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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
SJx1URNKwH | {
"title": "MetaPix: Few-Shot Video Retargeting",
"authors": [
"Jessica Lee",
"Deva Ramanan",
"Rohit Girdhar"
],
"authorids": [
"jl5@cs.cmu.edu",
"deva@cs.cmu.edu",
"rgirdhar@cs.cmu.edu"
],
"keywords": [
"Meta-learning",
"Few-shot Learning",
"Generative Adversarial Networ... | Accept (Poster) | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2
]
},
"scores": {
"Solid": null,
"Presentation": null,
"Novelty": null,
"Overa... | [
[
{
"role": "PC",
"data": {
"comment": "Three reviewers have assessed this paper and they have scored it 6/6/6 after rebuttal. Nonetheless, the reviewers have raised a number of criticisms and the authors are encouraged to resolve them for the camera-ready submission."
}
}
]
] | [
"In this paper, authors propose to address few shot video retargeting, where one should adapt a generic generative model of human actions to a specific person given a few samples of their appearance.",
"Overall, the paper is written with a good structure. I do like the problem setting and motivations in this pape... | [
[
10
],
[
1
],
[
2,
4,
8
],
[
7
],
[
6
],
[
5
],
[
9
],
[
0
],
[
3
]
] | [
"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": [
"CLAR-WRT"
]
},
... | benchmark/PDF/ICLR2020_SJx1URNKwH.pdf | openreview | benchmark/MD/ICLR2020_SJx1URNKwH.md | ICLR 2020 |
HylxE1HKwS | {
"authorids": [
"hancai@mit.edu",
"ganchuang1990@gmail.com",
"usedtobe@mit.edu",
"zhangzk@mit.edu",
"songhan@mit.edu"
],
"title": "Once-for-All: Train One Network and Specialize it for Efficient Deployment",
"authors": [
"Han Cai",
"Chuang Gan",
"Tianzhe Wang",
"Zhekai Zhang... | 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": "PC",
"data": {
"comment": "The authors propose a new method for neural architecture search, except it's not exactly that because model training is separated from architecture, which is the main point of the paper. Once this network is trained, sub-networks can be distilled from ... | [
"In this papers, the authors learn a Once-for-all net. This starts as a big neural network which is trained normally (albeit with input images of different resolutions). It is then fine-tuned while sampling sub-networks with progressively smaller kernels, then lower depth, then width (while still sampling larger ne... | [
[
3
],
[
8
],
[
4
],
[
5
],
[
7
],
[
16
],
[
23
],
[
9,
12
],
[
19
],
[
22
],
[
17
],
[
18
],
[
1
],
[
2
],
[
11
],
[
13
],
[
20
],
[
6
],
[
0
],
[
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",
"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
]
},
{
"role": "Au... | benchmark/PDF/ICLR2020_HylxE1HKwS.pdf | openreview | benchmark/MD/ICLR2020_HylxE1HKwS.md | ICLR 2020 |
S1g2skStPB | {
"title": "Causal Discovery with Reinforcement Learning",
"authors": [
"Shengyu Zhu",
"Ignavier Ng",
"Zhitang Chen"
],
"authorids": [
"zhushengyu@huawei.com",
"ignavierng@cs.toronto.edu",
"chenzhitang2@huawei.com"
],
"keywords": [
"causal discovery",
"structure learning",
... | Accept (Talk) | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7
]
},
"scores": {
"Solid": null,
... | [
[
{
"role": "PC",
"data": {
"comment": "This paper proposes an RL-based structure search method for causal discovery. The reviewers and AC think that the idea of applying reinforcement learning to causal structure discovery is novel and intriguing. While there were initially some concerns re... | [
"This work addresses the task of causal discovery. The proposed contribution is to apply prior work which uses reinforcement learning for combinatorial optimization to structure learning.",
"Specifically, the proposed optimization problem seeks to maximize a penalized score criterion subject to the acyclicity con... | [
[
24
],
[
22
],
[
23
],
[
16
],
[
18
],
[
19
],
[
1
],
[
5
],
[
7
],
[
20
],
[
2
],
[
3
],
[
14
],
[
15
],
[
17
],
[
21
],
[
0
],
[
4
],
[
6
],
[
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",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
... | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0,
1,
2
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
3,
4
]
},
{
... | benchmark/PDF/ICLR2020_S1g2skStPB.pdf | openreview | benchmark/MD/ICLR2020_S1g2skStPB.md | ICLR 2020 |
SygWvAVFPr | {
"title": "Neural Module Networks for Reasoning over Text",
"authors": [
"Nitish Gupta",
"Kevin Lin",
"Dan Roth",
"Sameer Singh",
