistency-based anomaly detection with adaptive multiple-hypotheses predictions",
"abstract": "In one-class-learning tasks, only the normal case can be modeled with data, whereas the variation of all possible anomalies is too large to be described sufficiently by samples. Thus, due to the lack of repr... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6
]
},
"scores": {
"Solid": null,
"Pres... | [
[
{
"role": "AC",
"data": {
"metareview": "This paper proposes an anomaly-detection approach by augmenting VAE encoder with a network multiple hypothesis network and then using a discriminator in the decoder to select one of the hypothesis. The idea is interesting although the reviewers foun... | [
"The paper proposes a new method for anomaly detection using deep learning. It works as follows.",
"The method is based on the recent Multiple-Hypotheses predictions (MHP) model, the impact of which is yet unclear/questionable. The idea in MHP is to represent the data using multiple models. Depending on the part ... | [
[
27
],
[
1
],
[
6
],
[
13
],
[
25
],
[
10
],
[
11
],
[
12
],
[
17,
24
],
[
18
],
[
21
],
[
22
],
[
23
],
[
26
],
[
19,
20
],
[
9,
15
],
[
14,
16
],
[
28
],
... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"i... | [
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"incorrect",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correc... | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1
]
},
{
"role": "Author",
"data": [
... | benchmark/PDF/ICLR2019_r1ledo0ctX.pdf | openreview | benchmark/MD/ICLR2019_r1ledo0ctX.md | ICLR 2019 |
ryeh4jA9F7 | {
"title": "Playing the Game of Universal Adversarial Perturbations",
"abstract": "We study the problem of learning classifiers robust to universal adversarial perturbations. While prior work approaches this problem via robust optimization, adversarial training, or input transformation, we instead phrase it as a tw... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8
]
},
"scores": {
... | [
[
{
"role": "AC",
"data": {
"metareview": "Reviewers mostly recommended to reject after engaging with the authors, with one reviewer slightly suggesting to accept, but with confidence 1. Please take reviewers' comments into consideration to improve your submission should you decide to resubm... | [
"The authors focus solely on universal adversarial perturbations, considering both epsilon ball attacks and universal adversarial patches. They propose a modified form of adversarial training inspired by game theory, whereby the training protocol includes adversarial examples from previous updates alongside up to d... | [
[
30
],
[
3
],
[
4,
9
],
[
14
],
[
31
],
[
17
],
[
1
],
[
5
],
[
2
],
[
6,
11
],
[
7
],
[
16,
29
],
[
26
],
[
13
],
[
15
],
[
10,
20
],
[
23
],
[
22
],
[
... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"c... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"c... | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1
]
},
{
"role": "Author",
"data": [
... | benchmark/PDF/ICLR2019_ryeh4jA9F7.pdf | openreview | benchmark/MD/ICLR2019_ryeh4jA9F7.md | ICLR 2019 |
HJeRm3Aqt7 | {
"title": "GenEval: A Benchmark Suite for Evaluating Generative Models",
"abstract": "Generative models are important for several practical applications, from low level image processing tasks, to model-based planning in robotics. More generally,\nthe study of generative models is motivated by the long-standing end... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
... | [
[
{
"role": "AC",
"data": {
"metareview": "The paper introduces a benchmark suite providing a series of synthetic distributions and metrics for the evaluation of generative models. While providing such a tool-kit is interesting and helpful and it extends existing approaches for evaluating ge... | [
"Overall, this is a thorough attempt at a system for evaluating various generative models on synthetic problems vaguely representative of the kinds of problems claimed to be covered by GANs. I think the approach and the conclusions drawn are mostly reasonable, with one major caveat discussed shortly.",
"I also th... | [
[
3
],
[
27
],
[
5
],
[
13
],
[
16
],
[
23
],
[
6
],
[
1
],
[
26
],
[
2,
12
],
[
4
],
[
8,
9
],
[
10
],
[
11
],
[
14
],
[
18
],
[
19
],
[
20
],
[
24
],
[
... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"cor... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"incorrect",
"correct",
"correct",
"correct",
"correct",
... | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1
]
},
{
"role": "Author",
"data": [
... | benchmark/PDF/ICLR2019_HJeRm3Aqt7.pdf | openreview | benchmark/MD/ICLR2019_HJeRm3Aqt7.md | ICLR 2019 |
HJeQToAqKQ | {
"title": "TherML: The Thermodynamics of Machine Learning",
"abstract": "In this work we offer an information-theoretic framework for representation learning that connects with a wide class of existing objectives in machine learning. We develop a formal correspondence between this work and thermodynamics and discu... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8
]
},
"scores": {
... | [
[
{
"role": "AC",
"data": {
"metareview": "Connecting different fields and bringing new insights to machine learning are always appreciated. But since it is challenging to do it needs to be done well. This paper falls short here. "
}
}
]
] | [
"This paper builds on the (Alemi et al 2018) ICML paper and presents a formal framework for representation learning. The authors use a graphical model for their representation learning task and use basic information theoretic inequalities to upper-bound their measure of performance which is a KL divergence. The aut... | [
[
2
],
[
4
],
[
18
],
[
1,
7,
9
],
[
3
],
[
5
],
[
6
],
[
10
],
[
11,
17
],
[
19
],
[
20
],
[
13
],
[
16,
21
],
[
22
],
[
15
],
[
12
],
[
0
],
[
8
],
[
... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct"
] | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct"
] | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1
]
},
{
"role": "Author",
"data": [
... | benchmark/PDF/ICLR2019_HJeQToAqKQ.pdf | openreview | benchmark/MD/ICLR2019_HJeQToAqKQ.md | ICLR 2019 |
B1gJOoRcYQ | {
"title": "S3TA: A Soft, Spatial, Sequential, Top-Down Attention Model",
"abstract": "We present a soft, spatial, sequential, top-down attention model (S3TA). This model uses a soft attention mechanism to bottleneck its view of the input. A recurrent core is used to generate query vectors, which actively select in... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10
]
},... | [
[
{
"role": "AC",
"data": {
"metareview": "1. Describe the strengths of the paper. As pointed out by the reviewers and based on your expert opinion.\n\nThe paper \n- tackles an interesting problem\n- makes a concerted effort to provide qualititative results that give insight into the models... | [
"This work presents a recurrent attention model as part of an RNN-based RL framework. The attention over the visual input is conditioned on the the model's state representation at time t.",
"Notably, this work incorporated multiple attention heads, each with differing behavior.",
"Pros:\n-Paper was easy to unde... | [
[
12
],
[
1
],
[
16
],
[
2
],
[
4
],
[
20
],
[
6
],
[
11
],
[
13
],
[
15
],
[
23
],
[
3
],
[
5
],
[
7
],
[
19
],
[
21
],
[
22
],
[
24
],
[
18
],
[
8
],
[
... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"cor... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"cor... | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0,
1
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
2
]
}
],
"category": [
"CLAR-WRT"... | benchmark/PDF/ICLR2019_B1gJOoRcYQ.pdf | openreview | benchmark/MD/ICLR2019_B1gJOoRcYQ.md | ICLR 2019 |
HJlEUoR9Km | {
"title": "Improved resistance of neural networks to adversarial images through generative pre-training",
"abstract": "We train a feed forward neural network with increased robustness against adversarial attacks compared to conventional training approaches. This is achieved using a novel pre-trained building block... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9
]
},
"scores... | [
[
{
"role": "AC",
"data": {
"metareview": "No reviewer has made a strong case for accepting this paper or championed it so I am recommending rejecting it. The unfavorable reviewers, although they mention real issues, have not highlighted some of the most important barriers to accepting this ... | [
"The recent work of Schott et al (which the authors compare results to) proposed the use of Bayes rule inversion as a more robust mechanism for classification under different types of adversarial attacks. The probabilities are approximated with variational autoencoders. During training the inference network is used... | [
[
10
],
[
5
],
[
7
],
[
12
],
[
14
],
[
15
],
[
18
],
[
1
],
[
2
],
[
3
],
[
4
],
[
8
],
[
9
],
[
11
],
[
16
],
[
17
],
[
19
],
[
13
],
[
0
],
[
6
]
] | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct"
] | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct"
] | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0,
1,
2
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
3
]
},
{
"role": "Au... | benchmark/PDF/ICLR2019_HJlEUoR9Km.pdf | openreview | benchmark/MD/ICLR2019_HJlEUoR9Km.md | ICLR 2019 |
r1xN5oA5tm | {
"title": "Phrase-Based Attentions",
"abstract": "Most state-of-the-art neural machine translation systems, despite being different\nin architectural skeletons (e.g., recurrence, convolutional), share an indispensable\nfeature: the Attention. However, most existing attention methods are token-based\nand ignore the... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
... | [
[
{
"role": "AC",
"data": {
"metareview": "All reviewers agree in their assessment that this paper does not meet the bar for ICLR. The area chair commends the authors for their detailed responses."
}
}
]
] | [
"The authors propose to include phrases (contiguous n-grams of wordpieces) in both the self-attention and encoder-decoder attention modules of the Transformer model (Vaswani et al., 2017). In standard multi-head attention, the logits of the attention distribution of each head is computed as the dot-product between ... | [
[
10
],
[
12
],
[
16
],
[
13
],
[
15
],
[
30
],
[
37
],
[
0,
24
],
[
6
],
[
22
],
[
25,
28
],
[
1
],
[
5,
26,
33
],
[
7
],
[
17
],
[
18
],
[
20
],
[
32
],
[... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"cor... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"i... | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0
]
}
],
"category": [
"ORIG-MTH"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1
]
},
{
"role": "Author",
"data"... | benchmark/PDF/ICLR2019_r1xN5oA5tm.pdf | openreview | benchmark/MD/ICLR2019_r1xN5oA5tm.md | ICLR 2019 |
SJl8gnAqtX | {
"title": "Prob2Vec: Mathematical Semantic Embedding for Problem Retrieval in Adaptive Tutoring",
"abstract": "We propose a new application of embedding techniques to problem retrieval in adaptive tutoring. The objective is to retrieve problems similar in mathematical concepts. There are two challenges: First, lik... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7
]
},
"scores": {
"Solid": null,
... | [
[
{
"role": "AC",
"data": {
"metareview": "I tend to agree with reviewers. This is a bit more of an applied type of work and does not lead to new insights in learning representations. \nLack of technical novelty\nDataset too small"
}
}
]
] | [
"This paper proposes a new application of embedding techniques for mathematical problem retrieval in adaptive tutoring. The proposed method performs much better than baseline sentence embedding methods.",
"Another contribution is on using negative pre-training to deal with an imbalanced training dataset.",
"To ... | [
[
4
],
[
17
],
[
13
],
[
5
],
[
8
],
[
1
],
[
6
],
[
9,
16
],
[
10
],
[
11
],
[
15
],
[
0,
3
],
[
2
],
[
14
],
[
7
],
[
12
]
] | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct"
] | [
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct"
] | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0,
1,
2
]
}
],
"category": [
"QUAL-MET"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
3,
4
]
},
{
... | benchmark/PDF/ICLR2019_SJl8gnAqtX.pdf | openreview | benchmark/MD/ICLR2019_SJl8gnAqtX.md | ICLR 2019 |
rkl3-hA5Y7 | {
"title": "Towards Decomposed Linguistic Representation with Holographic Reduced Representation",
"abstract": "The vast majority of neural models in Natural Language Processing adopt a form of structureless distributed representations. While these models are powerful at making predictions, the representational for... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
... | [
[
{
"role": "AC",
"data": {
"metareview": "This paper proposes the use of holographic reduced representations in language modeling, which allows for a cleaner decomposition of various linguistic traits in the representation. Results show improvements over baseline language models, and analys... | [
"Summary:\nTheis paper proposes a method for learning decomposable representations in the context of a language modeling task. Using holographic reduced representations (HRR), a word embedding is composed of a role and a filler. The embedding is then fed to an LSTM language model. There is also an extension to chun... | [
[
18
],
[
3
],
[
7
],
[
8
],
[
9
],
[
10
],
[
20
],
[
23
],
[
24
],
[
25
],
[
26
],
[
27
],
[
28
],
[
29
],
[
30
],
[
31
],
[
4,
33
],
[
5,
32,
38,
49,
58
... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"cor... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"cor... | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0,
1
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
2,
3
]
},
{
"role": "Au... | benchmark/PDF/ICLR2019_rkl3-hA5Y7.pdf | openreview | benchmark/MD/ICLR2019_rkl3-hA5Y7.md | ICLR 2019 |
rklhb2R9Y7 | {
"title": "Reinforced Imitation Learning from Observations",
"abstract": "Imitation learning is an effective alternative approach to learn a policy when the reward function is sparse. In this paper, we consider a challenging setting where an agent has access to a sparse reward function and state-only expert observ... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8
]
},
"scores": {
... | [
[
{
"role": "AC",
"data": {
"metareview": "This paper proposes to combine rewards obtained through IRL from rewards coming from the environment, and evaluate the algorithm on grid world environments. The problem setting is important and of interest to the ICLR community. While the revised pa... | [
"This paper proposes some new angles to the problem of imitation learning from state only observations (not state-action pairs which are more expensive).",
"Specifically, the paper proposes \"self exploration\", in which it mixes the imitation reward with environment reward from the MDP itself in a gradual manner... | [
[
19
],
[
20
],
[
21
],
[
16
],
[
17
],
[
18
],
[
22
],
[
1
],
[
2
],
[
4
],
[
6
],
[
7
],
[
10
],
[
14
],
[
15
],
[
3
],
[
8
],
[
9
],
[
11
],
[
12
],
[
... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct"
] | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct"
] | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0,
1,
2,
3,
4
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
5
]
},
... | benchmark/PDF/ICLR2019_rklhb2R9Y7.pdf | openreview | benchmark/MD/ICLR2019_rklhb2R9Y7.md | ICLR 2019 |
ByeNFoRcK7 | {
"title": "PA-GAN: Improving GAN Training by Progressive Augmentation",
"abstract": "Despite recent progress, Generative Adversarial Networks (GANs) still suffer from training instability, requiring careful consideration of architecture design choices and hyper-parameter tuning. The reason for this fragile trainin... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11
... | [
[
{
"role": "AC",
"data": {
"metareview": "The submission hypothesizes that in typical GAN training the discriminator is too strong, too fast, and thus suggests a modification by which they gradually increases the task difficulty of the discriminator. This is done by introducing (effectively... | [
"Authors argue that the main issue with stability in GANs is due to the discriminator becoming too powerful too quickly. To address this issue they propose to make the task progressively more difficult: Instead of providing only the samples to the discriminator, an additional (processed) bitstring is provided. The ... | [
[
2,
4,
8,
12
],
[
18
],
[
1
],
[
5
],
[
6
],
[
10,
15,
24
],
[
17
],
[
30
],
[
3
],
[
9
],
[
11
],
[
16
],
[
19
],
[
25
],
[
26
],
[
13,
21,
22
],
[
14
]... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"cor... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"i... | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1
]
}
],
"category": [
"ORIG-MTH"
]
},
... | benchmark/PDF/ICLR2019_ByeNFoRcK7.pdf | openreview | benchmark/MD/ICLR2019_ByeNFoRcK7.md | ICLR 2019 |
BJxLH2AcYX | {
"title": "Unsupervised Multi-Target Domain Adaptation: An Information Theoretic Approach",
"abstract": "Unsupervised domain adaptation (uDA) models focus on pairwise adaptation settings where there is a single, labeled, source and a single target domain. However, in many real-world settings one seeks to adapt to ... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2
]
},
"scores": {
"Solid": null,
"Presentation": null,
"Novelty": null,
"Overa... | [
[
{
"role": "AC",
"data": {
"metareview": "The paper proposes the unique setting of adapting to multiple target domains. The idea being that their approach may leverage commonality across domains to improve adaptation while maintaining domain specific parameters where needed. This idea and g... | [
"The biggest contribution is the setting part, where one seeks to adapt one source to multiple, but somewhat similar, target domains. It is interesting to explore such direction since in many real-world applications, applying the model to many different target domains are required.",
"It is also noted that there... | [
[
23
],
[
22
],
[
6
],
[
7
],
[
8
],
[
9
],
[
19
],
[
20
],
[
21
],
[
4
],
[
2,
3
],
[
15
],
[
0,
14
],
[
1
],
[
18
],
[
5
],
[
10
],
[
11
],
[
13
],
[
... | [
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct"
] | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"incorrect",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct"
] | [
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
0
]
}
],
"category": [
"ORIG-PROB"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1
]
},
{
"role": "Author",
"da... | benchmark/PDF/ICLR2019_BJxLH2AcYX.pdf | openreview | benchmark/MD/ICLR2019_BJxLH2AcYX.md | ICLR 2019 |
Ske25sC9FQ | {
"title": "Robustness and Equivariance of Neural Networks",
"abstract": "Neural networks models are known to be vulnerable to geometric transformations\nas well as small pixel-wise perturbations of input. Convolutional Neural Networks\n(CNNs) are translation-equivariant but can be easily fooled using rotations and... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
... | [
[
{
"role": "AC",
"data": {
"metareview": "Positives:\n\nThe paper proposes an interesting idea: to study the effect on vulnerability to adversarial attacks of training for invariance with respect to rotations.\nExperiments on MNIST, FashionMNIST, and CIFAR10.\nAn interesting hypothesis part... | [
"This paper empirically studies the robustness of equivariant CNNs to rotations as well as adversarial perturbations. It also studies their sample efficiency, parameter efficiency, and the effect of rotation- and adversarial augmentation during training and/or testing.\nThe main findings are:",
"1) Rotation-equiv... | [
[
8
],
[
11
],
[
7
],
[
18
],
[
5
],
[
14
],
[
20
],
[
22
],
[
4
],
[
9
],
[
10
],
[
12
],
[
13,
17
],
[
23
],
[
15
],
[
3
],
[
0
],
[
1
],
[
2
],
[
6
],
... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct"
] | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct"
] | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer",
"data": [
1
]
}
],
"category": [
"N/A"
]
},
{
... | benchmark/PDF/ICLR2019_Ske25sC9FQ.pdf | openreview | benchmark/MD/ICLR2019_Ske25sC9FQ.md | ICLR 2019 |
HygYqs0qKX | {
"title": "Conscious Inference for Object Detection",
"abstract": "Current Convolutional Neural Network (CNN)-based object detection models adopt strictly feedforward inference to predict the final detection results. However, the widely used one-way inference is agnostic to the global image context and the interpl... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8
]
},
"scores": {
... | [
[
{
"role": "AC",
"data": {
"metareview": "The paper presents an interesting idea, but there are significant concerns about the presentation issues and experimental results (e.g., comparisons with baselines). Overall, it is not ready for publication. "
}
}
]
] | [
"The paper proposes a iterative approach at inference time to improve object detections. The work relies on updating the feature activations and perform new feed forward passes to obtain improved results.",
"Pros:\n(+) The idea of iterative inference is potentially effective",
"(+) The paper is well written and... | [
[
14
],
[
2
],
[
11
],
[
13
],
[
1
],
[
7
],
[
5
],
[
10
],
[
3
],
[
8
],
[
9
],
[
12
],
[
16
],
[
4
],
[
15
],
[
0
],
[
6
]
] | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct"
] | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct"
] | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1
]
}
],
"category": [
"ORIG-MTH"
]
},
... | benchmark/PDF/ICLR2019_HygYqs0qKX.pdf | openreview | benchmark/MD/ICLR2019_HygYqs0qKX.md | ICLR 2019 |
SyxZOsA9tX | {
"title": "Accelerated Value Iteration via Anderson Mixing",
"abstract": "Acceleration for reinforcement learning methods is an important and challenging theme. We introduce the Anderson acceleration technique into the value iteration, developing an accelerated value iteration algorithm that we call Anderson Accel... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10
]
},... | [
[
{
"role": "AC",
"data": {
"metareview": "The paper proposes to use Anderson Mixing to accelerate value iteration and DQN. The idea is interesting, with some theoretical and empirical support. However, reviewers feel that the contribution is somewhat limited, and certain parts (e.g., the ... | [
"This paper seems like a nice idea, but I'm not sure if it's ready for publication. It seems that the main contribution of this paper is the DA2Q algorithm, since the A2VI algorithm is a straightforward application of AA to VI.",
"However the numerical examples are very weak, only 3 games are tested, and the resu... | [
[
3
],
[
14
],
[
15
],
[
19
],
[
7
],
[
16
],
[
21
],
[
2
],
[
23
],
[
0
],
[
10
],
[
1
],
[
11
],
[
12
],
[
13
],
[
20
],
[
5
],
[
9
],
[
4
],
[
6
],
[
... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"cor... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"cor... | [
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
0,
1,
2
]
},
{
"role": "Author",
"data": [
12,
13,
14,
15,
16,
17,
18,
29,
39,
... | benchmark/PDF/ICLR2019_SyxZOsA9tX.pdf | openreview | benchmark/MD/ICLR2019_SyxZOsA9tX.md | ICLR 2019 |
HyG1_j0cYQ | {
"title": "Pumpout: A Meta Approach for Robustly Training Deep Neural Networks with Noisy Labels",
"abstract": "It is challenging to train deep neural networks robustly on the industrial-level data, since labels of such data are heavily noisy, and their label generation processes are normally agnostic. To handle t... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5
]
},
"scores": {
"Solid": null,
"Presentation": null... | [
[
{
"role": "AC",
"data": {
"metareview": "The paper presents an approach to mitigate the presence of noisy labels during\ntraining by trying to forget wrong labels. Reviewers pointed out a few\nconcerns, including lack of novelty, lack of enough experimental support, and\nlack of theoretica... | [
"The paper proposes a meta algorithm to train a network with noisy labels.",
"It is not a general algorithm but a simple modification of two proposed methods. It is presented as a heuristics and it would be helpful to derive a theoretical framework or motivation for the proposed algorithm.",
"My main concern ... | [
[
10
],
[
12
],
[
13
],
[
14
],
[
19
],
[
7,
20
],
[
25
],
[
1
],
[
3
],
[
2
],
[
5,
6,
8
],
[
9
],
[
15
],
[
16
],
[
17
],
[
18
],
[
26
],
[
21
],
[
11
]... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"cor... | [
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"c... | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1
]
},
{
"role": "Author",
"data": [
... | benchmark/PDF/ICLR2019_HyG1_j0cYQ.pdf | openreview | benchmark/MD/ICLR2019_HyG1_j0cYQ.md | ICLR 2019 |
HklnzhR9YQ | {
"title": "Approximation and non-parametric estimation of ResNet-type convolutional neural networks via block-sparse fully-connected neural networks",
"abstract": "We develop new approximation and statistical learning theories of convolutional neural networks (CNNs) via the ResNet-type structure where the channel ... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
... | [
[
{
"role": "AC",
"data": {
"metareview": "The paper presents an interesting treatment of transforming a block-sparse fully connected neural networks to a ResNet-type Convolutional Network. Equipped with recent development on approximations of function classes (Barron, Holder) via block-spar... | [
"The paper studies approximation and estimation properties of CNNs with residual blocks in the context",
"of non-parametric regression, by constructing equivalent fully-connected architectures (with a block-sparse structure),",
"and leveraging previous approximation results for such functions.",
"Explicit ris... | [
[
5
],
[
6
],
[
9
],
[
10
],
[
11
],
[
12
],
[
13
],
[
4
],
[
7
],
[
8
],
[
16
],
[
17,
21
],
[
18
],
[
23
],
[
1
],
[
15
],
[
3
],
[
20
],
[
22
],
[
2
],... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct"
] | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct"
] | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0,
1,
2,
3,
4,
5,
6,
7,
8
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1"... | benchmark/PDF/ICLR2019_HklnzhR9YQ.pdf | openreview | benchmark/MD/ICLR2019_HklnzhR9YQ.md | ICLR 2019 |
ByGVui0ctm | {
"title": "Three continual learning scenarios and a case for generative replay",
"abstract": "Standard artificial neural networks suffer from the well-known issue of catastrophic forgetting, making continual or lifelong learning problematic. Recently, numerous methods have been proposed for continual learning, but... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
... | [
[
{
"role": "AC",
"data": {
"metareview": "The authors have proposed 3 continual learning variants which are all based on MNIST and which vary in terms of whether task ids are given and what the classification task is, and they have proposed a method which incorporates a symmetric VAE for ge... | [
"This paper points out a important issue in current continual learning literature: Due to the different settings and different evaluation protocols of each method, comparison between methods are usually not fair, and lead to distinct conclusions.",
"The paper is in general easy to understand except a few drawback... | [
[
1
],
[
10
],
[
14
],
[
20
],
[
0,
2
],
[
4
],
[
11
],
[
22
],
[
24
],
[
3
],
[
6
],
[
7
],
[
8
],
[
9
],
[
15
],
[
21
],
[
5
],
[
12
],
[
13
],
[
16
],
... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"cor... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"cor... | [
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
0
]
}
],
"category": [
"ORIG-ANL"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1
]
}
],
"category": [
"CLAR-WRT"
]... | benchmark/PDF/ICLR2019_ByGVui0ctm.pdf | openreview | benchmark/MD/ICLR2019_ByGVui0ctm.md | ICLR 2019 |
r1lM_sA5Fm | {
"title": "Assumption Questioning: Latent Copying and Reward Exploitation in Question Generation",
"abstract": "Question generation is an important task for improving our ability to process natural language data, with additional challenges over other sequence transformation tasks. Recent approaches use modificatio... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8
]
},
"scores": {
... | [
[
{
"role": "AC",
"data": {
"metareview": "This paper investigates copying mechanisms and reward functions in sequence to sequence models for question generation. The key findings are threefold: (1) when the alignments between input and output are weak, it is better to use latent copying mec... | [
"In the paper, author investigate the use of copy mechanisms for the question generation task. It evaluates on the SQuAD dataset. The model is a popular seq2seq/encoder-decoder model with copy mechanisms using pointer networks.",
"Pros:\nIt is well motivated. For the question generation task, a word to be predict... | [
[
18
],
[
13
],
[
2,
9
],
[
10
],
[
15
],
[
1
],
[
3
],
[
7
],
[
11
],
[
4
],
[
5
],
[
6
],
[
17
],
[
12
],
[
14
],
[
0
],
[
8
],
[
16
]
] | [
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct"
] | [
"correct",
"correct",
"correct",
"correct",
"incorrect",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect"
] | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1
]
}
],
"category": [
"ORIG-ANL"
]
},
... | benchmark/PDF/ICLR2019_r1lM_sA5Fm.pdf | openreview | benchmark/MD/ICLR2019_r1lM_sA5Fm.md | ICLR 2019 |
Syez3j0cKX | {
"title": "Dissecting an Adversarial framework for Information Retrieval",
"abstract": "Recent advances in Generative Adversarial Networks facilitated by improvements to the framework and successful application to various problems has resulted in extensions to multiple domains. IRGAN attempts to leverage the frame... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
... | [
[
{
"role": "AC",
"data": {
"metareview": "The manuscript centers on a critique of IRGAN, a recently proposed extension of GANs to the information retrieval setting, and introduces a competing procedure. \n\nReviewers found the findings and the proposed alternative to be interesting and in o... | [
"This paper trains an information retrieval (IR) model by contrasting the joint query-document distributions, p(q, d) with negative samples drawn from a resampling of the product of marginals, p(q) x p(d). They use a second discriminator to provide the re-weighting (I believe picking to top negative sample from the... | [
[
7
],
[
10
],
[
13,
30
],
[
14
],
[
37
],
[
3,
8
],
[
4
],
[
9,
12
],
[
11
],
[
20
],
[
35
],
[
2
],
[
19,
23,
25,
29,
31
],
[
1
],
[
5
],
[
15
],
[
27
]... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"cor... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"cor... | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0,
1
]
}
],
"category": [
"QUAL-CMP"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
2
]
},
{
"role": "Author",
... | benchmark/PDF/ICLR2019_Syez3j0cKX.pdf | openreview | benchmark/MD/ICLR2019_Syez3j0cKX.md | ICLR 2019 |
BklACjAqFm | {
"title": "Successor Uncertainties: exploration and uncertainty in temporal difference learning",
"abstract": "We consider the problem of balancing exploration and exploitation in sequential decision making problems. This trade-off naturally lends itself to probabilistic modelling. For a probabilistic approach to ... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
... | [
[
{
"role": "AC",
"data": {
"metareview": "Pros:\n- interesting algorithmic idea for using successor features to propagate uncertainty for use in epxloration\n- clarity\n\nCons:\n- moderate novelty\n- initially only simplistic experiments (later complemented with Atari results)\n- initially ... | [
"This paper tackles the classical exploration / exploitation problem in reinforcement learning.",
"The paper argues that it is necessary to propagate uncertainty correctly and argue that they can do so using the successor representation to compute the Bayesian posterior over Q-values conditioned on the data alrea... | [
[
12
],
[
4
],
[
21
],
[
1
],
[
3
],
[
8
],
[
9
],
[
13
],
[
14,
23
],
[
2
],
[
7
],
[
10
],
[
11,
22
],
[
5
],
[
6
],
[
16
],
[
17
],
[
18
],
[
19
],
[
... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct"
] | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct"
] | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0,
1
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
2
]
},
{
"role": "Author",
... | benchmark/PDF/ICLR2019_BklACjAqFm.pdf | openreview | benchmark/MD/ICLR2019_BklACjAqFm.md | ICLR 2019 |
HylKJhCcKm | {
"title": "Generalized Capsule Networks with Trainable Routing Procedure",
"abstract": "CapsNet (Capsule Network) was first proposed by Sabour et al. (2017) and lateranother version of CapsNet was proposed by Hinton et al. (2018). CapsNet hasbeen proved effective in modeling spatial features with much fewer param... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
... | [
[
{
"role": "AC",
"data": {
"metareview": "The paper proposes to replace dynamic routing in Capsule networks with a trainable layer that produces routing coefficients. The goal is to improve their scalability. This is promising as a research direction but reviewers have raised several concer... | [
"Pros:\nThe paper claims to make CapsuleNet's routing trainable. The proposed G-CapsNet have two variants (within feature map and across feature map).",
"It presents evaluation of G-CapsNet in terms of robustness and generalization. It is interesting that Capsule networks are as bad as traditional CNNs for strong... | [
[
11
],
[
13
],
[
3
],
[
21
],
[
27
],
[
33
],
[
2
],
[
10
],
[
12
],
[
18
],
[
31
],
[
34
],
[
0,
29
],
[
7
],
[
9
],
[
14
],
[
15
],
[
16,
28
],
[
1
],
[
... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"cor... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"incorrect",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
... | [
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
0
]
}
],
"category": [
"ORIG-MTH"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1
]
}
],
"category": [
"ORIG-NEG"
]... | benchmark/PDF/ICLR2019_HylKJhCcKm.pdf | openreview | benchmark/MD/ICLR2019_HylKJhCcKm.md | ICLR 2019 |
Skluy2RcK7 | {
"title": "Selectivity metrics can overestimate the selectivity of units: a case study on AlexNet",
"abstract": "Various methods of measuring unit selectivity have been developed in order to understand the representations learned by neural networks (NNs). Here we undertake a comparison of four such measures on Al... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7
]
},
"scores": {
"Solid": null,
... | [
[
{
"role": "AC",
"data": {
"metareview": "The paper examined the folk-knowledge that there are highly selective units in popular CNN architectures, and performed a detailed analysis of recent measures of unit selectivity, as well as introducing a novel one. The finding that units are not ex... | [
"This is a paper with scattered potentially interesting ideas.",
"But the execution is limited and the writing poor with critical details lacking. A major limitation of the paper is that it is not clear what contribution it makes. Some of the analyses are indeed interesting but 1) these analyses are mostly descr... | [
[
7
],
[
11
],
[
0
],
[
12
],
[
4
],
[
5
],
[
6
],
[
17
],
[
14
],
[
1
],
[
8
],
[
9
],
[
10
],
[
15
],
[
16
],
[
2
],
[
18
],
[
3
],
[
13
]
] | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct"
] | [
"incorrect",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct"
] | [
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
0,
1,
2,
3,
4,
6
]
},
{
"role": "Author",
"data": [
8,
9,
12,
20,
21,
22,
... | benchmark/PDF/ICLR2019_Skluy2RcK7.pdf | openreview | benchmark/MD/ICLR2019_Skluy2RcK7.md | ICLR 2019 |
BJxYEsAqY7 | {
"title": "FEED: Feature-level Ensemble Effect for knowledge Distillation",
"abstract": "This paper proposes a versatile and powerful training algorithm named Feature-level Ensemble Effect for knowledge Distillation(FEED), which is inspired by the work of factor transfer. The factor transfer is one of the knowledg... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
... | [
[
{
"role": "AC",
"data": {
"metareview": "The paper describes knowledge distillation methods. As noted by all reviewers, the methods are very similar to the prior art, so there is not enough novelty for the paper to be accepted. The reviewers' opinion didn't change after the rebuttal."