"Matt Gardner"
],
"authorids": [
"gnnitish@gmail.com",
"kevinlin@eecs.berkeley.edu",
"danroth@seas.upenn.edu",
"sameer@uci.edu",
"mattg@allen... | Accept (Poster) | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3
]
},
"scores": {
"Solid": null,
"Presentation": null,
"Novelty": null,
... | [
[
{
"role": "PC",
"data": {
"comment": "This work extends the previously introduced NMN for VQA for handling reasoning over text using symbolic reasoning components that can perform counting, sorting etc and can be compositionally combined. Moreover, to successfully train the model, the auth... | [
"This paper proposes a model and a training framework for question answering which requires compositional reasoning over the input text, by building executable neural modules and training based on additional auxiliary supervision signals.",
"I really like this paper and the approach taken: tackling complex QA ta... | [
[
5
],
[
23
],
[
2
],
[
15,
17,
25
],
[
19
],
[
20
],
[
22
],
[
26
],
[
7,
13
],
[
8
],
[
9
],
[
18
],
[
21
],
[
1
],
[
4
],
[
6,
12
],
[
10
],
[
0
],
[
... | [
"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
]
}
],
"category": [
"SIGN-BRD"
]
},
... | benchmark/PDF/ICLR2020_SygWvAVFPr.pdf | openreview | benchmark/MD/ICLR2020_SygWvAVFPr.md | ICLR 2020 |
SJeLIgBKPS | {
"authorids": [
"vfleaking@gmail.com",
"lijian83@mail.tsinghua.edu.cn"
],
"title": "Gradient Descent Maximizes the Margin of Homogeneous Neural Networks",
"authors": [
"Kaifeng Lyu",
"Jian Li"
],
"pdf": "/pdf/1961d0a01f9b41951a88793dcc0e818a80d108fe.pdf",
"TL;DR": "We study the implicit b... | Accept (Talk) | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9
]
},
"scores... | [
[
{
"role": "PC",
"data": {
"comment": "This paper studies the implicit regularization of the gradient descent in homogeneous and shows that when the training loss falls below a threshold, then the smoothed. This study generalizes some of the earlier related works by relying on weaker assump... | [
"This paper studies the implicit regularization phenomenon.",
"More precisely, given separable data the authors ask whether homogenous functions (including neural networks) trained by gradient flow/descent converge to the max-margin solution. The authors show that the limit points of gradient descent are KKT poi... | [
[
4
],
[
17
],
[
1,
3
],
[
10
],
[
14
],
[
15
],
[
16
],
[
2,
6
],
[
9
],
[
11
],
[
12
],
[
13
],
[
0
],
[
5
],
[
7
],
[
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",
"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
]
}
],
"category": [
"ORIG-ANL"... | benchmark/PDF/ICLR2020_SJeLIgBKPS.pdf | openreview | benchmark/MD/ICLR2020_SJeLIgBKPS.md | ICLR 2020 |
S1lJv0VYDr | {
"authorids": [
"kriswu8021@gmail.com",
"tinghanf@princeton.edu",
"ramadge@princeton.edu",
"haosu@eng.ucsd.edu"
],
"title": "Model Imitation for Model-Based Reinforcement Learning",
"authors": [
"Yueh-Hua Wu",
"Ting-Han Fan",
"Peter J. Ramadge",
"Hao Su"
],
"pdf": "/pdf/23c4... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
... | [
[
{
"role": "PC",
"data": {
"comment": "This paper addresses challenges in offline model learning, i.e., in the setting where some trajectories are given and can be used for learning a model, which in turn serves to train an RL agent or plan action sequences in simulation. A key issue in thi... | [
"Review for \"Model Imitation for Model-Based Reinforcement Learning\".",
"The paper proposes a type of model-based RL that relies on matching the distribution of (s,a,s') tuples rather than using supervised learning to learn an autoregressive model using supervised learning.",
"I vote to reject the paper for... | [
[
5
],
[
6
],
[
17
],
[
19
],
[
9
],
[
18
],
[
20
],
[
21
],
[
23
],
[
4
],
[
2
],
[
25
],
[
13,
15
],
[
22
],
[
26
],
[
10
],
[
1
],
[
11,
16
],
[
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",
"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,
1
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer",
"data": [
2
]
}
],
"category": [
"QUAL-MET"
... | benchmark/PDF/ICLR2020_S1lJv0VYDr.pdf | openreview | benchmark/MD/ICLR2020_S1lJv0VYDr.md | ICLR 2020 |
ryg48p4tPH | {
"title": "Action Semantics Network: Considering the Effects of Actions in Multiagent Systems",
"authors": [
"Weixun Wang",
"Tianpei Yang",
"Yong Liu",
"Jianye Hao",
"Xiaotian Hao",
"Yujing Hu",
"Yingfeng Chen",
"Changjie Fan",
"Yang Gao"
],
"authorids": [
"wxwang@tju.ed... | 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": "PC",
"data": {
"comment": "The authors address the challenge of sample-efficient learning in multi-agent systems. They propose a model that distinguishes actions in terms of their semantics, specifically in terms of whether they influence the acting agent and environment or whet... | [
"This paper proposes a neural network architecture that provides an agent-agent based embeddings that are used for actions that directly affect specific agent. Proposed architectural choice exploits (implicitly assumed) independence of some actions wrt. observations of agents that it is not directly affecting. Auth... | [
[
5
],
[
6
],
[
7
],
[
2
],
[
8
],
[
9
],
[
13
],
[
4
],
[
16
],
[
1
],
[
17
],
[
3
],
[
12
],
[
18
],
[
15
],
[
0
],
[
10
],
[
11
],
[
14
]
] | [
"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",
"incorrect",
"incorrect",
"correct"
] | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1
]
}
],
"category": [
"QUAL-EXP"
]
},
... | benchmark/PDF/ICLR2020_ryg48p4tPH.pdf | openreview | benchmark/MD/ICLR2020_ryg48p4tPH.md | ICLR 2020 |
rJeB36NKvB | {
"title": "How much Position Information Do Convolutional Neural Networks Encode?",
"authors": [
"Md Amirul Islam*",
"Sen Jia*",
"Neil D. B. Bruce"
],
"authorids": [
"amirul@scs.ryerson.ca",
"sen.jia@ryerson.ca",
"bruce@ryerson.ca"
],
"keywords": [
"network understanding",
"... | Accept (Spotlight) | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8
]
},
"scores": {
... | [
[
{
"role": "PC",
"data": {
"comment": "This paper analyzes the weights associated with filters in CNNs and finds that they encode positional information (i.e. near the edges of the image). A detailed discussion and analysis is performed, which shows where this positional information comes ... | [
"The paper investigates to what degree Convolutional Neural Networks (CNNs) learn to encode positional information.",
"Rather interesting finding is the not only they do encode this information, but that it is to a large degree function of the padding commonly used in the CNN architectures.",
"The problem the p... | [
[
9
],
[
5
],
[
1,
16
],
[
4
],
[
3
],
[
2,
13
],
[
7
],
[
10
],
[
15
],
[
6
],
[
11
],
[
17
],
[
14
],
[
0
],
[
8
],
[
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",
"incorrect",
"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": [
"ORIG-ANL"
]
},
... | benchmark/PDF/ICLR2020_rJeB36NKvB.pdf | openreview | benchmark/MD/ICLR2020_rJeB36NKvB.md | ICLR 2020 |
rkeu30EtvS | {
"authorids": [
"yechengxi@gmail.com",
"mevanusa@umd.edu",
"huah@umd.edu",
"amitrokh@umd.edu",
"tomg@cs.umd.edu",
"yorke@umd.edu",
"fer@umiacs.umd.edu",
"yiannis@cs.umd.edu"
],
"title": "Network Deconvolution",
"authors": [
"Chengxi Ye",
"Matthew Evanusa",
"Hua He",
... | Accept (Spotlight) | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6
]
},
"scores": {
"Solid": null,
"Pres... | [
[
{
"role": "PC",
"data": {
"comment": "This paper presents a feature normalization method for CNNs by decorrelating channel-wise and spatial correlation simultaneously. Overall all reviewers are positive to the acceptance and I support their opinions. The idea and implementation is relative... | [
"This paper proposes \"network deconvolution\", a neural network primitive aimed at whitening the activations of each layer of the network. The method is a generalization of batch normalization that not only whitens per channel, but also removes correlations between channels and across spatial locations. Experiment... | [
[
17
],
[
15
],
[
18
],
[
1
],
[
16
],
[
3
],
[
5
],
[
6
],
[
11
],
[
14
],
[
2
],
[
4
],
[
8
],
[
12
],
[
9
],
[
0
],
[
7
],
[
10
],
[
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",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct"
] | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1
]
}
],
"category": [
"ORIG-MTH",
"QUAL... | benchmark/PDF/ICLR2020_rkeu30EtvS.pdf | openreview | benchmark/MD/ICLR2020_rkeu30EtvS.md | ICLR 2020 |
SJeQEp4YDH | {
"authorids": [
"xy4cm@virginia.edu",
"skolouri@hrl.com",
"gustavo@virginia.edu"
],
"title": "GAT: Generative Adversarial Training for Adversarial Example Detection and Robust Classification",
"authors": [
"Xuwang Yin",
"Soheil Kolouri",
"Gustavo K Rohde"
],
"pdf": "/pdf/d5c437c5244... | Accept (Poster) | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3
]
},
"scores": {
"Solid": null,
"Presentation": null,
"Novelty": null,
... | [
[
{