... | [
"In this paper, the authors present two methods, Sequential and Parallel-FEED for learning student networks that share architectures with their teacher.",
"Firstly, it would be a good idea to cite https://arxiv.org/abs/1312.6184, it precedes knowledge distillation and is basically the same thing minus a temperatu... | [
[
2,
25
],
[
9,
21
],
[
22
],
[
3,
10,
16,
18,
27
],
[
13
],
[
30
],
[
1
],
[
4,
24
],
[
5
],
[
11
],
[
23,
26
],
[
6
],
[
19
],
[
8
],
[
28
],
[
31
],
[
... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"cor... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"cor... | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1
]
}
],
"category": [
"QUAL-CMP"
]
},
... | benchmark/PDF/ICLR2019_BJxYEsAqY7.pdf | openreview | benchmark/MD/ICLR2019_BJxYEsAqY7.md | ICLR 2019 |
SkgzYiRqtX | {
"title": "Graph Neural Networks with Generated Parameters for Relation Extraction",
"abstract": "Recently, progress has been made towards improving relational reasoning in machine learning field. Among existing models, graph neural networks (GNNs) is one of the most effective approaches for multi-hop relational r... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
... | [
[
{
"role": "AC",
"data": {
"metareview": "+ experiments on an interesting task: inferring relations which are not necessarily explicitly mentioned in a sentence but need to be induced relying on other relations\n+ the idea to frame the relation prediction task as an inference task on a grap... | [
"This work proposes a method for parametrising relation matrices in graph neural networks (GNNs) using text. The model is applied in a relation extraction task, and specific dataset subsections are identified to test and analyse “hopping” behaviour: a model’s ability to combine multiple relations for inferring a ne... | [
[
31
],
[
34
],
[
8
],
[
11
],
[
14
],
[
20
],
[
21
],
[
30,
37
],
[
32
],
[
33
],
[
36
],
[
4
],
[
9
],
[
12
],
[
13
],
[
16
],
[
39,
40
],
[
2
],
[
3
],
[... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"cor... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"c... | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1
]
}
],
"category": [
"SIGN-SOT"
]
},
... | benchmark/PDF/ICLR2019_SkgzYiRqtX.pdf | openreview | benchmark/MD/ICLR2019_SkgzYiRqtX.md | ICLR 2019 |
rJxug2R9Km | {
"title": "Meta-Learning for Contextual Bandit Exploration",
"abstract": "We describe MÊLÉE, a meta-learning algorithm for learning a good exploration policy in the interactive contextual bandit setting. Here, an algorithm must take actions based on contexts, and learn based only on a reward signal from the action... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10
]
},... | [
[
{
"role": "AC",
"data": {
"metareview": "This paper provides an interesting strategy for learning to explore, by first training on fully supervised data before deploying that policy to an online setting. There are some concerns, however, on the realism and utility of this setting that shou... | [
"This paper investigates a meta-learning approach for the contextual bandit problem. The goal is to learn a generic exploration policy from datasets, and then to apply the exploration policy to contextual bandit tasks. The authors have adapted an algorithm proposed for imitation learning (Ross & Bagnell 2014) to th... | [
[
18
],
[
6
],
[
7
],
[
11
],
[
14,
23
],
[
19
],
[
22
],
[
8,
9,
17,
21
],
[
4
],
[
13
],
[
1
],
[
2,
10
],
[
3,
20
],
[
16
],
[
5,
15
],
[
12
],
[
0
]
] | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct"
] | [
"correct",
"incorrect",
"incorrect",
"correct",
"incorrect",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"corr... | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1,
2
]
},
{
"role": "Author",
... | benchmark/PDF/ICLR2019_rJxug2R9Km.pdf | openreview | benchmark/MD/ICLR2019_rJxug2R9Km.md | ICLR 2019 |
B1xnPsA5KX | {
"title": "Modular Deep Probabilistic Programming",
"abstract": "Modularity is a key feature of deep learning libraries but has not been fully exploited for probabilistic programming. We propose to improve modularity of probabilistic programming language by offering not only plain probabilistic distributions but a... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
... | [
[
{
"role": "AC",
"data": {
"metareview": "This paper presents a probabilistic programming language where models are constructed out of building blocks which specify both the distribution and an inference procedure. As a demonstration, they show how a GP-LVM can be implemented.\n\nThe paper ... | [
"The paper presents an extension of the MXFusion language that allows the use of probabilistic modules. These modules are defined as a set of random variables and a specific probabilistic distribution. The modules also contain dedicated inference methods. Using these modules, one can use probabilistic distributions... | [
[
5
],
[
7
],
[
8
],
[
9
],
[
10
],
[
11
],
[
12
],
[
13
],
[
26
],
[
30
],
[
35
],
[
36
],
[
37
],
[
14
],
[
19
],
[
1,
18
],
[
31
],
[
32
],
[
21
],
[
22
... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"cor... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"cor... | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0,
1
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
2,
4
]
},
{
"role": "Au... | benchmark/PDF/ICLR2019_B1xnPsA5KX.pdf | openreview | benchmark/MD/ICLR2019_B1xnPsA5KX.md | ICLR 2019 |
rygo9iR9F7 | {
"title": "Progressive Weight Pruning Of Deep Neural Networks Using ADMM",
"abstract": "Deep neural networks (DNNs) although achieving human-level performance in many domains, have very large model size that hinders their broader applications on edge computing devices. Extensive research work have been conducted o... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7
]
},
"scores": {
"Solid": null,
... | [
[
{
"role": "AC",
"data": {
"metareview": "The paper proposes a progressive pruning technique that achieves high pruning ratio. Reviewers have a consensus on rejection. Reviewer 1 pointed out that the experimental results are weak. Reviewer 2 is also concerned about the proposed method and e... | [
"This paper proposed a progressive weight pruning approach to compress the learned weights in DNN. My major concerns about the paper are as follows:",
"1. Novelty: The proposed approach heavily relies on the one in (Zhang et. al. 2018b) as shown in Sec. 2.2 for 1 page, making the paper as being an incremental wor... | [
[
4
],
[
10
],
[
3
],
[
1
],
[
5
],
[
13
],
[
17
],
[
9
],
[
15
],
[
20
],
[
2
],
[
6
],
[
12
],
[
19
],
[
21
],
[
14
],
[
11
],
[
0
],
[
7
],
[
8
],
[
... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct"
] | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct"
] | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1
]
},
{
"role": "Author",
"data": [
... | benchmark/PDF/ICLR2019_rygo9iR9F7.pdf | openreview | benchmark/MD/ICLR2019_rygo9iR9F7.md | ICLR 2019 |
H1gZV30qKQ | {
"title": "Transfer Value or Policy? A Value-centric Framework Towards Transferrable Continuous Reinforcement Learning",
"abstract": "Transferring learned knowledge from one environment to another is an important step towards practical reinforcement learning (RL). In this paper, we investigate the problem of trans... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
... | [
[
{
"role": "AC",
"data": {
"metareview": "The paper studies whether the best strategy for transfer learning in RL is to transfer value estimates or policy probabilities. The paper also presents a model-based value-centric (MVC) framework for continuous RL. The reviewers raised concerns rega... | [
"This paper proposes a model-based value-centric (MVC) framework for transfer learning in continuous RL problems, and an algorithm within that framework. The paper attempts to answer two questions: (1) \"why are current RL algorithms so inefficient in transfer learning\" and (2) \"what kind of RL algorithms could b... | [
[
12
],
[
33
],
[
10,
26
],
[
11
],
[
13
],
[
14
],
[
15
],
[
16
],
[
17
],
[
21,
29
],
[
27
],
[
1
],
[
36
],
[
35
],
[
8
],
[
32
],
[
5
],
[
18
],
[
24
],
... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"cor... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"incorrect",
... | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0,
1
]
}
],
"category": [
"SIGN-SOT"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
2
]
}
],
"category": [
"ORIG... | benchmark/PDF/ICLR2019_H1gZV30qKQ.pdf | openreview | benchmark/MD/ICLR2019_H1gZV30qKQ.md | ICLR 2019 |
Bkeuz20cYm | {
"title": "Double Neural Counterfactual Regret Minimization",
"abstract": "Counterfactual regret minimization (CRF) is a fundamental and effective technique for solving imperfect information games. However, the original CRF algorithm only works for discrete state and action spaces, and the resulting strategy is ma... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
... | [
[
{
"role": "AC",
"data": {
"metareview": "The reviewers agreed that there are some promising ideas in this work, and useful empirical analysis to motivate the approach. The main concern is in the soundness of the approach (for example, comments about cumulative learning and negative samples... | [
"========= Summary =========\nThe authors propose \"Double Neural CFR\", which uses neural network function approximation in place of the tabular update in CFR. CFR is the leading method for finding equilibria in imperfect information games.",
"However it is typically employed with a tabular policy, limiting its ... | [
[
16
],
[
17
],
[
20
],
[
22
],
[
19,
25
],
[
21
],
[
6,
23
],
[
8
],
[
10
],
[
15
],
[
18
],
[
5
],
[
9
],
[
11,
12
],
[
13
],
[
14
],
[
26
],
[
27,
28
],
... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"cor... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
... | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0,
1,
2
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer",
"data": [
3
]
}
],
"category": [
... | benchmark/PDF/ICLR2019_Bkeuz20cYm.pdf | openreview | benchmark/MD/ICLR2019_Bkeuz20cYm.md | ICLR 2019 |
HJM4rsRqFX | {
"title": "Neural Variational Inference For Embedding Knowledge Graphs",
"abstract": "Recent advances in Neural Variational Inference allowed for a renaissance in latent variable models in a variety of domains involving high-dimensional data. In this paper, we introduce two generic Variational Inference frameworks... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
... | [
[
{
"role": "AC",
"data": {
"metareview": "The paper proposes a novel variational inference framework for knowledge graphs which is evaluated on link prediction benchmark sets and is competitive to previous generative approaches.\nWhile the idea is interstnig and technically correct, the ori... | [
"The paper presents two variational inference frameworks for generative models of knowledge graphs. Such models are based respectively on latent fact model and latent information model.",
"The authors argue that with the presented framework the underlying probabilistic semantics can be discovered. Experiments sho... | [
[
6
],
[
13
],
[
14
],
[
29
],
[
5
],
[
7
],
[
30
],
[
31
],
[
1
],
[
3
],
[
4,
20
],
[
8
],
[
11
],
[
15
],
[
16
],
[
17
],
[
25
],
[
28
],
[
22
],
[
24
... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"cor... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"cor... | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0,
1
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
2
]
}
],
"category": [
"CLAR-WRT"... | benchmark/PDF/ICLR2019_HJM4rsRqFX.pdf | openreview | benchmark/MD/ICLR2019_HJM4rsRqFX.md | ICLR 2019 |
HJglg2A9FX | {
"title": "Iteratively Learning from the Best",
"abstract": "We study a simple generic framework to address the issue of bad training data; both bad labels in supervised problems, and bad samples in unsupervised ones. Our approach starts by fitting a model to the whole training dataset, but then iteratively improv... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10
]
},... | [
[
{
"role": "AC",
"data": {
"metareview": "This paper addresses the problem of learning with outliers, which many reviewers agree is an important direction. However, reviewers point to issues with the experiments (missing baselines, ablations, etc.) and are concerned that the assumptions in ... | [
"This paper provides an algorithm that excludes the bad training data in the training process and obtain a more accurate model for both supervised and unsupervised learning problem. The paper gives the theoretical guarantee for mixed linear regression and Gaussian mixture model, and also conducts the experiments fo... | [
[
4
],
[
9
],
[
10
],
[
1
],
[
3,
16
],
[
5
],
[
20
],
[
7
],
[
8
],
[
11,
17
],
[
12
],
[
15
],
[
18
],
[
23
],
[
2
],
[
19
],
[
21
],
[
22
],
[
14
],
[
... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct"
] | [
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct"
] | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1,
2,
3
]
},
{
"role": "Au... | benchmark/PDF/ICLR2019_HJglg2A9FX.pdf | openreview | benchmark/MD/ICLR2019_HJglg2A9FX.md | ICLR 2019 |
SyxXhsAcFQ | {
"title": "Cohen Welling bases & SO(2)-Equivariant classifiers using Tensor nonlinearity.",
"abstract": "In this paper we propose autoencoder architectures for learning a Cohen-Welling\n(CW)-basis for images and their rotations. We use the learned CW-basis to build\na rotation equivariant classifier to classify im... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
... | [
[
{
"role": "AC",
"data": {
"metareview": "This paper studies group equivariant neural network representations by building on the work by [Cohen and Welling, '14], which introduced learning of group irreducible representations, and [Kondor'18], who introduced tensor product non-linearities o... | [
"Review: This paper deals with the issue of learning rotation invariant autoencoders and classifiers.",
"While this problem is well motivated, I found that this paper was fairly weak experimentally, and I also found it difficult to determine what the exact algorithm was.",
"For example, how the optimization was... | [
[
35
],
[
4
],
[
5
],
[
14
],
[
21,
38
],
[
23
],
[
25
],
[
26
],
[
28
],
[
17
],
[
6
],
[
7
],
[
13
],
[
20
],
[
32
],
[
34
],
[
0,
19
],
[
8
],
[
18
],
[
... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"cor... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"cor... | [
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
0,
1
]
},
{
"role": "Author",
"data": [
14,
15,
18,
142,
143,
144,
145,
146,
147,
1... | benchmark/PDF/ICLR2019_SyxXhsAcFQ.pdf | openreview | benchmark/MD/ICLR2019_SyxXhsAcFQ.md | ICLR 2019 |
rkxfjjA5Km | {
"title": "Auto-Encoding Knockoff Generator for FDR Controlled Variable Selection",
"abstract": "A new statistical procedure (Candès,2018) has provided a way to identify important factors using any supervised learning method controlling for FDR. This line of research has shown great potential to expand the horiz... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4
]
},
"scores": {
"Solid": null,
"Presentation": null,
"No... | [
[
{
"role": "AC",
"data": {
"metareview": "The paper presents a novel strategy for statistically motivated feature selection i.e. aimed at controlling the false discovery rate. This is achieved by extending knockoffs to complex predictive models and complex distributions; specifically using ... | [
"In the paper , the authors proposed the use of autoencoder for Model-X knockoffs. The authors proved that, if there exists latent factors, and if the encoders and the decoders can approximate conditional distributions well, the autoencoder can be used for approximating Model-X knockoff random variables: one can fi... | [
[
23
],
[
19
],
[
20
],
[
25
],
[
26,
28
],
[
27
],
[
29
],
[
10
],
[
11
],
[
14,
18
],
[
16
],
[
24
],
[
15,
22
],
[
9
],
[
1
],
[
2
],
[
3
],
[
7,
12,
17
... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"cor... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
... | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1
]
}
],
"category": [
"QUAL-MET",
"QUAL... | benchmark/PDF/ICLR2019_rkxfjjA5Km.pdf | openreview | benchmark/MD/ICLR2019_rkxfjjA5Km.md | ICLR 2019 |
B1fysiAqK7 | {
"title": "Probabilistic Binary Neural Networks",
"abstract": "Low bit-width weights and activations are an effective way of combating the increasing need for both memory and compute power of Deep Neural Networks. In this work, we present a probabilistic training method for Neural Network with both binary weights ... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8
]
},
"scores": {
... | [
[
{
"role": "AC",
"data": {
"metareview": "The paper proposes a probabilistic training method for binary Neural Network with stochastic versions of Batch Normalization and max pooling.\n\nThe reviewers and AC note the following potential weaknesses: (1) limited novelty and (2) preliminary ex... | [
"## Summary\nThis work presents a probabilistic training method for binary Neural Network with stochastic versions of Batch Normalization and max pooling. By sampling from the weight distribution an ensemble of Binary Neural Networks could further improve the performance. In the experimental section, the authors co... | [
[
6
],
[
1
],
[
7,
9,
12,
13
],
[
14
],
[
5,
16
],
[
17
],
[
2
],
[
3
],
[
4
],
[
10
],
[
15
],
[
11
],
[
0
],
[
8
]
] | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct"
] | [
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct"
] | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1
]
},
{
"role": "Author",
"data": [
... | benchmark/PDF/ICLR2019_B1fysiAqK7.pdf | openreview | benchmark/MD/ICLR2019_B1fysiAqK7.md | ICLR 2019 |
S14g5s09tm | {
"title": "Unseen Action Recognition with Unpaired Adversarial Multimodal Learning",
"abstract": "In this paper, we present a method to learn a joint multimodal representation space that allows for the recognition of unseen activities in videos. We compare the effect of placing various constraints on the embedding... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
... | [
[
{
"role": "AC",
"data": {