"role": "PC",
"data": {
"comment": "This work addresses the problem of detecting an adversarial attack. This is a challenging problem as the detection mechanism itself is also vulnerable to attack. The paper proposes asymmetrical adversarial training as a robust solution. This approach p... | [
"Summary:\nThis paper studies the adversarial detection problem within the robust optimization framework. They propose an adversarial detection and a generative modeling technique called asymmetrical adversarial training (AAT). With one detector for each class discriminating natural data from adversarially perturbe... | [
[
11
],
[
1
],
[
4
],
[
6
],
[
7
],
[
9
],
[
10
],
[
3
],
[
0
],
[
2
],
[
5
],
[
8
]
] | [
"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,
2
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
3
]
},
{
"role": "Au... | benchmark/PDF/ICLR2020_SJeQEp4YDH.pdf | openreview | benchmark/MD/ICLR2020_SJeQEp4YDH.md | ICLR 2020 |
HyxyIgHFvr | {
"authorids": [
"goldblumcello@gmail.com",
"jonas.geiping@uni-siegen.de",
"avi1@umd.edu",
"michael.moeller@uni-siegen.de",
"tomg@cs.umd.edu"
],
"title": "Truth or backpropaganda? An empirical investigation of deep learning theory",
"authors": [
"Micah Goldblum",
"Jonas Geiping",
... | Accept (Spotlight) | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5
]
},
"scores": {
"Solid": null,
"Presentation": null... | [
[
{
"role": "AC",
"data": {
"comment": "The work Bayesian Neural Network Ensembles by Pearce et al (https://arxiv.org/abs/1811.12188) proposes to use an ensemble of neural networks that are each trained with L2 regularization from a normal distribution sample. They show this is form of appro... | [
"The authors seek to challenge some presumptions about training deep neural networks, such as the robustness of low rank linear layers and the existence of suboptimal local minima. They provide analytical insight as well as a few experiments.",
"I give this paper an accept. They analytically explore four relevant... | [
[
12
],
[
2
],
[
5,
11
],
[
8
],
[
10
],
[
4
],
[
9
],
[
17
],
[
14
],
[
3
],
[
13
],
[
7
],
[
16
],
[
0
],
[
1
],
[
6
],
[
15
],
[
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"
] | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer",
"data": [
1
]
}
],
"category": [
"N/A"
]
},
{
... | benchmark/PDF/ICLR2020_HyxyIgHFvr.pdf | openreview | benchmark/MD/ICLR2020_HyxyIgHFvr.md | ICLR 2020 |
r1lZ7AEKvB | {
"authorids": [
"pbarcelo@gmail.com",
"egor.kostylev@cs.ox.ac.uk",
"mikael.monet@imfd.cl",
"jorge.perez.rojas@gmail.com",
"juan.reutter@gmail.com",
"jpsilvapena@gmail.com"
],
"title": "The Logical Expressiveness of Graph Neural Networks",
"authors": [
"Pablo Barceló",
"Egor V. K... | Accept (Spotlight) | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6
]
},
"scores": {
"Solid": null,
"Pres... | [
[
{
"role": "PC",
"data": {
"comment": "The paper focuses on characterizing the expressiveness of graph neural networks. The reviewers were satisfied that the authors answered their questions suffciiently and uniformly agree that this is a strong paper that should be accepted."
}
}... | [
"The paper utilizes recent insights into the relationship between the Weisfeiler-Lehman (WL) test for checking graph isomorphism and Graph Neural Networks (GNNs) in order to characterize the class of node classifiers that can be captured by a specific GNN architecture, called aggregate-combine GNN (AC-GNN).",
"Th... | [
[
6
],
[
8
],
[
0,
5,
7
],
[
1,
3
],
[
9
],
[
2
],
[
4
],
[
10
]
] | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct"
] | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct"
] | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0,
1
]
}
],
"category": [
"ORIG-ANL"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
2
]
},
{
"role": "Author",
... | benchmark/PDF/ICLR2020_r1lZ7AEKvB.pdf | openreview | benchmark/MD/ICLR2020_r1lZ7AEKvB.md | ICLR 2020 |
rkgg6xBYDH | {
"authorids": [
"zhtu3055@uni.sydney.edu.au",
"fengxiang.he@sydney.edu.au",
"dacheng.tao@sydney.edu.au"
],
"title": "Understanding Generalization in Recurrent Neural Networks",
"authors": [
"Zhuozhuo Tu",
"Fengxiang He",
"Dacheng Tao"
],
"pdf": "/pdf/9d02fe4b684d11816e76d578c498b29a... | Accept (Poster) | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10
]
},... | [
[
{
"role": "PC",
"data": {
"comment": "This paper presents a generalization bound for RNNs based on matrix-1 norm and Fisher-Rao norm. As the initial bound relies on non-signularity of input covariance, which may not always hold in practice, the authors present additional analysis by noise ... | [
"This paper proposes a new generalization bound for vanilla RNN with ReLU activation in terms of matrix-1 norm and Fisher-Rao norm. This bound has no explicit dependence on the size of networks.",
"I am actually not familiar with the generalization theorem on RNN.",
"Nevertheless, according to the demonstration... | [
[
26
],
[
27
],
[
5
],
[
9
],
[
2
],
[
15
],
[
22
],
[
24
],
[
28
],
[
6
],
[
11
],
[
13
],
[
17,
19
],
[
20
],
[
4
],
[
8
],
[
0
],
[
1
],
[
3
],
[
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",
"correct",
"correct",
"correct",
"correct",
"cor... | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer",
"data": [
1
]
}
],
"category": [
"N/A"
]
},
{
... | benchmark/PDF/ICLR2020_rkgg6xBYDH.pdf | openreview | benchmark/MD/ICLR2020_rkgg6xBYDH.md | ICLR 2020 |
SJxUjlBtwB | {
"authorids": [
"zhonge@mit.edu",
"tbepler@mit.edu",
"jhdavis@mit.edu",
"bab@mit.edu"
],
"title": "Reconstructing continuous distributions of 3D protein structure from cryo-EM images",
"authors": [
"Ellen D. Zhong",
"Tristan Bepler",
"Joseph H. Davis",
"Bonnie Berger"
],
"TL... | Accept (Spotlight) | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5
]
},
"scores": {
"Solid": null,
"Presentation": null... | [
[
{
"role": "PC",
"data": {
"comment": "The paper introduces a generative approach to reconstruct 3D images for cryo-electron microscopy (cryo-EM).\n\nAll reviewers really liked the paper, appreciate the challenging problem tackled and the proposed solution.\n\nAcceptance is therefore recomm... | [
"The authors introduce cryoDRGN, a VAE neural network architecture to reconstruct 3D protein structure from 2D cryo-EM images.",
"The paper offers for a good read and diagrams are informative.",
"Below are comments for improvement and clarification.",
"> Consider explaining cryoSPARC in detail given that is t... | [
[
2
],
[
8
],
[
5
],
[
1
],
[
6
],
[
7
],
[
3
],
[
15
],
[
16
],
[
14
],
[
9
],
[
10
],
[
17
],
[
0
],
[
4
],
[
11
],
[
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",
"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/ICLR2020_SJxUjlBtwB.pdf | openreview | benchmark/MD/ICLR2020_SJxUjlBtwB.md | ICLR 2020 |
r1eowANFvr | {
"title": "Towards Fast Adaptation of Neural Architectures with Meta Learning",
"authors": [
"Dongze Lian",
"Yin Zheng",
"Yintao Xu",
"Yanxiong Lu",
"Leyu Lin",
"Peilin Zhao",
"Junzhou Huang",
"Shenghua Gao"
],
"authorids": [
"liandz@shanghaitech.edu.cn",
"yzheng3xg@gmai... | 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": "PC",
"data": {
"comment": "This paper introduces T-NAS, a neural architecture search (NAS) method that can quickly adapt architectures to new datasets based on gradient-based meta-learning. It is a combination of the NAS method DARTS and the meta-learning method MAML.\n\nAll rev... | [
"Summary:\nCurrent neural architecture search (NAS) methods work in the mode of what is called S1 in this paper: given a dataset, search for a new architecture from scratch for that particular dataset. This paper proposes using MAML-style metalearning for learning meta-architectures across many meta-training tasks ... | [
[
12
],
[
3
],
[
5
],
[
7
],
[
1,
8
],
[
9
],
[
11
],
[
6
],
[
13
],
[
14
],
[
4
],
[
0
],
[
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",
"correct",
"correct"
] | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0,
1
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
2
]
},
{
"role": "Author",
... | benchmark/PDF/ICLR2020_r1eowANFvr.pdf | openreview | benchmark/MD/ICLR2020_r1eowANFvr.md | ICLR 2020 |
rkeIIkHKvS | {
"authorids": [
"yfhou@cse.cuhk.edu.hk",
"jzhang@cse.cuhk.edu.hk",
"jcheng@cse.cuhk.edu.hk",
"klma@cse.cuhk.edu.hk",
"tbma@comp.nus.edu.sg",
"hzchen@cse.cuhk.edu.hk",
"mcyang@cse.cuhk.edu.hk"
],
"title": "Measuring and Improving the Use of Graph Information in Graph Neural Networks",
... | Accept (Poster) | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2
]
},
"scores": {
"Solid": null,
"Presentation": null,
"Novelty": null,
"Overa... | [
[
{
"role": "PC",
"data": {
"comment": "Two reviewers are positive about this paper while the other reviewer is negative. The low-scoring reviewer did not respond to discussions. I also read the paper and found it interesting. Thus an accept is recommended."