"metareview": "The paper received mixed reviews. The proposed ideas are reasonable and it shows that unpaired data can improve the performance of unseen video (action) classification tasks and other related tasks. The authors rightfully argue that the main contrib... | [
"This paper proposes a joint embedding model that aligns video sequences with sentences describing the context (caption) in a shared embedding space. With the space, various tasks such as zero-shot activity recognition and unseen video captioning can be performed. The problem tackled in this paper is interesting.",... | [
[
7
],
[
22
],
[
3
],
[
11
],
[
18
],
[
1,
4,
16
],
[
5
],
[
12
],
[
20
],
[
21
],
[
8
],
[
9
],
[
10
],
[
13
],
[
17
],
[
19
],
[
2
],
[
15
],
[
0
],
[
... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct"
] | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"incorrect",
"incorrect",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct"
] | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1,
2
]
},
{
"role": "Author",
... | benchmark/PDF/ICLR2019_S14g5s09tm.pdf | openreview | benchmark/MD/ICLR2019_S14g5s09tm.md | ICLR 2019 |
rkeqCoA5tX | {
"title": "LEARNING GENERATIVE MODELS FOR DEMIXING OF STRUCTURED SIGNALS FROM THEIR SUPERPOSITION USING GANS",
"abstract": "Recently, Generative Adversarial Networks (GANs) have emerged as a popular alternative for modeling complex high dimensional distributions. Most of the existing works implicitly assume that t... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
... | [
[
{
"role": "AC",
"data": {
"metareview": "The paper proposes two simple generator architecture variants enabling the use of GAN training for the tasks of denoising (from known noise types) and demixing (of two added sources). While the denoising approach is very similar to AmbientGAN and co... | [
"Quality is good, just a handful of typos.",
"Claritys above average in explaining the problem setting.\nOriginality: scan refs...",
"Significance: medium\nPros: the authors develop a novel GAN-based approach to denoising, demixing, and in the process train generators for the various components (not just infere... | [
[
0
],
[
1
],
[
2,
9
],
[
3
],
[
6
],
[
4,
13
],
[
5
],
[
10
],
[
11
],
[
12
],
[
14
],
[
7
],
[
8
]
] | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct"
] | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct"
] | [
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
0
]
},
{
"role": "Author",
"data": [
16
]
}
],
"category": [
"CLAR-WRT"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"da... | benchmark/PDF/ICLR2019_rkeqCoA5tX.pdf | openreview | benchmark/MD/ICLR2019_rkeqCoA5tX.md | ICLR 2019 |
Hkl-di09FQ | {
"title": "Decoupling feature extraction from policy learning: assessing benefits of state representation learning in goal based robotics",
"abstract": "Scaling end-to-end reinforcement learning to control real robots from vision presents a series of challenges, in particular in terms of sample efficiency. Against... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7
]
},
"scores": {
"Solid": null,
... | [
[
{
"role": "AC",
"data": {
"metareview": "This paper proposes a new method for combining previous state representation learning methods and compares to end-to-end learning without without separately learning a state representation. The topic is important, and the authors have made an extens... | [
"The paper is easy to read and the presentation is clear, and I really appreciate this.",
"The authors address the very important topic of feature extraction and state representation learning. New results in this area are always valuable and welcome.",
"However, my feeling is that the paper falls short in terms... | [
[
9
],
[
0
],
[
12
],
[
13
],
[
14
],
[
15
],
[
3
],
[
4
],
[
2
],
[
10
],
[
16
],
[
17
],
[
18
],
[
5
],
[
8,
11
],
[
22
],
[
23
],
[
25
],
[
26
],
[
27
... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"cor... | [
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
... | [
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
0
]
}
],
"category": [
"CLAR-WRT"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1
]
}
],
"category": [
"SIGN-BRD"
]... | benchmark/PDF/ICLR2019_Hkl-di09FQ.pdf | openreview | benchmark/MD/ICLR2019_Hkl-di09FQ.md | ICLR 2019 |
B1gHjoRqYQ | {
"title": "An Efficient and Margin-Approaching Zero-Confidence Adversarial Attack",
"abstract": "There are two major paradigms of white-box adversarial attacks that attempt to impose input perturbations. The first paradigm, called the fix-perturbation attack, crafts adversarial samples within a given perturbation... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6
]
},
"scores": {
"Solid": null,
"Pres... | [
[
{
"role": "AC",
"data": {
"metareview": "The paper proposes a new method for adversarial attacks, MarginAttack, which finds adversarial examples with small distortion and runs faster than the CW baseline, but slower than other methods. The authors provide theoretical guarantees and a broad... | [
"i have change my rating from 5 to 6 after reading the numerous and thorough rebuttals from the authors. I hope they will incorporate these clarifications and additional experiments into the final version of the paper if accepted.",
"The purpose of this paper is presumably to approximate the margin of a sample as... | [
[
15
],
[
1,
8,
11
],
[
10
],
[
13
],
[
7
],
[
14
],
[
5
],
[
6
],
[
9
],
[
17
],
[
19
],
[
20
],
[
21
],
[
2,
22
],
[
12
],
[
3
],
[
0
],
[
4
],
[
16
],
... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct"
] | [
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect"
] | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1,
2
]
},
{
"role": "Author",
... | benchmark/PDF/ICLR2019_B1gHjoRqYQ.pdf | openreview | benchmark/MD/ICLR2019_B1gHjoRqYQ.md | ICLR 2019 |
H1xmqiAqFm | {
"title": "Investigating CNNs' Learning Representation under label noise",
"abstract": "Deep convolutional neural networks (CNNs) are known to be robust against label noise on extensive datasets. However, at the same time, CNNs are capable of memorizing all labels even if they are random, which means they can memo... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
... | [
[
{
"role": "AC",
"data": {
"metareview": "The paper analyzes the performance of CNN models when data is mislabelled in different manners.\n\nThe reviewers and AC note the critical limitation of novelty of this paper to meet the high standard of ICLR.\n\nAC thinks the proposed method has pot... | [
"This paper demonstrates that CNNs are more robust to class-relevant label noise. They argue that real-world noise should be class-relevant.\nPros:",
"1. The authors find a new angle to exploit robust learning with noisy labels.",
"2. The authors perform numerical experiments to demonstrate the effectiveness of... | [
[
11
],
[
4
],
[
5
],
[
17
],
[
6
],
[
7
],
[
16
],
[
1
],
[
2
],
[
8
],
[
10
],
[
12
],
[
13
],
[
9
],
[
0
],
[
3
],
[
14
],
[
15
]
] | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct"
] | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct"
] | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1
]
}
],
"category": [
"ORIG-MTH"
]
},
... | benchmark/PDF/ICLR2019_H1xmqiAqFm.pdf | openreview | benchmark/MD/ICLR2019_H1xmqiAqFm.md | ICLR 2019 |
HkGGfhC5Y7 | {
"title": "Towards a better understanding of Vector Quantized Autoencoders",
"abstract": "Deep neural networks with discrete latent variables offer the promise of better symbolic reasoning, and learning abstractions that are more useful to new tasks. There has been a surge in interest in discrete latent variable ... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
... | [
[
{
"role": "AC",
"data": {
"metareview": "Strengths:\n\n- well-written \n- strong results for non-autoregressive NMT\n- a novel soft EM version of VQ-VAE\n\nWeaknesses:\n\n- as pointed out by reviewers, the improvements are mostly not due to the VQ-VAE modification rather due to orthogonal... | [
"This paper introduces a new way of interpreting the VQ-VAE,",
"and proposes a new training algorithm based on the soft EM clustering.",
"I think the technical aspect of this paper is written concisely.",
"Introducing the interpretation as hard EM seems natural for me, and the extension",
"to the soft EM tr... | [
[
23
],
[
37
],
[
44
],
[
48
],
[
25
],
[
26
],
[
27
],
[
1
],
[
4
],
[
5,
15,
18
],
[
7,
11,
12
],
[
17
],
[
20
],
[
22
],
[
24
],
[
33
],
[
41
],
[
45
],
... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"cor... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"cor... | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0,
1
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
2
]
}
],
"category": [
"CLAR-WRT"... | benchmark/PDF/ICLR2019_HkGGfhC5Y7.pdf | openreview | benchmark/MD/ICLR2019_HkGGfhC5Y7.md | ICLR 2019 |
ByeTHsAqtX | {
"title": "Gradient Descent Happens in a Tiny Subspace",
"abstract": "We show that in a variety of large-scale deep learning scenarios the gradient dynamically converges to a very small subspace after a short period of training. The subspace is spanned by a few top eigenvectors of the Hessian (equal to the number ... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
... | [
[
{
"role": "AC",
"data": {
"metareview": "The paper is overally interesting and addresses an important problem, however reviewers ask for more rigorous empirical study and less restrictive settings."
}
}
]
] | [
"This paper shows that gradient descent mostly happens in a tiny subspace which is spanned by the top eigenvectors of the Hessian. Empirical results are shown to support the claim. This finding is interesting and provides us some insights to design more efficient optimization algorithms.",
"Overall, this paper is... | [
[
1,
9
],
[
21
],
[
0,
12,
17
],
[
10
],
[
20
],
[
2
],
[
4
],
[
5
],
[
6
],
[
7
],
[
15
],
[
16
],
[
19
],
[
13
],
[
18
],
[
8
],
[
14
],
[
3
],
[
11
]
] | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct"
] | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct"
] | [
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
0
]
}
],
"category": [
"ORIG-ANL"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1
]
}
],
"category": [
"CLAR-WRT"
]... | benchmark/PDF/ICLR2019_ByeTHsAqtX.pdf | openreview | benchmark/MD/ICLR2019_ByeTHsAqtX.md | ICLR 2019 |
HJx4KjRqYQ | {
"title": "Ergodic Measure Preserving Flows",
"abstract": "Training probabilistic models with neural network components is intractable in most cases and requires to use approximations such as Markov chain Monte Carlo (MCMC), which is not scalable and requires significant hyper-parameter tuning, or mean-field varia... | Reject | [
[
{
"role": "Reviewer",
"data": {
"value": {
"comment": [
0,
1,
2,
3,
4,
5
]
}
}
},
{
"role": "Author",
"data": {
"value": {
"comment": [
... | [
[
{
"role": "AC",
"data": {
"metareview": "This paper proposes to a simple method for tuning parameters of HMC by maximizing the log density under the final sample of the MCMC, and apply it for training VAE. The reviews and discussion raises some critical concerns and questions, which unfor... | [
"The comparison to results from the literature might be slightly misleading. Please clarify.",
"* The toughest baseline that is compared to in section 4.2 is from 2015. Please include some more recent results for reference. There exist several better results.",
"* You write\n\"In (Salimans et al., 2015), the te... | [
[
9
],
[
16
],
[
33
],
[
34
],
[
0
],
[
1
],
[
21
],
[
22
],
[
23,
32
],
[
26
],
[
24
],
[
3
],
[
2,
4,
7
],
[
5
],
[
10,
11
],
[
27
],
[
8,
12
],
[
13,
... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"cor... | [
"correct",
"correct",
"incorrect",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
... | [
{
"sentences": [
{
"role": "Reviewer 1 Further Reply",
"data": [
0
]
},
{
"role": "Author",
"data": [
6,
8,
9,
10,
11,
12,
13,
14
]
}
],
"cate... | benchmark/PDF/ICLR2019_HJx4KjRqYQ.pdf | openreview | benchmark/MD/ICLR2019_HJx4KjRqYQ.md | ICLR 2019 |
S1gBz2C9tX | {
"title": "Importance Resampling for Off-policy Policy Evaluation",
"abstract": "Importance sampling is a common approach to off-policy learning in reinforcement learning. While it is consistent and unbiased, it can result in high variance updates to the parameters for the value function. Weighted importance samp... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4
]
},
"scores": {
"Solid": null,
"Presentation": null,
"No... | [
[
{
"role": "AC",
"data": {
"metareview": "The paper proposes to use importance resampling (IR) as an alternative to the more popular importance sampling (IS) approach to off-policy RL. The hope is to reduce variance, as shown in experiments. However, there is no analysis why/when IR will ... | [
"The authors propose to use importance resampling (IR) in place of importance sampling (IS) for policy evaluation tasks. The method proposed by the authors definitely seems valid, but it isn’t quite clear when this is applicable.",
"IR is often used in the case of particle filters and other SMC is often used to c... | [
[
9
],
[
10
],
[
11
],
[
15
],
[
7,
13,
14
],
[
18
],
[
0
],
[
2
],
[
6
],
[
8
],
[
1
],
[
4
],
[
5
],
[
12
],
[
17
],
[
3
],
[
16
]
] | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct"
] | [
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct"
] | [
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
0
]
},
{
"role": "Author",
"data": [
6,
7,
8,
9,
10,
11,
12,
14,
18,
19
]
}
... | benchmark/PDF/ICLR2019_S1gBz2C9tX.pdf | openreview | benchmark/MD/ICLR2019_S1gBz2C9tX.md | ICLR 2019 |
SyNPk2R9K7 | {
"title": "Learning to Describe Scenes with Programs",
"abstract": "Human scene perception goes beyond recognizing a collection of objects and their pairwise relations. We understand higher-level, abstract regularities within the scene such as symmetry and repetition. Current vision recognition modules and scene r... | Accept (Poster) | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
... | [
[
{
"role": "AC",
"data": {
"comment": "Thank you all for your substantial reviews. As you can see, the authors have responded with detailed comments and rebuttals of their own, and revised the paper. It is imperative that you examine these rebuttals, the revisions made to the paper, and eac... | [
"This paper presents a system that infers programs describing 3D scenes composed of simple primitives. The system consists of three stages each of which is trained separately.",
"First, the perceptual module extracts object masks and their attributes. The objects are then are split into several groups.",
"Final... | [
[
1
],
[
10
],
[
12
],
[
13
],
[
16
],
[
15,
18
],
[
22
],
[
27
],
[
21
],
[
9
],
[
19
],
[
25
],
[
28
],
[
3,
6
],
[
4
],
[
5
],
[
7
],
[
8
],
[
11
],
[
... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"cor... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"incorrect",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct"... | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0,
1,
2
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
3
]
}
],
"category": [
... | benchmark/PDF/ICLR2019_SyNPk2R9K7.pdf | openreview | benchmark/MD/ICLR2019_SyNPk2R9K7.md | ICLR 2019 |
H1Gfx3Rqtm | {
"title": "End-to-End Hierarchical Text Classification with Label Assignment Policy",
"abstract": "We present an end-to-end reinforcement learning approach to hierarchical text classification where documents are labeled by placing them at the right positions in a given hierarchy.\nWhile existing “global” methods c... | Reject | [
[
{
"role": "Author",
"data": {
"value": {
"comment": [
0
]
}
}
}
],
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
1,
2,
... | [
[
{
"role": "AC",
"data": {
"metareview": "This paper presents a reinforcement learning approach to hierarchical text classification.\n\nPros: A potentially interesting idea to drive the search process over a hierachical set of labels using reinforcement learning.\n\nCons: The major concensu... | [
"We released our code in an anonymized repo: https://github.com/hi-label-assignment-policy/HiLAP",
"This paper presents an end to end rl approach for hierarchical text classification. The paper proposes a label assignment policy for determining the appropropriate positioning of a document in a hierarchy. It is ba... | [
[
10
],
[
11,
21
],
[
12
],
[
14
],
[
30
],
[
31
],
[
32
],
[
33
],
[
0
],
[
28
],
[
4
],
[
5,
7
],
[
6
],
[
13,
13,
23,
23
],
[
29
],
[
3
],
[
16,
25
],
[
... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"cor... | [
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"c... | [
{
"sentences": [
{
"role": "Reviewer 2",
"data": [
1
]
}
],
"category": [
"ORIG-MTH",
"QUAL-EXP"
]
},
{
"sentences": [
{
"role": "Reviewer",
"data": [
2
]
}
],
"category": [
... | benchmark/PDF/ICLR2019_H1Gfx3Rqtm.pdf | openreview | benchmark/MD/ICLR2019_H1Gfx3Rqtm.md | ICLR 2019 |
B1lxH20qtX | {
"title": "Learning to control self-assembling morphologies: a study of generalization via modularity",
"abstract": "Much of contemporary sensorimotor learning assumes that one is already given a complex agent (e.g., a robotic arm) and the goal is to learn to control it. In contrast, this paper investigates a modu... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10
]
},... | [
[
{
"role": "AC",
"data": {
"comment": "We are coming to the end of the discussion phase. \nThank you for the discussion around R1 comments.\nHearing back from R3 would be very useful.\nWe do realize that everyone's time is limited.\n-- area chair\n"
}
}
],
[
{
"role... | [
"Summary:\nThe paper considers the problem of constructing compositional robotic morphologies that can solve different continuous control tasks in a (multi-agent) reinforcement learning setting. The authors created an environment where the actor consists of a number of primitive components which interface with each... | [
[
4
],
[
5
],
[
11
],
[
9
],
[
19
],
[
1
],
[
21
],
[
13
],
[
14
],
[
15
],
[
16
],
[
17
],
[
7
],
[
3
],
[
2
],
[
6
],
[
8
],
[
22
],
[
12
],
[
20
],
[
... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct"
] | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct"
] | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1
]
}
],
"category": [
"ORIG-PROB"
]
},
... | benchmark/PDF/ICLR2019_B1lxH20qtX.pdf | openreview | benchmark/MD/ICLR2019_B1lxH20qtX.md | ICLR 2019 |
SJl8J30qFX | {
"title": "Learning Global Additive Explanations for Neural Nets Using Model Distillation",
"abstract": "Interpretability has largely focused on local explanations, i.e. explaining why a model made a particular prediction for a sample. These explanations are appealing due to their simplicity and local fidelity. Ho... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8
]
},
"scores": {
... | [
[
{
"role": "AC",
"data": {
"metareview": "This paper introduces a distillation approach for black-box classifiers that trains generalized additive models (GAM), an additive model over feature shapes, thus providing global explanations for the model. Given the importance of interpretability,... | [
"This paper is of high quality and clarity. I think it's originality is at least decent. Whether it is significant or not depends on how significant one thinks fully connected neural networks are as these are the models for which this explanation model makes sense.\nGood things:",
"- It is a very elegant method. ... | [
[
6
],
[
26
],
[
5
],
[
9,
17
],
[
12,
15,
19
],
[
1
],
[
10,
13,
16,
18,
21
],
[
16
],
[
18,
22
],
[
24
],
[
3
],
[
7
],
[
11
],
[
14
],
[
20
],
[
23
],
[
... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"cor... | [
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
... | [
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
0
]
},
{
"role": "Author",
"data": [
10,
11,
12,
13,
58
]
}
],
"category": [
"SIGN-BRD"
]
},
{
"sentenc... | benchmark/PDF/ICLR2019_SJl8J30qFX.pdf | openreview | benchmark/MD/ICLR2019_SJl8J30qFX.md | ICLR 2019 |
H1e6ij0cKQ | {
"title": "EFFICIENT SEQUENCE LABELING WITH ACTOR-CRITIC TRAINING",
"abstract": "Neural approaches to sequence labeling often use a Conditional Random Field (CRF) to model their output dependencies, while Recurrent Neural Networks (RNN) are used for the same purpose in other tasks. We set out to establish RNNs as ... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
... | [
[
{
"role": "AC",
"data": {
"metareview": "this is an interesting approach to use reinforcement learning to replace CRF for sequence tagging, which would potentially be beneficial when the tag set is gigantic. unfortunately the conducted experiments do not really show this, which makes it di... | [
"I found the paper difficult to follow. The method proposed is not well motivated, and the literature review explains well the novelty. Here are some questions/points for discussion:",
"- the token-level MLE training is not what causes the exposure bias: one can train with MLE and still avoid it by generating ap... | [
[
7
],
[
0,
21
],
[
25
],
[
2,
3,
27
],
[
4,
10
],
[
11
],
[
12
],
[
15
],
[
16
],
[
22
],
[
26
],
[
5,
17
],
[
9
],
[
1
],
[
6
],
[
8,
18
],
[
13
],
[
... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"cor... | [
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"c... | [
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
0
]
}
],
"category": [
"CLAR-WRT"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1
]
}
],
"category": [
"QUAL-MET"
]... | benchmark/PDF/ICLR2019_H1e6ij0cKQ.pdf | openreview | benchmark/MD/ICLR2019_H1e6ij0cKQ.md | ICLR 2019 |
S1gd7nCcF7 | {
"title": "Self-Supervised Generalisation with Meta Auxiliary Learning",
"abstract": "Auxiliary learning has been shown to improve the generalisation performance of a principal task. But typically, this requires manually-defined auxiliary tasks based on domain knowledge. In this paper, we consider that it may be p... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7
]
},
"scores": {
"Solid": null,
... | [
[
{
"role": "AC",
"data": {
"metareview": "This paper proposes a framework for generating auxiliary tasks as a means to regularize learning. The idea is interesting, and the method is simple. Two of the three reviewers found the paper to be well-written. The experiment include a promising re... | [
"Summary:\nThe role of auxiliary tasks is to improve the generalization performance of the principal task of interest.",
"So far, hand-crafted auxiliary tasks are generated, tailored for a problem of interest. The current work addresses a meta-learning approach to automatically generate auxiliary tasks suited to ... | [
[
9
],
[
15
],
[
2,
23
],
[
7
],
[
10
],
[
20
],
[
3
],
[
1,
17,
19
],
[
4
],
[
12
],
[
13
],
[
14
],
[
18
],
[
22
],
[
8,
11
],
[
5
],
[
21
],
[
0
],
[
... | [
"correct",
"incorrect",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct"
] | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct"
] | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0,
1
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
2
]
}
],
"category": [
"ORIG-MTH"... | benchmark/PDF/ICLR2019_S1gd7nCcF7.pdf | openreview | benchmark/MD/ICLR2019_S1gd7nCcF7.md | ICLR 2019 |
rkMhusC5Y7 | {
"title": "Learning to Coordinate Multiple Reinforcement Learning Agents for Diverse Query Reformulation",
"abstract": "We propose a method to efficiently learn diverse strategies in reinforcement learning for query reformulation in the tasks of document retrieval and question answering. In the proposed framework ... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11
... | [
[
{
"role": "AC",
"data": {
"metareview": "\npros:\n- The paper is clear and easy to read\n- Both Reviewer 1 and Reviewer 2 found the empirical evaluation to be good\n\ncons:\n- Some of the reviewers felt that the proposed approach lacked novelty (e.g. with respect to Nogueira and Cho)\n- So... | [
"In this paper, authors proposed an ensemble approach for query reformulation (QR). The basic idea is that 1) train a bunch of models/sub-agents on subsets, e.g., randomly partitioned, of the training data; 2) and then train an additional meta model/meta-agent to aggregate the results from the step 1). They condu... | [
[
8
],
[
13
],
[
15
],
[
1,
17
],
[
11
],
[
12
],
[
24
],
[
14
],
[
5,
19,
22
],
[
4
],
[
10
],
[
28
],
[
3
],
[
7
],
[
18
],
[
25
],
[
27
],
[
29
],
[
23
... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"cor... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"c... | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1
]
}
],
"category": [
"CLAR-WRT"
]
},
... | benchmark/PDF/ICLR2019_rkMhusC5Y7.pdf | openreview | benchmark/MD/ICLR2019_rkMhusC5Y7.md | ICLR 2019 |
r1eiqi09K7 | {
"title": "Riemannian Adaptive Optimization Methods",
"abstract": "Several first order stochastic optimization methods commonly used in the Euclidean domain such as stochastic gradient descent (SGD), accelerated gradient descent or variance reduced methods have already been adapted to certain Riemannian settings. ... | Accept (Poster) | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
... | [
[
{
"role": "AC",
"data": {
"metareview": "Dear authors,\n\nAll reviewers agreed that your work sheds new light on a popular class of algorithms and should thus be presented at ICLR.\n\nPlease make sure to implement all their comments in the final version."
}
}
]
] | [
"This paper presents Riemannian versions of adaptive optimization methods, including ADAGRAD, ADAM, AMSGRAD and ADAMNC. There are no natural coordinates on a manifold.",
"Therefore, the authors resort to product of manifolds and view each manifold component as a coordinate. Convergence analyses for those methods ... | [
[
5
],
[
1
],
[
2
],
[
4
],
[
12
],
[
14
],
[
15
],
[
0
],
[
3
],
[
13
],
[
6
],
[
11
],
[
10
],
[
8
],
[
9
],
[
16
],
[
18
],
[
7
],
[
17
]
] | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct"
] | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct"
] | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0,
1,
2,
3
]
}
],
"category": [
"CLAR-WRT"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
4,
5
]
... | benchmark/PDF/ICLR2019_r1eiqi09K7.pdf | openreview | benchmark/MD/ICLR2019_r1eiqi09K7.md | ICLR 2019 |
rJxcHnRqYQ | {
"title": "Local Binary Pattern Networks for Character Recognition",
"abstract": "Memory and computation efficient deep learning architectures are crucial to the continued proliferation of machine learning capabilities to new platforms and systems. Binarization of operations in convolutional neural networks has sh... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6
]
},
"scores": {
"Solid": null,
"Pres... | [
[
{
"role": "AC",
"data": {
"metareview": "This paper proposed a LBPNet for character recognition, which introduces the LBP feature extraction into deep learning. Reviewers are confused on implementation and not convinced on experiments. The only score 6 reviewer is also concerned \"Empirica... | [
"This paper proposed a LBPNet for character recognition, which introduces the LBP feature extraction into deep learning. Personally I think that this idea is interesting for improving the efficiency of CNNs, as traditionally LBP has been demonstrated its good performance and efficiency in some vision tasks such as ... | [
[
11
],
[
2
],
[
15
],
[
0
],
[
16
],
[
7
],
[
17
],
[
18
],
[
10
],
[
19
],
[
3
],
[
4
],
[
5
],
[
6,
14,
20
],
[
8
],
[
9
],
[
12
],
[
13
],
[
1
]
] | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct"
] | [
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct"
] | [
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
0
]
}
],
"category": [
"ORIG-COM"
]
},
{
"sentences": [
{
"role": "Reviewer",
"data": [
1
]
}
],
"category": [
"N/A"
]
},
... | benchmark/PDF/ICLR2019_rJxcHnRqYQ.pdf | openreview | benchmark/MD/ICLR2019_rJxcHnRqYQ.md | ICLR 2019 |
r1f78iAcFm | {
"title": "GRAPH TRANSFORMATION POLICY NETWORK FOR CHEMICAL REACTION PREDICTION",
"abstract": "We address a fundamental problem in chemistry known as chemical reaction product prediction. Our main insight is that the input reactant and reagent molecules can be jointly represented as a graph, and the process of gen... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
... | [
[
{
"role": "AC",
"data": {
"metareview": "The reviewers and authors participated in modest discussion, with the authors providing direct responses to reviewer comments. However, this did not appreciably change the overall ratings of the paper (one reviewer raised their rating, while another... | [
"The paper provides a new system that combines a number of neural networks to predict chemical reactions. The paper brings together a number of interesting methods to create a system that outperforms the state of the art.\nGood about this paper:",
"- reported performance: the authors report a small but very consi... | [
[
27
],
[
7
],
[
13
],
[
4
],
[
9
],
[
10
],
[
12
],
[
17
],
[
26
],
[
28
],
[
31,
36
],
[
2
],
[
30
],
[
37
],
[
20
],
[
15
],
[
8
],
[
21
],
[
32
],
[
35
... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"cor... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
... | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1
]
}
],
"category": [
"SIGN-SOT"
]
},
... | benchmark/PDF/ICLR2019_r1f78iAcFm.pdf | openreview | benchmark/MD/ICLR2019_r1f78iAcFm.md | ICLR 2019 |
HyxnZh0ct7 | {
"title": "Meta-learning with differentiable closed-form solvers",
"abstract": "Adapting deep networks to new concepts from a few examples is challenging, due to the high computational requirements of standard fine-tuning procedures.\nMost work on few-shot learning has thus focused on simple learning techniques fo... | Accept (Poster) | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
... | [
[
{
"role": "AC",
"data": {
"metareview": "The reviewers disagree strongly on this paper. Reviewer 2 was the most positive, believing it to be an interesting contribution with strong results. Reviewer 3 however, was underwhelmed by the results. Reviewer 1 does not believe that the contributi... | [
"This paper proposes a meta-learning approach for the problem of few-shot classification. Their method, based on parametrizing the learner for each task by a closed-form solver, strikes an interesting compromise between not performing any adaptation for each new task (as is the case in pure metric learning methods)... | [
[
15
],
[
6
],
[
7
],
[
8
],
[
14
],
[
0,
13
],
[
11,
12
],
[
3
],
[
5
],
[
9
],
[
16
],
[
4
],
[
1
],
[
2
],
[
10
]
] | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct"
] | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"incorrect",
"correct",
"correct",
"correct",
"incorrect"
] | [
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
0
]
}
],
"category": [
"ORIG-MTH"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1
]
}
],
"category": [
"QUAL-MET"
]... | benchmark/PDF/ICLR2019_HyxnZh0ct7.pdf | openreview | benchmark/MD/ICLR2019_HyxnZh0ct7.md | ICLR 2019 |
HyevnsCqtQ | {
"title": "Integral Pruning on Activations and Weights for Efficient Neural Networks",
"abstract": "With the rapidly scaling up of deep neural networks (DNNs), extensive research studies on network model compression such as weight pruning have been performed for efficient deployment. This work aims to advance the ... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
... | [
[
{
"role": "AC",
"data": {
"metareview": "This paper proposes to compress the deep learning model using both activation pruning and weight pruning. The reviewers have a consensus on rejection due to lack of novelty. "
}
}
]
] | [
"This article presents a novel approach called Integral Pruning (IP) to reduce the computation cost of Deep neural networks (DNN) by integrating activation pruning along with weight pruning. The authors show that common techniques of exclusive weight pruning does compress the model size, but increases the number of... | [
[
18
],
[
3
],
[
11
],
[
19
],
[
32
],
[
7,
26
],
[
0
],
[
10,
12,
13,
14,
17
],
[
20
],
[
21
],
[
22
],
[
27,
28
],
[
2,
5,
24
],
[
6
],
[
9
],
[
16
],
[
... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"cor... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"cor... | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0,
1
]
}
],
"category": [
"ORIG-MTH"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
2
]
}
],
"category": [
"SIGN... | benchmark/PDF/ICLR2019_HyevnsCqtQ.pdf | openreview | benchmark/MD/ICLR2019_HyevnsCqtQ.md | ICLR 2019 |
S1xiOjC9F7 | {
"title": "Graph Matching Networks for Learning the Similarity of Graph Structured Objects",
"abstract": "This paper addresses the challenging problem of retrieval and matching of graph structured objects, and makes two key contributions. First, we demonstrate how Graph Neural Networks (GNN), which have emerged a... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
... | [
[
{
"role": "AC",
"data": {
"metareview": "This is a tough choice as it is a reasonably strong paper.\nI am similar to another reviewer quite confused how this graph matching can \"only focus on important nodes in the graph\"\nThis seems counter-intuitive and the only reason given in the reb... | [
"The authors present two methods for learning a similarity score between pairs of graphs. They first is to use a shared GNN for each graph to produce independent graph embeddings on which a similarity score is computed. The authors improve this model using pairs of graphs as input and utilizing a cross-graph attent... | [
[
2,
11
],
[
4
],
[
9
],
[
1
],
[
22
],
[
5
],
[
12
],
[
24
],
[
0
],
[
3
],
[
7
],
[
10
],
[
15
],
[
17,
20
],
[
18
],
[
19
],
[
21
],
[
23
],
[
8
],
[
... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"cor... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
... | [
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
0
]
},
{
"role": "Author",
"data": [
28,
29,
30,
31,
32
]
}
],
"category": [
"QUAL-EXP"
]
},
{
"sentenc... | benchmark/PDF/ICLR2019_S1xiOjC9F7.pdf | openreview | benchmark/MD/ICLR2019_S1xiOjC9F7.md | ICLR 2019 |
r1gGpjActQ | {
"title": "Hint-based Training for Non-Autoregressive Translation",
"abstract": "Machine translation is an important real-world application, and neural network-based AutoRegressive Translation (ART) models have achieved very promising accuracy. Due to the unparallelizable nature of the autoregressive factorization... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
... | [
[
{
"role": "AC",
"data": {
"metareview": "\n\n+ sufficiently strong results\n\n+ a fast / parallelizable model\n\n\n- Novelty with respect to previous work is not as great (see AnonReviewer1 and AnonReviewer2's comments)\n\n- The same reviewers raised concerns about the discussion of relate... | [
"This work proposes a non-autoregressive Neural Machine Translation model which the authors call NART, as opposed to an autoregressive model which is referred to as an ART model. The main idea behind this work is to leverage a well trained ART model to inform the hidden states and the word alignment of NART models.... | [
[
1,
27
],
[
20
],
[
4
],
[
28
],
[
2
],
[
3
],
[
5
],
[
16,
21
],
[
17,
22
],
[
18
],
[
19
],
[
25
],
[
29
],
[
30
],
[
6
],
[
7
],
[
8
],
[
9,
10,
12
],... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"cor... | [
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"c... | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0,
1,
2,
3
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
4,
5,
6,
... | benchmark/PDF/ICLR2019_r1gGpjActQ.pdf | openreview | benchmark/MD/ICLR2019_r1gGpjActQ.md | ICLR 2019 |
Ske1-209Y7 | {
"title": "Probabilistic Model-Based Dynamic Architecture Search",
"abstract": "The architecture search methods for convolutional neural networks (CNNs) have shown promising results. These methods require significant computational resources, as they repeat the neural network training many times to evaluate and sea... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4
]
},
"scores": {
"Solid": null,
"Presentation": null,
"No... | [
[
{
"role": "AC",
"data": {
"metareview": "The paper presents an architecture search method which jointly optimises the architecture and its weights. As noted by reviewers, the method is very close to Shirakawa et al., with the main innovation being the use of categorical distributions to mo... | [
"This paper presents a joint optimization approach for the continuous weights and categorical structures of neural networks. The idea is the standard stochastic relaxation of introducing a parametrised distribution over the categorical parameters and marginalising it. The method then follows by alternating gradient... | [