}
}
]
] | [
"The authors study how neighbor information on graphs can be used in Graph Neural Networks. It proposes measures on whether the data in neighboring nodes are useful in terms of labels or features. It also provides a new Graph Neural Network algorithm that is a modification of attention-based models incorporating th... | [
[
1
],
[
11
],
[
5
],
[
9
],
[
10
],
[
10
],
[
10
],
[
10
],
[
13
],
[
8
],
[
3
],
[
6
],
[
4
],
[
12
],
[
0
],
[
2
],
[
7
]
] | [
"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"
] | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1,
2
]
},
{
"role": "Author",
... | benchmark/PDF/ICLR2020_rkeIIkHKvS.pdf | openreview | benchmark/MD/ICLR2020_rkeIIkHKvS.md | ICLR 2020 |
ryxjnREFwH | {
"authorids": [
"xinyun.chen@berkeley.edu",
"crazydonkey@google.com",
"adamsyuwei@google.com",
"dennyzhou@google.com",
"dawnsong.travel@gmail.com",
"qvl@google.com"
],
"title": "Neural Symbolic Reader: Scalable Integration of Distributed and Symbolic Representations for Reading Comprehens... | Accept (Spotlight) | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
... | [
[
{
"role": "PC",
"data": {
"comment": "Main content:\n\nBlind review #1 summarizes it well:\n\nThis paper presents a semantic parser that operates over passages of text instead of a structured data source. This is the first time anyone has demonstrated such a semantic parser (Siva Reddy an... | [
"This paper presents a semantic parser that operates over passages of text instead of a structured data source. This is the first time anyone has demonstrated such a semantic parser (Siva Reddy and several others have essentially used unstructured text as an information source for a semantic parser, similar to Ope... | [
[
17
],
[
12
],
[
13
],
[
16
],
[
20
],
[
22
],
[
0
],
[
1,
2
],
[
7,
19
],
[
9
],
[
10
],
[
3
],
[
4
],
[
5
],
[
6,
11
],
[
14
],
[
15
],
[
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",
"incorrect",
"incorrect",
"correct",
"correct",
"correct",
"incorrect",
"incorrect",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct"
] | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0
]
}
],
"category": [
"ORIG-MTH"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1,
2,
3
]
},
{
"role"... | benchmark/PDF/ICLR2020_ryxjnREFwH.pdf | openreview | benchmark/MD/ICLR2020_ryxjnREFwH.md | ICLR 2020 |
rJe2syrtvS | {
"authorids": [
"henryzhu@berkeley.edu",
"justinvyu@berkeley.edu",
"abhigupta@berkeley.edu",
"shah@eecs.berkeley.edu",
"kristian.hartikainen@gmail.com",
"avisingh@cs.berkeley.edu",
"vikashplus@gmail.com",
"svlevine@eecs.berkeley.edu"
],
"title": "The Ingredients of Real World Robo... | Accept (Spotlight) | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
... | [
[
{
"role": "PC",
"data": {
"comment": "This is a very interesting paper which discusses practical issues and solutions around deploying RL on real physical robotic systems, specifically involving questions on the use of raw sensory data, crafting reward functions, and not having resets at t... | [
"The paper takes seriously the question of having a robotic system learning continuously without manual reset nor state or reward engineering. The authors propose a first approach using vison-based SAC, shown visual goals and VICE, and show that it does not provide a satisfactory solution.",
"Then they add a rand... | [
[
3
],
[
7
],
[
18
],
[
6
],
[
19,
22
],
[
20
],
[
1
],
[
2
],
[
4
],
[
5
],
[
8
],
[
10
],
[
23
],
[
24
],
[
14
],
[
15
],
[
16
],
[
17
],
[
0
],
[
9
],
... | [
"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,
2
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
3
]
},
{
"role": "Au... | benchmark/PDF/ICLR2020_rJe2syrtvS.pdf | openreview | benchmark/MD/ICLR2020_rJe2syrtvS.md | ICLR 2020 |
rkgbYyHtwB | {
"title": "Disagreement-Regularized Imitation Learning",
"authors": [
"Kiante Brantley",
"Wen Sun",
"Mikael Henaff"
],
"authorids": [
"kdbrant@cs.umd.edu",
"wen.sun@microsoft.com",
"mihenaff@microsoft.com"
],
"keywords": [
"imitation learning",
"reinforcement learning",
... | Accept (Spotlight) | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
... | [
[
{
"role": "PC",
"data": {
"comment": "This paper presents an approach for interactive imitation learning while avoiding an adversarial optimization by using ensembles. The reviewers agreed that the contributions were significant and the results were compelling. Hence, the paper should be a... | [
"* Summary:\nThe paper aims to address the covariate shift issue of behavior cloning (BC). The main idea of the paper is to learn a policy by minimizing a BC loss and an uncertainty loss. This uncertainty loss is defined as a variance of a policy posterior given by demonstration. To approximate this posterior, the ... | [
[
32
],
[
19
],
[
21
],
[
22
],
[
23
],
[
33
],
[
35
],
[
38
],
[
10
],
[
24
],
[
30
],
[
31
],
[
34
],
[
3
],
[
6
],
[
11
],
[
13
],
[
14
],
[
27
],
[
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",
"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",
"data": [
1,
2
]
}
],
"category": [
"QUAL-MET"
... | benchmark/PDF/ICLR2020_rkgbYyHtwB.pdf | openreview | benchmark/MD/ICLR2020_rkgbYyHtwB.md | ICLR 2020 |
ryxdEkHtPS | {
"title": "A Closer Look at Deep Policy Gradients",
"authors": [
"Andrew Ilyas",
"Logan Engstrom",
"Shibani Santurkar",
"Dimitris Tsipras",
"Firdaus Janoos",
"Larry Rudolph",
"Aleksander Madry"
],
"authorids": [
"ailyas@mit.edu",
"engstrom@mit.edu",
"shibani@mit.edu",
... | Accept (Talk) | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6
]
},
"scores": {
"Solid": null,
"Pres... | [
[
{
"role": "PC",
"data": {
"comment": "The paper empirically studies the behaviour of deep policy gradient algorithms, and reveals several unexpected observations that are not explained by the current theory. All three reviewers are excited about this work and recommend acceptance."