[
10
],
[
16
],
[
1,
6,
19
],
[
4
],
[
9
],
[
21
],
[
2
],
[
7
],
[
3
],
[
12
],
[
13
],
[
20
],
[
23
],
[
11
],
[
14
],
[
22
],
[
8
],
[
17
],
[
18
],
[
... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct"
] | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct"
] | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1
]
},
{
"role": "Author",
"data": [
... | benchmark/PDF/ICLR2019_Ske1-209Y7.pdf | openreview | benchmark/MD/ICLR2019_Ske1-209Y7.md | ICLR 2019 |
SyxknjC9KQ | {
"title": "Dense Morphological Network: An Universal Function Approximator",
"abstract": "Artificial neural networks are built on the basic operation of linear combination and non-linear activation function. Theoretically this structure can approximate any continuous function with three layer architecture. But in ... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
... | [
[
{
"role": "AC",
"data": {
"metareview": "This work presents an interesting take on how to combine basic functions to lead to better activation functions. While the experiments in the paper show that the approach works well compared to the baselines that are used as reference, reviewers not... | [
"The authors introduce Morph-Net, a single layer neural network where",
"the mapping is performed using morphological dilation and erosion.",
"I was expecting something applied to convolutional networks as such operators",
"are very popular in image processing, so the naming is a bit misleading.",
"It is sh... | [
[
6
],
[
1
],
[
3,
26
],
[
7
],
[
9
],
[
13
],
[
16
],
[
2,
30
],
[
0,
15
],
[
17,
27
],
[
20
],
[
8,
14
],
[
23,
28,
31
],
[
10
],
[
11
],
[
12
],
[
18
]... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"cor... | [
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
... | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0,
1
]
}
],
"category": [
"ORIG-MTH"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
2,
3
]
},
{
"role"... | benchmark/PDF/ICLR2019_SyxknjC9KQ.pdf | openreview | benchmark/MD/ICLR2019_SyxknjC9KQ.md | ICLR 2019 |
r1znKiAcY7 | {
"title": "Few-shot Classification on Graphs with Structural Regularized GCNs",
"abstract": "We consider the fundamental problem of semi-supervised node classification in attributed graphs with a focus on \\emph{few-shot} learning. Here, we propose Structural Regularized Graph Convolutional Networks (SRGCN), novel... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8
]
},
"scores": {
... | [
[
{
"role": "AC",
"data": {
"metareview": "A new regularized graph CNN approach is proposed for semi-supervised learning on graphs. The conventional Graph CNN is concatenated with a Transposed Network, which is used to supplement the supervised loss w.r.t. the labeled part of the graph with... | [
"Edited: I raised the score by 1 point after the authors revised the paper significantly.",
"This paper proposes a regularization approach for improving GCN when the training examples are very few. The regularization is the reconstruction loss of the node features under an autoencoder. The encoder is the usual GC... | [
[
4
],
[
21
],
[
22
],
[
23
],
[
2
],
[
6
],
[
11
],
[
17
],
[
19
],
[
12,
13
],
[
14
],
[
15
],
[
5
],
[
9
],
[
10
],
[
24
],
[
25
],
[
26
],
[
27
],
[
1
... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"cor... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"cor... | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0,
1
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
2
]
}
],
"category": [
"QUAL-MET"... | benchmark/PDF/ICLR2019_r1znKiAcY7.pdf | openreview | benchmark/MD/ICLR2019_r1znKiAcY7.md | ICLR 2019 |
BkxSHsC5FQ | {
"title": "SupportNet: solving catastrophic forgetting in class incremental learning with support data",
"abstract": "A plain well-trained deep learning model often does not have the ability to learn new knowledge without forgetting the previously learned knowledge, which is known as catastrophic forgetting. Here ... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8
]
},
"scores": {
... | [
[
{
"role": "AC",
"data": {
"metareview": "The authors propose using a SVM, trained as a last layer of a neural network, to identify exemplars (support vectors) to save and use to prevent forgetting as the model is trained on further tasks. The method is effective on several supervised bench... | [
"This paper presents a continual learning method that aims to overcome the catastrophic forgetting problem by holding out small number of samples for each task to be used in training for new tasks. Specifially, these representative samples for each task are selected as support vectors of a SVM trained on it. The pr... | [
[
37
],
[
34
],
[
35
],
[
30
],
[
36
],
[
12,
27
],
[
28
],
[
2,
8,
13
],
[
3
],
[
4
],
[
11
],
[
14
],
[
17
],
[
18
],
[
19
],
[
20
],
[
10
],
[
15,
21
],
... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"cor... | [
"correct",
"correct",
"correct",
"incorrect",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
... | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1
]
}
],
"category": [
"QUAL-MET"
]
},
... | benchmark/PDF/ICLR2019_BkxSHsC5FQ.pdf | openreview | benchmark/MD/ICLR2019_BkxSHsC5FQ.md | ICLR 2019 |
HJlmhs05tm | {
"title": "EnGAN: Latent Space MCMC and Maximum Entropy Generators for Energy-based Models",
"abstract": "Unsupervised learning is about capturing dependencies between variables and is driven by the contrast between the probable vs improbable configurations of these variables, often either via a generative model w... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
... | [
[
{
"role": "AC",
"data": {
"metareview": "The proposed method is an extension of Kim & Bengio (2016)'s energy-based GAN. The novel contributions are to approximate the entropy regularizer using a mutual information estimator, and to try to clean up the model samples using some Langevin step... | [
"Thank you for an interesting read.",
"The paper proposes an approximate training technique for energy-based models (EBMs). More specifically, the samples used negative phase gradient in EBM training is approximated by samples from another generator. This \"approximate generator\" is a composition of a decoder (w... | [
[
8
],
[
9
],
[
10,
28
],
[
14,
24
],
[
2
],
[
17
],
[
25
],
[
18
],
[
22
],
[
26
],
[
3,
4,
13,
16,
20,
23
],
[
6,
15
],
[
7
],
[
11
],
[
5
],
[
27
],
[
... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"inc... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"cor... | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1
]
}
],
"category": [
"SIGN-SOT"
]
},
... | benchmark/PDF/ICLR2019_HJlmhs05tm.pdf | openreview | benchmark/MD/ICLR2019_HJlmhs05tm.md | ICLR 2019 |
BkGiPoC5FX | {
"title": "Efficient Convolutional Neural Network Training with Direct Feedback Alignment",
"abstract": "There were many algorithms to substitute the back-propagation (BP) in the deep neural network (DNN) training. However, they could not become popular because their training accuracy and the computational efficie... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9
]
},
"scores... | [
[
{
"role": "AC",
"data": {
"metareview": "This paper proposes a training algorithm for ConvNet architectures in which the final few layers are fully connected. The main idea is to use direct feedback alignment with carefully chosen binarized (±1) weights to train the fully connected layers... | [
"This manuscript extends the direct feedback alignment (DFA) approach to convolutional neural networks (CNN) by (1) only applying DFA to FC layers with backpropagation (BP) in place for convolutional layers (2) using binary numbers for feedback matrix.",
"Originality wise, I think (1) is a very straightforward ex... | [
[
5
],
[
9
],
[
15
],
[
16
],
[
17
],
[
18
],
[
23
],
[
21
],
[
1,
6
],
[
25
],
[
14
],
[
2
],
[
3
],
[
4
],
[
7
],
[
8,
10,
20
],
[
11
],
[
13
],
[
27
],... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"cor... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"cor... | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1
]
},
{
"role": "Author",
"data": [
... | benchmark/PDF/ICLR2019_BkGiPoC5FX.pdf | openreview | benchmark/MD/ICLR2019_BkGiPoC5FX.md | ICLR 2019 |
SklcFsAcKX | {
"title": "Deep Denoising: Rate-Optimal Recovery of Structured Signals with a Deep Prior",
"abstract": "Deep neural networks provide state-of-the-art performance for image denoising, where the goal is to recover a near noise-free image from a noisy image.\nThe underlying principle is that neural networks trained o... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8
]
},
"scores": {
... | [
[
{
"role": "AC",
"data": {
"metareview": "The paper analyzes the interesting problem of image denoising with neural networks by imposing simplifying assumptions on the Gaussianity and independence of the prior. A bound is established from the analysis of (Hand & Voroninksi, 2018) that can ... | [
"The paper studies the standard denoising problem under the assumption that the unknown n-dimensional signal can be written as the output of a known d-layer neural network G mapping k dimensions to n dimensions. The paper specifies an algorithm to perform this denoising and the algorithm is based on a variant of th... | [
[
1,
14
],
[
5
],
[
16
],
[
7,
11
],
[
15
],
[
19
],
[
3,
10
],
[
4
],
[
12
],
[
17
],
[
18
],
[
2
],
[
6
],
[
9
],
[
0
],
[
8
],
[
13
]
] | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct"
] | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct"
] | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1
]
}
],
"category": [
"ORIG-ANL"
]
},
... | benchmark/PDF/ICLR2019_SklcFsAcKX.pdf | openreview | benchmark/MD/ICLR2019_SklcFsAcKX.md | ICLR 2019 |
SkghN205KQ | {
"title": "Search-Guided, Lightly-supervised Training of Structured Prediction Energy Networks",
"abstract": " In structured output prediction tasks, labeling ground-truth training output is often expensive. However, for many tasks, even when the true output is unknown, we can evaluate predictions using a scalar ... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
... | [
[
{
"role": "AC",
"data": {
"metareview": "This paper proposes search-guided training for structured prediction energy networks (SPENs).\n\nThe reviewers found some interest in this approach, though were somewhat underwhelmed by the experimental comparison and the details provided about the ... | [
"The paper proposes to use a reward function to guide the learning of energy-based models for structured prediction. The idea is to update the energy function based on a random search algorithm guided by a reward function. At each iteration, the SPEN proposes a solution, then a better one is found by the search alg... | [
[
1
],
[
12
],
[
7
],
[
16
],
[
20
],
[
2
],
[
4
],
[
9
],
[
19
],
[
3
],
[
5
],
[
6
],
[
8
],
[
10
],
[
14
],
[
13,
18
],
[
21
],
[
15
],
[
0
],
[
11
],
... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct"
] | [
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct"
] | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1
]
},
{
"role": "Author",
"data": [
... | benchmark/PDF/ICLR2019_SkghN205KQ.pdf | openreview | benchmark/MD/ICLR2019_SkghN205KQ.md | ICLR 2019 |
HJguLo0cKQ | {
"title": "Strength in Numbers: Trading-off Robustness and Computation via Adversarially-Trained Ensembles",
"abstract": "While deep learning has led to remarkable results on a number of challenging problems, researchers have discovered a vulnerability of neural networks in adversarial settings, where small but ca... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
... | [
[
{
"role": "AC",
"data": {
"metareview": "The work brings little novelty compared to existing literature. "
}
}
]
] | [
"Summary. The paper considers the robustness of neural nets against adversarial attacks.",
"More precisely, the authors experimentally investigate the robustness of ensembles of neural nets. They empirically show that adversarially trained ensembles of 2 neural nets are more robust than ensembles of 2 adversarial... | [
[
29
],
[
2,
23
],
[
8
],
[
11
],
[
19
],
[
26
],
[
30
],
[
32
],
[
21,
31
],
[
28,
33,
36
],
[
37
],
[
10
],
[
13
],
[
16
],
[
20
],
[
3
],
[
4,
22,
27
],
... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"cor... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"c... | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0,
1
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer",
"data": [
2
]
}
],
"category": [
"N/A"
]
... | benchmark/PDF/ICLR2019_HJguLo0cKQ.pdf | openreview | benchmark/MD/ICLR2019_HJguLo0cKQ.md | ICLR 2019 |
rkGcYi09Km | {
"title": "NUTS: Network for Unsupervised Telegraphic Summarization",
"abstract": "Extractive summarization methods operate by ranking and selecting the sentences which best encapsulate the theme of a given document. They do not fare well in domains like fictional narratives where there is no central theme and cor... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5
]
},
"scores": {
"Solid": null,
"Presentation": null... | [
[
{
"role": "AC",
"data": {
"metareview": "This paper presents methods for telegraphic summarization, a task that generates extremely short summaries.\n\nThere are concerns about the utility of the task in general, and also the novelty of the modeling framework.\n\nThere is overall consensus... | [
"The paper explores unsupervised deep learning model for extractive telegraphic summaries, which extracts text fragments (e.g., fragments of a sentence) as summaries. The paper is in general well structured and is easy to follow.",
"However, I think the submission does not have enough content to be accepted to th... | [
[
18
],
[
0
],
[
7
],
[
2
],
[
6
],
[
10,
13
],
[
4
],
[
5
],
[
14
],
[
16
],
[
17
],
[
3
],
[
8
],
[
11
],
[
15
],
[
1
],
[
9
],
[
19
],
[
20
],
[
21
],
... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct"
] | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"incorrect",
"incorrect",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct"
] | [
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
0
]
}
],
"category": [
"CLAR-WRT"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1
]
}
],
"category": [
"QUAL-MET"
]... | benchmark/PDF/ICLR2019_rkGcYi09Km.pdf | openreview | benchmark/MD/ICLR2019_rkGcYi09Km.md | ICLR 2019 |
BkfbpsAcF7 | {
"title": "Excessive Invariance Causes Adversarial Vulnerability",
"abstract": "Despite their impressive performance, deep neural networks exhibit striking failures on out-of-distribution inputs. One core idea of adversarial example research is to reveal neural network errors under such distribution shifts. We dec... | Accept (Poster) | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3
]
},
"scores": {
"Solid": null,
"Presentation": null,
"Novelty": null,
... | [
[
{
"role": "AC",
"data": {
"metareview": "This paper studies the roots of the existence of adversarial perspective from a new perspective. This perspective is quite interesting and thought-provoking. However, some of the contributions rely on fairly restrictive assumptions and/or are not pr... | [
"The paper focuses on adversarial vulnerability of neural networks, and more specifically on perturbation-based versus invariance-based adversarial examples and how using bijective networks (with so-called metameric sampling) may help overcoming issues related to invariance. The approach is used to get around insuf... | [
[
7
],
[
3
],
[
11
],
[
1
],
[
14
],
[
2
],
[
16
],
[
9
],
[
18
],
[
19
],
[
6
],
[
10
],
[
13,
13,
15,
15
],
[
5
],
[
8
],
[
0
],
[
4
],
[
12
],
[
17
],
... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct"
] | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct"
] | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1,
2
]
},
{
"role": "Author",
... | benchmark/PDF/ICLR2019_BkfbpsAcF7.pdf | openreview | benchmark/MD/ICLR2019_BkfbpsAcF7.md | ICLR 2019 |
HJMRvsAcK7 | {
"title": "Dynamic Pricing on E-commerce Platform with Deep Reinforcement Learning",
"abstract": "In this paper we develop an approach based on deep reinforcement learning (DRL) to address dynamic pricing problem on E-commerce platform. We models real-world E-commerce dynamic pricing problem as Markov Decision Pr... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11
... | [
[
{
"role": "AC",
"data": {
"metareview": "\nThis is an interesting topic but the reviewers had substantial concerns on the clarity and significance of the contribution.\n"
}
}
]
] | [
"In this paper, the authors study the problem of Dynamic Pricing. This is a well-studied problem in Economics, Operations Research, and Computer Science. The basic problem is to find the right price for a product based on repeated interaction with the market. The significant challenge in this problem is to figure t... | [
[
36
],
[
37
],
[
38
],
[
1,
8,
20
],
[
2
],
[
21,
30,
31,
32,
33
],
[
6
],
[
7
],
[
34
],
[
35
],
[
40
],
[
22
],
[
13,
16
],
[
3,
4
],
[
5
],
[
12,
15
]... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"cor... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
... | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0,
1
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
2
]
},
{
"role": "Author",
... | benchmark/PDF/ICLR2019_HJMRvsAcK7.pdf | openreview | benchmark/MD/ICLR2019_HJMRvsAcK7.md | ICLR 2019 |
SyfXKoRqFQ | {
"title": "Ada-Boundary: Accelerating the DNN Training via Adaptive Boundary Batch Selection",
"abstract": "Neural networks can converge faster with help from a smarter batch selection strategy. In this regard, we propose Ada-Boundary, a novel adaptive-batch selection algorithm that constructs an effective mini-ba... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9
]
},
"scores... | [
[
{
"role": "AC",
"data": {
"metareview": "This paper introduced an adaptive importance sampling strategy to select mini-batches to speed up the convergence of network training. The method is well motivated and easy to follow.\n\nThe main concerns raised by the reviewers are limited novelty ... | [
"This paper attempts to speed up convergence of deep neural networks by intelligently selecting batches. The experiments show this method works moderately well.",
"This paper appears quite similar to the recent work \"Active Bias\" [1].",
"The motivation for the technique and setting appear very similar, while ... | [
[
9
],
[
10
],
[
14
],
[
20
],
[
5
],
[
7
],
[
11
],
[
1,
4
],
[
15
],
[
19
],
[
3
],
[
8,
12
],
[
16
],
[
17,
22
],
[
2
],
[
18
],
[
21
],
[
0
],
[
6
],
... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct"
] | [
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct"
] | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1,
2
]
},
{
"role": "Author",
... | benchmark/PDF/ICLR2019_SyfXKoRqFQ.pdf | openreview | benchmark/MD/ICLR2019_SyfXKoRqFQ.md | ICLR 2019 |
SyezvsC5tX | {
"title": "The loss landscape of overparameterized neural networks",
"abstract": "We explore some mathematical features of the loss landscape of overparameterized neural networks. A priori one might imagine that the loss function looks like a typical function from $\\mathbb{R}^n$ to $\\mathbb{R}$ - in particular,... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6
]
},
"scores": {
"Solid": null,
"Pres... | [
[
{
"role": "AC",
"data": {
"comment": "I am wondering whether the results imply that all the global minima form a *connected* n-d dimensional manifold? "
}
},
{
"role": "Author",
"data": {
"value": {
"comment": "That is an excellent question, than... | [
"The paper shows that the set of global minimums of an overparametrized network with smooth activation function is almost surely a high dimensional manifold.",
"In particular, the dimension of this manifold is exactly n-d, where n is the number of parameters and d is the number of datas. To the best of my knowled... | [
[
13
],
[
23
],
[
24
],
[
25
],
[
26
],
[
27
],
[
28
],
[
29
],
[
32,
33
],
[
1,
12
],
[
2
],
[
5
],
[
8,
9,
10,
11
],
[
21
],
[
22
],
[
15
],
[
3
],
[
4
... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"cor... | [
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
... | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1
]
}
],
"category": [
"ORIG-ANL"
]
},
... | benchmark/PDF/ICLR2019_SyezvsC5tX.pdf | openreview | benchmark/MD/ICLR2019_SyezvsC5tX.md | ICLR 2019 |
r1f0YiCctm | {
"title": "Minimal Random Code Learning: Getting Bits Back from Compressed Model Parameters",
"abstract": "While deep neural networks are a highly successful model class, their large memory footprint puts considerable strain on energy consumption, communication bandwidth, and storage requirements. Consequently, mo... | Accept (Poster) | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7
]
},
"scores": {
"Solid": null,
... | [
[
{
"role": "AC",
"data": {
"metareview": "This paper proposes a novel coding scheme for compressing neural network weights using Shannon-style coding and a variational distribution over weights. This approach is shown to improve over existing schemes for LeNet-5 on MNIST and VGG-16 on CIFA... | [
"This paper considers the compression of the model parameters in deep neural networks. The authors propose minimal random code learning (MIRACLE), which uses a random sample of weights and the variational framework interpreted by the bits-back argument. The authors introduce two theorems characterizing the properti... | [
[
4
],
[
1
],
[
2,
9
],
[
5
],
[
6
],
[
3,
7
],
[
11
],
[
12
],
[
10
],
[
8
],
[
0
],
[
13
]
] | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct"
] | [
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct"
] | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1
]
}
],
"category": [
"ORIG-MTH",
"QUAL... | benchmark/PDF/ICLR2019_r1f0YiCctm.pdf | openreview | benchmark/MD/ICLR2019_r1f0YiCctm.md | ICLR 2019 |
B1eSg3C9Ym | {
"title": "MEAN-FIELD ANALYSIS OF BATCH NORMALIZATION",
"abstract": "Batch Normalization (BatchNorm) is an extremely useful component of modern neural network architectures, enabling optimization using higher learning rates and achieving faster convergence. In this paper, we use mean-field theory to analytically q... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9
]
},
"scores... | [
[
{
"role": "AC",
"data": {
"metareview": "This paper presents a mean field analysis of the effect of batch norm on optimization. Assuming the weights and biases are independent Gaussians (an assumption that's led to other interesting analysis), they propagate various statistics through the ... | [
"In this paper, the effect of batch normalization to the maximum eigenvalue of the Fisher information is analyzed. The techinique is mostly developed by Karakida et al. (2018). The main result is an informal bound of the maximum eigenvalue, which is given without proof.",
"Though, the numerical result corresponds... | [
[
13
],
[
18
],
[
0
],
[
2
],
[
3
],
[
5
],
[
8
],
[
11,
14
],
[
12
],
[
21
],
[
15
],
[
1
],
[
9
],
[
23
],
[
24
],
[
4
],
[
10
],
[
17
],
[
19
],
[
6
],... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"cor... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"c... | [
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
0
]
},
{
"role": "Author",
"data": [
22,
24
]
}
],
"category": [
"CLAR-NOT"
]
},
{
"sentences": [
{
"role": "Reviewer 1... | benchmark/PDF/ICLR2019_B1eSg3C9Ym.pdf | openreview | benchmark/MD/ICLR2019_B1eSg3C9Ym.md | ICLR 2019 |
rkl4M3R5K7 | {
"title": "Optimal Attacks against Multiple Classifiers",
"abstract": "We study the problem of designing provably optimal adversarial noise algorithms that induce misclassification in settings where a learner aggregates decisions from multiple classifiers. Given the demonstrated vulnerability of state-of-the-art m... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4
]
},
"scores": {
"Solid": null,
"Presentation": null,
"No... | [
[
{
"role": "AC",
"data": {
"metareview": "Four reviewers have evaluated this paper. The reviewers have raised concerns about the specific formulation used for adversarial example generation which requires further clarity in motivation and interpretation. The reviewers have also made the poi... | [
"Summary:\nThe paper considers finding the most adversarial random noise given multiple classifiers. They formulate the problem as the standard min-max game and apply the multiplicative weight updates. The technical contribution is to clarify the computational complexity of implementing/approximating the response o... | [
[
7
],
[
18
],
[
19
],
[
20
],
[
8
],
[
3,
12,
17
],
[
9
],
[
1
],
[
5
],
[
10
],
[
14,
24
],
[
15
],
[
4
],
[
21,
25
],
[
16
],
[
0
],
[
2
],
[
6
],
[
... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"cor... | [
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
... | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1,
2,
3,
4
]
},
{
... | benchmark/PDF/ICLR2019_rkl4M3R5K7.pdf | openreview | benchmark/MD/ICLR2019_rkl4M3R5K7.md | ICLR 2019 |
ryxjH3R5KQ | {
"title": "Single Shot Neural Architecture Search Via Direct Sparse Optimization",
"abstract": "Recently Neural Architecture Search (NAS) has aroused great interest in both academia and industry, however it remains challenging because of its huge and non-continuous search space. Instead of applying evolutionary al... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
... | [
[
{
"role": "AC",
"data": {
"comment": "Latency is a practical measurement of performance. FLOPS is not.\n\nThe complicated learned architecture and many branches in Figure 4 make me concerned about the actual latency of the model and the practicality of deploying it, despite the low reporte... | [
"The authors present an architecture search method where connections are removed with sparse regularization. It produces good network blocks relatively quickly that perform well on CIFAR/ImageNet.",
"There are a few grammatical/spelling errors that need ironing out.",
"e.g. \"In specific\" --> \"Specifically\" ... | [
[
4
],
[
33
],
[
1,
14,
37,
38,
39
],
[
2
],
[
3
],
[
30
],
[
22
],
[
28
],
[
7
],
[
18
],
[
29
],
[
36,
41
],
[
12
],
[
5
],
[
9
],
[
17
],
[
19
],
[
20
... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"cor... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"c... | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1,
2
]
},
{
"role": "Author",
... | benchmark/PDF/ICLR2019_ryxjH3R5KQ.pdf | openreview | benchmark/MD/ICLR2019_ryxjH3R5KQ.md | ICLR 2019 |
S1xBioR5KX | {
"title": "Parameter efficient training of deep convolutional neural networks by dynamic sparse reparameterization",
"abstract": "Modern deep neural networks are highly overparameterized, and often of huge sizes. A number of post-training model compression techniques, such as distillation, pruning and quantizatio... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5
]
},
"scores": {
"Solid": null,
"Presentation": null... | [
[
{
"role": "AC",
"data": {
"metareview": "\nThe authors presents a technique for training neural networks, through dynamic sparse reparameterization. The work builds on previous work notably SET (Mocanu et al., 18), but the authors propose to use an adaptive threshold for and a heuristic fo... | [
"The paper provides a dynamic sparse reparameterization method allowing small networks to be trained at a comparable accuracy as pruned network with (initially) large parameter spaces. Improper initialization along with a fewer number of parameters requires a large parameter model, to begin with (Frankle and Carbin... | [
[
29
],
[
31
],
[
6,
17
],
[
14
],
[
20
],
[
2,
26
],
[
3,
8,
9,
23,
25
],
[
10,
22
],
[
12
],
[
16
],
[
27
],
[
30
],
[
4,
13,
21
],
[
11
],
[
15
],
[
7,
... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"cor... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"c... | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1
]
}
],
"category": [
"SIGN-BRD"
]
},
... | benchmark/PDF/ICLR2019_S1xBioR5KX.pdf | openreview | benchmark/MD/ICLR2019_S1xBioR5KX.md | ICLR 2019 |
Byxz4n09tQ | {
"title": "Model Compression with Generative Adversarial Networks",
"abstract": "More accurate machine learning models often demand more computation and memory at test time, making them difficult to deploy on CPU- or memory-constrained devices. Model compression (also known as distillation) alleviates this burden ... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9
]
},
"scores... | [
[
{
"role": "AC",
"data": {
"metareview": "The authors propose a scheme to compress models using student-teacher distillation, where training data are augmented using examples generated from a conditional GAN.\nThe reviewers were generally in agreement that 1) that the experimental results g... | [
"I like this paper. What the authors have done is of high quality. It is well written and clear.",
"However, quite a lot of experiments are necessary to make this paper publishable in my opinion.\nStrenghts:",
"- The idea to use a GAN for model compression is something that many must have considered. It is good... | [
[
11
],
[
13
],
[
22
],
[
41
],
[
42
],
[
2
],
[
33
],
[
10
],
[
5
],
[
26
],
[
1
],
[
4
],
[
7,
44
],
[
8
],
[
12
],
[
14
],
[
17,
18,
19,
45
],
[
20
],
[
... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"cor... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"c... | [
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
0
]
}
],
"category": [
"QUAL-MET",
"CLAR-WRT"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1
]
},
{
"role": "Aut... | benchmark/PDF/ICLR2019_Byxz4n09tQ.pdf | openreview | benchmark/MD/ICLR2019_Byxz4n09tQ.md | ICLR 2019 |
rkVOXhAqY7 | {
"title": "The Conditional Entropy Bottleneck",
"abstract": "We present a new family of objective functions, which we term the Conditional Entropy Bottleneck (CEB). These objectives are motivated by the Minimum Necessary Information (MNI) criterion. We demonstrate the application of CEB to classification tasks. We... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
... | [
[
{
"role": "AC",
"data": {
"metareview": "This paper proposes a criterion for representation learning, minimum necessary information, which states that for a task defined by some joint probability distribution P(X,Y) and the goal of (for example) predicting Y from X, a learned representatio... | [
"Update: see comments \"On revisions\" below.",
"This paper essentially introduces a label-dependent regularization to the VIB framework, matching the encoder distribution of one computed from labels. The authors show good performance in generalization, such that their approach is relatively robust in a number of... | [
[
10
],
[
38
],
[
2
],
[
3
],
[
20
],
[
22
],
[
24
],
[
30
],
[
31
],
[
34
],
[
36
],
[
17
],
[
18
],
[
19
],
[
0,
4,
16
],
[
1,
9,
13
],
[
5
],
[
6,
33,
... | [
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"c... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
... | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0,
1,
2
]
}
],
"category": [
"ORIG-MTH"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
3
]
}
],
"category": ... | benchmark/PDF/ICLR2019_rkVOXhAqY7.pdf | openreview | benchmark/MD/ICLR2019_rkVOXhAqY7.md | ICLR 2019 |
BygANhA9tQ | {
"title": "Cost-Sensitive Robustness against Adversarial Examples",
"abstract": "Several recent works have developed methods for training classifiers that are certifiably robust against norm-bounded adversarial perturbations. These methods assume that all the adversarial transformations are equally important, whic... | Accept (Poster) | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
... | [
[
{
"role": "AC",
"data": {
"metareview": "This paper studies the notion of certified cost-sensitive robustness against adversarial examples, by building from the recent [Wong & Koller'18]. Its main contribution is to adapt the robust classification objective to a 'cost-sensitive' objective,... | [
"** review score incremented following discussion below **",
"Strengths:\nWell written and clear paper",
"Intuition is strong: not all source-target class pairs are as beneficial to find adversarial examples for",
"Weaknesses:\nCost matrices choices feel a bit arbitrary in experiments",
"CIFAR experiments s... | [
[
14
],
[
7
],
[
1
],
[
6
],
[
10,
16,
18
],
[
12
],
[
13
],
[
22
],
[
2
],
[
20,
23,
26
],
[
27
],
[
3,
9
],
[
4,
11
],
[
8
],
[
21
],
[
24
],
[
30,
33
]... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"cor... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"c... | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1
]
}
],
"category": [
"CLAR-WRT"
]
},
... | benchmark/PDF/ICLR2019_BygANhA9tQ.pdf | openreview | benchmark/MD/ICLR2019_BygANhA9tQ.md | ICLR 2019 |
B1xVTjCqKQ | {
"title": "A Data-Driven and Distributed Approach to Sparse Signal Representation and Recovery",
"abstract": "In this paper, we focus on two challenges which offset the promise of sparse signal representation, sensing, and recovery. First, real-world signals can seldom be described as perfectly sparse vectors in a... | Accept (Poster) | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6
]
},
"scores": {
"Solid": null,
"Pres... | [
[
{
"role": "AC",
"data": {
"metareview": "This paper studies deep convolutional architectures to perform compressive sensing of natural images, demonstrating improved empirical performance with an efficient pipeline. \nReviewers reached a consensus that this is an interesting contribution t... | [
"Authors case the problem of finding informative measurements by using a maximum likelihood formulation and show how a data-driven dimensionality reduction protocol is built for sensing signals using convolutional architectures. A novel parallelization scheme is discussed and analyzed for speeding up the signal rec... | [
[
8
],
[
9,
16
],
[
10
],
[
26
],
[
2
],
[
5
],
[
7
],
[
11
],
[
14
],
[
6
],
[
19
],
[
18
],
[
1,
22
],
[
23
],
[
24
],
[
4
],
[
12
],
[
25
],
[
3
],
[
... | [
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"c... | [
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
... | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1,
2,
3
]
},
{
"role": "Au... | benchmark/PDF/ICLR2019_B1xVTjCqKQ.pdf | openreview | benchmark/MD/ICLR2019_B1xVTjCqKQ.md | ICLR 2019 |
rJVoEiCqKQ | {
"title": "Deep Perm-Set Net: Learn to predict sets with unknown permutation and cardinality using deep neural networks",
"abstract": "Many real-world problems, e.g. object detection, have outputs that are naturally expressed as sets of entities. This creates a challenge for traditional deep neural networks which ... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
... | [
[
{
"role": "AC",
"data": {
"metareview": "Strengths:\nThe method extends [21], which proposes an unordered set prediction model for multi-class classification.\nThe submission proposes a formulation to learn the distribution over unobservable permutation variables based on deep networks and... | [
"This paper looks to predict \"unstructured\" set output data. It extends Rezatofighi et al 2018 by modeling a latent permutation.",
"Unfortunately, there is a bit of an identity crisis happening in this paper. There are several choices that do not follow based on the data the paper considers.",
"1) The paper c... | [
[
6
],
[
9
],
[
14
],
[
15
],
[
16
],
[
18,
21,
34
],
[
4
],
[
5
],
[
8
],
[
17
],
[
20
],
[
38
],
[
29
],
[
22,
26,
33
],
[
25
],
[
28
],
[
24
],
[
39
],
[... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"cor... | [
"correct",
"incorrect",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
... | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0,
1
]
}
],
"category": [
"QUAL-MET"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
2
]
},
{
"role": "Unknown",
... | benchmark/PDF/ICLR2019_rJVoEiCqKQ.pdf | openreview | benchmark/MD/ICLR2019_rJVoEiCqKQ.md | ICLR 2019 |
S1M6Z2Cctm | {
"title": "Harmonic Unpaired Image-to-image Translation",
"abstract": "The recent direction of unpaired image-to-image translation is on one hand very exciting as it alleviates the big burden in obtaining label-intensive pixel-to-pixel supervision, but it is on the other hand not fully satisfactory due to the pres... | Accept (Poster) | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
... | [
[
{
"role": "AC",
"data": {
"metareview": "The proposed method introduces a method for unsupervised image-to-image mapping, using a new term into the objective function that enforces consistency in similarity between image patches across domains. Reviewers left constructive and detailed comm... | [