}... | [
"The paper explores a critical divergence between theory and practice, emphasizing that while deep policy gradient algorithms seem to work in certain cases, they don't seem to be working foor the reasons underlying their derivations. It particularly looks at how closely the sample-based approximation of the objecti... | [
[
3
],
[
4
],
[
5
],
[
9
],
[
10
],
[
13
],
[
2
],
[
6
],
[
14
],
[
15
],
[
16
],
[
17
],
[
18
],
[
1
],
[
8
],
[
11
],
[
12
],
[
0
],
[
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",
"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/ICLR2020_ryxdEkHtPS.pdf | openreview | benchmark/MD/ICLR2020_ryxdEkHtPS.md | ICLR 2020 |
rkg1ngrFPr | {
"authorids": [
"piotr.sokol@stonybrook.edu",
"memming.park@stonybrook.edu"
],
"title": "Information Geometry of Orthogonal Initializations and Training",
"authors": [
"Piotr Aleksander Sokół",
"Il Memming Park"
],
"pdf": "/pdf/518fe69dd850c85be75561849db48ae9cbadf710.pdf",
"TL;DR": "near... | Accept (Poster) | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4
]
},
"scores": {
"Solid": null,
"Presentation": null,
"No... | [
[
{
"role": "PC",
"data": {
"comment": "I've gone over this paper carefully and think it's above the bar for ICLR.\n\nThe paper proves a relationship between the eigenvalues of the Fisher information matrix and the singular values of the network Jacobian. The main step is bounding the eigenv... | [
"This paper analyzes the connection between the spectrum of the layer-to-layer or input-output Jacobian matrices and the spectrum of the Fisher information matrix / Neural Tangent Kernel. By bounding the maximum eigenvalue of the Fisher in terms of the maximum squared singular value of the input-output Jacobian, th... | [
[
12
],
[
5
],
[
11
],
[
6
],
[
7
],
[
2
],
[
10
],
[
14
],
[
1
],
[
9
],
[
13
],
[
0
],
[
3
],
[
4
],
[
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",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct"
] | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0,
1
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
2
]
},
{
"role": "Author",
... | benchmark/PDF/ICLR2020_rkg1ngrFPr.pdf | openreview | benchmark/MD/ICLR2020_rkg1ngrFPr.md | ICLR 2020 |
S1e2agrFvS | {
"authorids": [
"gspeihongbing@163.com",
"bwei6@illinois.edu",
"kcchang@illinois.edu",
"csylei@comp.polyu.edu.hk",
"ybo@jlu.edu.cn"
],
"title": "Geom-GCN: Geometric Graph Convolutional Networks",
"authors": [
"Hongbin Pei",
"Bingzhe Wei",
"Kevin Chen-Chuan Chang",
"Yu Lei",
... | Accept (Spotlight) | [
[
{
"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": "Nice work!\nI have some questions about the details and model.\n1. If the high-level aggregation function q is concatenation, the dimension of $m^{l+1}_v$ is not fixed, which may change with the number of virtual nodes. So how can you solve this? To my... | [
"This work proposes geometric aggregation scheme for GCNs, which aims to overcome the limitations in traditional GCNs; those are lacking long distance dependencies and structure information in nodes.",
"In particular, each node is transformed into a latent space. To overcome the first limitation, some nodes that ... | [
[
19
],
[
29
],
[
30
],
[
7
],
[
25
],
[
26
],
[
2
],
[
3
],
[
4
],
[
9
],
[
27
],
[
28
],
[
0
],
[
1
],
[
22
],
[
5
],
[
6
],
[
11
],
[
14
],
[
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",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"c... | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0,
1
]
}
],
"category": [
"ORIG-MTH"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
2
]
}
],
"category": [
"ORIG... | benchmark/PDF/ICLR2020_S1e2agrFvS.pdf | openreview | benchmark/MD/ICLR2020_S1e2agrFvS.md | ICLR 2020 |
r1eyceSYPr | {
"authorids": [
"yixuanq@andrew.cmu.edu",
"lingsong@purdue.edu",
"wangxiao@purdue.edu"
],
"title": "Unbiased Contrastive Divergence Algorithm for Training Energy-Based Latent Variable Models",
"authors": [
"Yixuan Qiu",
"Lingsong Zhang",
"Xiao Wang"
],
"pdf": "/pdf/657b4d2249efb2827... | Accept (Spotlight) | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
... | [
[
{
"role": "PC",
"data": {
"comment": "Main content:\n\nBlind review #1 summarizes it well:\n\nThe paper proposes an algorithmic improvement that significantly simplifies training of energy-based models, such as the Restricted Boltzmann Machine. The key issue in training such models is comp... | [
"The paper proposes an algorithmic improvement that significantly simplifies training of energy-based models, such as the Restricted Boltzmann Machine. The key issue in training such models is computing the gradient of the log partition function, which can be framed as computing the expected value of f(x) = dE(x; t... | [
[
7
],
[
17
],
[
18
],
[
19
],
[
20
],
[
25
],
[
3
],