"Summary: The paper proposes a new smoothness constraint in the original cycle-gan formulation. The cycle-gan formulation minimizes reconstruction error on the input, and there is no criterion other than the adversarial loss function to ensure that it produce a good output (this is in sync with the observations fro... | [
[
12
],
[
16
],
[
17
],
[
24
],
[
23
],
[
26
],
[
31
],
[
8
],
[
11,
14,
15
],
[
19
],
[
21
],
[
2
],
[
3
],
[
10
],
[
20
],
[
22
],
[
28
],
[
29
],
[
1
],
... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"cor... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"incorrect",
"incorrect",
"correct",
"incorrect",
"correct",
"correct",
"inco... | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1
]
}
],
"category": [
"QUAL-MET"
]
},
... | benchmark/PDF/ICLR2019_S1M6Z2Cctm.pdf | openreview | benchmark/MD/ICLR2019_S1M6Z2Cctm.md | ICLR 2019 |
Sye7qoC5FQ | {
"title": "Adversarial Attacks on Node Embeddings",
"abstract": "The goal of network representation learning is to learn low-dimensional node embeddings that capture the graph structure and are useful for solving downstream tasks. However, despite the proliferation of such methods there is currently no study of th... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11
... | [
[
{
"role": "AC",
"data": {
"metareview": "The paper provides a novel analysis of the robustness to adversarial attacks in network representation learning. It appears to be a useful contribution for important class of models; however, the detailed reviews (1 and 2) raise some concerns that ... | [
"Brief Summary:\nThe authors present a novel adversarial attack on node embedding method based on random walks. They focus on perturbing the structure of the network. Because the bi-level optimization problem can be highly challenging, they refer to factorize a random walk matrix which is proved equivalent to DeepW... | [
[
1
],
[
2
],
[
16
],
[
3
],
[
6
],
[
7
],
[
5
],
[
8
],
[
9
],
[
10
],
[
15
],
[
17
],
[
4
],
[
13
],
[
14
],
[
12
],
[
0
],
[
11
]
] | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct"
] | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"incorrect",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct"
] | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1
]
}
],
"category": [
"CLAR-WRT"
]
},
... | benchmark/PDF/ICLR2019_Sye7qoC5FQ.pdf | openreview | benchmark/MD/ICLR2019_Sye7qoC5FQ.md | ICLR 2019 |
HkGTwjCctm | {
"title": "Pyramid Recurrent Neural Networks for Multi-Scale Change-Point Detection",
"abstract": "Many real-world time series, such as in activity recognition, finance, or climate science, have changepoints where the system's structure or parameters change. Detecting changes is important as they may indicate crit... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9
]
},
"scores... | [
[
{
"role": "AC",
"data": {
"metareview": "This paper studies change-point detection in time series using a multiscale neural network architecture which contains recurrent connections across different time scales. \n\nReviewers were mixed in this submission. They found the paper generally cl... | [
"1. This papers leverages the concept of wavelet transform within a deep architecture to solve the classic problem (especially for wavelet analysis) of change point detection. The authors do a reasonably comprehensive job of demonstrating the efficacy of the proposed framework using various synthetic and real data ... | [
[
4
],
[
8
],
[
15
],
[
7
],
[
1
],
[
2
],
[
3
],
[
13
],
[
14
],
[
18
],
[
19
],
[
0
],
[
9
],
[
10
],
[
11
],
[
17
],
[
16
],
[
5
],
[
12
],
[
6
]
] | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct"
] | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct"
] | [
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
0
]
}
],
"category": [
"QUAL-EXP"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1
]
},
{
"role": "Author",
"dat... | benchmark/PDF/ICLR2019_HkGTwjCctm.pdf | openreview | benchmark/MD/ICLR2019_HkGTwjCctm.md | ICLR 2019 |
r1xwKoR9Y7 | {
"title": "GamePad: A Learning Environment for Theorem Proving",
"abstract": "In this paper, we introduce a system called GamePad that can be used to explore the application of machine learning methods to theorem proving in the Coq proof assistant. Interactive theorem provers such as Coq enable users to construct ... | Accept (Poster) | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10
]
},... | [
[
{
"role": "AC",
"data": {
"metareview": "This paper provides an RL environment defined over Coq, allowing for RL agents and other such systems to to be trained to propose tactics during the running of an ITP. I really like this general line of work, and the reviewers broadly speaking did a... | [
"Summary: This paper mixes automated theorem proving with machine learning models. The final goal, of course, is to be able to train a model that works in conjunction with an automated theorem proving system to efficiently prove theorems, and, ideally, in a way that resembles the way humans prove theorems. This is ... | [
[
12
],
[
13
],
[
2
],
[
4
],
[
6
],
[
15
],
[
11
],
[
9
],
[
16
],
[
5
],
[
7
],
[
14
],
[
1,
3
],
[
0
],
[
8
],
[
10
]
] | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct"
] | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct"
] | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0,
1,
2
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
3
]
}
],
"category": [
... | benchmark/PDF/ICLR2019_r1xwKoR9Y7.pdf | openreview | benchmark/MD/ICLR2019_r1xwKoR9Y7.md | ICLR 2019 |
B1VWtsA5tQ | {
"title": "PPO-CMA: Proximal Policy Optimization with Covariance Matrix Adaptation",
"abstract": "Proximal Policy Optimization (PPO) is a highly popular model-free reinforcement learning (RL) approach. However, in continuous state and actions spaces and a Gaussian policy -- common in computer animation and robotic... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2
]
},
"scores": {
"Solid": null,
"Presentation": null,
"Novelty": null,
"Overa... | [
[
{
"role": "AC",
"data": {
"metareview": "This paper proposes to improve the exploration in the PPO algorithm by applying CMA-ES. Major concerns of the paper include: paper editing can be improved; the choices of baselines used in the paper may be not reasonable; flaws in comparisons with S... | [
"I have to say that this paper is not well organized. It describes the advantage function and CMA-ES, but it does not describe PPO and PPO-CMA very well. I goes through the paper twice, but I couldn't really get how the policy variance is adapted.",
"Though the title of section 4 is \"PPO-CMA\", only the first pa... | [
[
0
],
[
13,
20
],
[
1
],
[
9,
18
],
[
4
],
[
7
],
[
11
],
[
16
],
[
17
],
[
10
],
[
12
],
[
19
],
[
15
],
[
2
],
[
3
],
[
5
],
[
6
],
[
8
],
[
14
]
] | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct"
] | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct"
] | [
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
0,
1
]
},
{
"role": "Author",
"data": [
7,
8,
73,
75
]
}
],
"category": [
"CLAR-WRT"
]
},
{
"sentences"... | benchmark/PDF/ICLR2019_B1VWtsA5tQ.pdf | openreview | benchmark/MD/ICLR2019_B1VWtsA5tQ.md | ICLR 2019 |
ByxmXnA9FQ | {
"title": "A Variational Dirichlet Framework for Out-of-Distribution Detection",
"abstract": "With the recently rapid development in deep learning, deep neural networks have been widely adopted in many real-life applications. However, deep neural networks are also known to have very little control over its uncerta... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7
]
},
"scores": {
"Solid": null,
... | [
[
{
"role": "AC",
"data": {
"metareview": "The paper proposes a new framework for out-of-distribution detection, based on variational inference and a prior Dirichlet distribution.\n\nThe reviewers and AC note the following potential weaknesses: (1) arguable and not well justified choices of ... | [
"This paper provides a new method that approximates the confidence distribution of classification probability, which is useful for novelty detection. The variational inference with Dirichlet family is a natural choice.",
"Though it is principally insightful to introduce the “higher-order” uncertainty, I do see th... | [
[
10
],
[
27
],
[
4,
5
],
[
2
],
[
7
],
[
8
],
[
9
],
[
16
],
[
22
],
[
25
],
[
1
],
[
12
],
[
18
],
[
18
],
[
17
],
[
19
],
[
21
],
[
3
],
[
6
],
[
13,
... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"cor... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"c... | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1
]
},
{
"role": "Author",
"data": [
... | benchmark/PDF/ICLR2019_ByxmXnA9FQ.pdf | openreview | benchmark/MD/ICLR2019_ByxmXnA9FQ.md | ICLR 2019 |
S1EERs09YQ | {
"title": "Discovery of Natural Language Concepts in Individual Units of CNNs",
"abstract": "Although deep convolutional networks have achieved improved performance in many natural language tasks, they have been treated as black boxes because they are difficult to interpret. Especially, little is known about how t... | Accept (Poster) | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3
]
},
"scores": {
"Solid": null,
"Presentation": null,
"Novelty": null,
... | [
[
{
"role": "AC",
"data": {
"metareview": "Important problem (making NN more transparent); reasonable approach for identifying which linguistic concepts different neurons are sensitive to; rigorous experiments. Paper was reviewed by three experts. Initially there were some concerns but after... | [
"The paper is well written and structured, presenting the problem clearly and accurately. It contains considerable relevant references and enough background knowledge. It nicely motivates the proposed approach, locates the contributions in the state-of-the-art and reviews related work. It is also very honest in ter... | [
[
28,
47
],
[
19
],
[
26
],
[
0
],
[
5,
15
],
[
9,
24
],
[
10
],
[
11
],
[
14
],
[
18
],
[
27
],
[
29
],
[
30
],
[
31
],
[
32
],
[
33
],
[
34
],
[
35
],
[
3... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"cor... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"cor... | [
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
0
]
}
],
"category": [
"CLAR-WRT",
"QUAL-CMP"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1
]
}
],
"category": [
... | benchmark/PDF/ICLR2019_S1EERs09YQ.pdf | openreview | benchmark/MD/ICLR2019_S1EERs09YQ.md | ICLR 2019 |
S1GcHsAqtm | {
"title": "Adaptive Pruning of Neural Language Models for Mobile Devices",
"abstract": "Neural language models (NLMs) exist in an accuracy-efficiency tradeoff space where better perplexity typically comes at the cost of greater computation complexity. In a software keyboard application on mobile devices, this tran... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
... | [
[
{
"role": "AC",
"data": {
"metareview": "The area chair agrees with the authors and the reviewers that the topic of this work is relevant and important. The area chair however shares the concerns of the reviewers about the setup and the empirical evaluation:\n- Having one model that can be... | [
"In this paper, the authors investigate the accuracy-efficiency tradeoff for neural language models.",
"In particular, they explore how different compression strategies impact the accuracy (and flops), and more interestingly, also how it impacts the power use for a RaspberryPi. The authors consider the QRNNs and ... | [
[
1
],
[
11
],
[
15
],
[
3
],
[
4
],
[
19
],
[
20
],
[
5
],
[
6
],
[
7
],
[
8
],
[
13
],
[
14
],
[
16
],
[
17
],
[
18
],
[
9
],
[
12
],
[
0
],
[
2
],
[
... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct"
] | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct"
] | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1,
2
]
}
],
"category": [
"ORIG-MTH"... | benchmark/PDF/ICLR2019_S1GcHsAqtm.pdf | openreview | benchmark/MD/ICLR2019_S1GcHsAqtm.md | ICLR 2019 |
Sklv5iRqYX | {
"title": "Multi-Domain Adversarial Learning",
"abstract": "Multi-domain learning (MDL) aims at obtaining a model with minimal average risk across multiple domains. Our empirical motivation is automated microscopy data, where cultured cells are imaged after being exposed to known and unknown chemical perturbations... | Accept (Poster) | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
... | [
[
{
"role": "AC",
"data": {
"metareview": "This paper extends the single source H-divergence theory for domain adaptation to the case of multiple domains. Thus, drawing on the known connection between H-divergence and learning the domain classifier for adversarial adaptation, the authors pro... | [
"Summary:\nThe manuscript proposes a multi-domain adversarial learning (MDL) method called MULANN, to leverage multiple datasets with overlapping but distinct class sets, in a semi-supervised setting. The authors define a new discrimination task to discriminate, within each domain, labeled samples from unlabeled on... | [
[
3
],
[
15
],
[
18
],
[
2
],
[
17
],
[
1,
16
],
[
7
],
[
11
],
[
19
],
[
12
],
[
13
],
[
4
],
[
5,
6
],
[
8
],
[
20
],
[
9
],
[
10
],
[
0
],
[
14
]
] | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct"
] | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct"
] | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1
]
}
],
"category": [
"ORIG-MTH"
]
},
... | benchmark/PDF/ICLR2019_Sklv5iRqYX.pdf | openreview | benchmark/MD/ICLR2019_Sklv5iRqYX.md | ICLR 2019 |
HJxB5sRcFQ | {
"title": "LayoutGAN: Generating Graphic Layouts with Wireframe Discriminators",
"abstract": "Layout is important for graphic design and scene generation. We propose a novel Generative Adversarial Network, called LayoutGAN, that synthesizes layouts by modeling geometric relations of different types of 2D elements.... | Accept (Poster) | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3
]
},
"scores": {
"Solid": null,
"Presentation": null,
"Novelty": null,
... | [
[
{
"role": "AC",
"data": {
"metareview": "Reviewers agree the paper should be accepted.\nSee reviews below."
}
}
]
] | [
"The authors present a GAN based framework for Graphic Layouts.",
"Instead of considering a graphic layout as a collection of pixels, they treat it as a collection of primitive objects like polygons. The objective is to create an alignment of these objects that mimics some real data distribution.",
"The novelty... | [
[
16
],
[
21
],
[
14
],
[
15
],
[
24
],
[
0
],
[
1,
6,
25
],
[
23
],
[
9
],
[
10
],
[
11,
20
],
[
17
],
[
2
],
[
3
],
[
7
],
[
8
],
[
19
],
[
27
],
[
12
]... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"cor... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"cor... | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0,
1
]
}
],
"category": [
"ORIG-MTH"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
2
]
}
],
"category": [
"ORIG... | benchmark/PDF/ICLR2019_HJxB5sRcFQ.pdf | openreview | benchmark/MD/ICLR2019_HJxB5sRcFQ.md | ICLR 2019 |
SyzjBiR9t7 | {
"title": "MANIFOLDNET: A DEEP NEURAL NETWORK FOR MANIFOLD-VALUED DATA",
"abstract": "Developing deep neural networks (DNNs) for manifold-valued data sets\nhas gained much interest of late in the deep learning research\ncommunity. Examples of manifold-valued data include data from\nomnidirectional cameras on auto... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
... | [
[
{
"role": "AC",
"data": {
"metareview": "This manuscript proposes an extension of convolution operations for manifold-valued data. The primary contributions include the development and description of the approach and implementation and evaluation on real data.\n\nThe reviewers and AC expre... | [
"Brief summary:\nThe paper considers a generalization of convolutional neural networks (CNNs) to data residing on Riemannian manifolds. The idea is to replace convolutions with weighted averages, which are implemented intrinsically on the manifold. It is shown that this operator is equivariant to isometric group ac... | [
[
9
],
[
33
],
[
7
],
[
21
],
[
27
],
[
0,
8
],
[
2
],
[
3
],
[
13
],
[
16
],
[
19
],
[
25,
28
],
[
34
],
[
30
],
[
38
],
[
37
],
[
36
],
[
1
],
[
6
],
[
... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"cor... | [
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"incorrect",
"incorrect",
"correct",
"incorrect",
"incorrect",
"correct",
"incorrect",
"correct",
"... | [
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
0
]
},
{
"role": "Author",
"data": [
32,
33,
89,
95
]
}
],
"category": [
"CLAR-WRT",
"QUAL-EXP"
]
},
{
"sen... | benchmark/PDF/ICLR2019_SyzjBiR9t7.pdf | openreview | benchmark/MD/ICLR2019_SyzjBiR9t7.md | ICLR 2019 |
S1E3Ko09F7 | {
"title": "L-Shapley and C-Shapley: Efficient Model Interpretation for Structured Data",
"abstract": "Instancewise feature scoring is a method for model interpretation, which yields, for each test instance, a vector of importance scores associated with features. Methods based on the Shapley score have been propose... | Accept (Poster) | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10
]
},... | [
[
{
"role": "AC",
"data": {
"metareview": "The paper presents two new methods for model-agnostic interpretation of instance-wise feature importance. \n\nPros:\nUnlike previous approaches based on the Shapley value, which had an exponential complexity in the number of features, the proposed m... | [
"The paper proposes two approximations to the Shapley value used for generating feature scores for interpretability. Both exploit a graph structure over the features by considering only subsets of neighborhoods of features (rather than all subsets). The authors give some approximation guarantees under certain Marko... | [
[
2
],
[
5
],
[
1
],
[
3
],
[
14
],
[
16
],
[
11
],
[
4
],
[
6
],
[
7
],
[
8
],
[
9
],
[
10
],
[
12
],
[
17
],
[
18
],
[
15
],
[
0
],
[
13
]
] | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct"
] | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct"
] | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1
]
},
{
"role": "Author",
"data": [
... | benchmark/PDF/ICLR2019_S1E3Ko09F7.pdf | openreview | benchmark/MD/ICLR2019_S1E3Ko09F7.md | ICLR 2019 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.