[
9
],
[
23
],
[
24
],
[
21
],
[
29
],
[
30
],
[
34
],
[
5
],
[
6
],
[
11
],
[
12
],
[
14
],
[
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": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
2
]
}
],
"category": [
"QUAL-MET"... | benchmark/PDF/ICLR2020_r1eyceSYPr.pdf | openreview | benchmark/MD/ICLR2020_r1eyceSYPr.md | ICLR 2020 |
rylwJxrYDS | {
"authorids": [
"alexei.b@gmail.com",
"stes@fb.com",
"michael.auli@gmail.com"
],
"title": "vq-wav2vec: Self-Supervised Learning of Discrete Speech Representations",
"authors": [
"Alexei Baevski",
"Steffen Schneider",
"Michael Auli"
],
"pdf": "/pdf/0a5ac6d85b01d047385eff9fc4507ef6fc0... | 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": "PC",
"data": {
"comment": "This paper proposes a new self-supervised pre-trained speech model that improves speech recognition performance. \n The idea combines an earlier pre-training approach (wav2vec) with discretization followed by BERT-style masked reconstruction. The resu... | [
"Overview:\nThis paper considers unsupervised (or self-supervised) discrete representation learning of speech using a combination of a recent vector quantized neural network discritization method and future time step prediction. Discrete representations are fine-tuned by using these as input to a BERT model; the re... | [
[
3
],
[
4
],
[
7
],
[
9
],
[
10
],
[
11
],
[
23
],
[
1
],
[
5
],
[
6
],
[
2
],
[
15
],
[
22
],
[
28
],
[
12
],
[
13
],
[
14
],
[
8
],
[
18,
21
],
[
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",
"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": [
"ORIG-COM",
"SIGN... | benchmark/PDF/ICLR2020_rylwJxrYDS.pdf | openreview | benchmark/MD/ICLR2020_rylwJxrYDS.md | ICLR 2020 |
rke2P1BFwS | {
"title": "Tensor Decompositions for Temporal Knowledge Base Completion",
"authors": [
"Timothée Lacroix",
"Guillaume Obozinski",
"Nicolas Usunier"
],
"authorids": [
"timothee.lax@gmail.com",
"guillaume.obozinski@epfl.ch",
"usunier@fb.com"
],
"keywords": [
"knowledge base comple... | Accept (Poster) | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7
]
},
"scores": {
"Solid": null,
... | [
[
{
"role": "PC",
"data": {
"comment": "The authors propose a new algorithm based on tensor decompositions for the problem of knowledge base completion. They also introduce new regularisers to augment their method. They also propose an new dataset for temporal KB completion. \n\nAll the rev... | [
"In this paper, the authors study an important problem, i.e., time-aware link prediction in a knowledge base.",
"Specifically, the authors focus on predicting the missing link in a quadruple, i.e., (subject, predicate, ?, timestamp).",
"In particular, the authors design a new tensor (order 4) factorization base... | [
[
23
],
[
5
],
[
14
],
[
32
],
[
34
],
[
1,
8,
33
],
[
3
],
[
7
],
[
9,
28
],
[
26
],
[
36
],
[
4
],
[
15
],
[
10
],
[
11
],
[
13
],
[
18,
27
],
[
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",
"cor... | [
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"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/ICLR2020_rke2P1BFwS.pdf | openreview | benchmark/MD/ICLR2020_rke2P1BFwS.md | ICLR 2020 |
Skgq1ANFDB | {
"title": "Curvature-based Robustness Certificates against Adversarial Examples",
"authors": [
"Sahil Singla",
"Soheil Feizi"
],
"authorids": [
"ssingla@cs.umd.edu",
"sfeizi@cs.umd.edu"
],
"keywords": [
"Adversarial examples",
"Robustness certificates",
"Adversarial attacks",
... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6
]
},
"scores": {
"Solid": null,
"Pres... | [
[
{
"role": "PC",
"data": {
"comment": "This paper presents a upper bound on the curvature of a deep network. After the discussion, the author has addressed some concerns of reviwers, but the results are not very strong, there is some limitation on the applications. There is no strong suppor... | [
"Summary\nThis paper proposes the Curvature-based Robustness Certificate (CRC) and Curvature-based Robust Training (CRT) for robustness certificate against adversarial examples. The proposed techniques are theoretically formulated and empirically justified. The authors showed that, when the curvature (Hessian) of t... | [
[
6,
11
],
[
10
],
[
5
],
[
1
],
[
2
],
[
4
],
[
7,
20,
23
],
[
12
],
[
24
],
[
28
],
[
16
],
[
18
],
[
0
],
[
3
],
[
8,
21
],
[
9
],
[
29
],
[
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",
"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-EXP"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
2
]
},
{
"role": "Author",
... | benchmark/PDF/ICLR2020_Skgq1ANFDB.pdf | openreview | benchmark/MD/ICLR2020_Skgq1ANFDB.md | ICLR 2020 |
r1gelyrtwH | {
"title": "Physics-aware Difference Graph Networks for Sparsely-Observed Dynamics",
"authors": [
"Sungyong Seo*",
"Chuizheng Meng*",
"Yan Liu"
],
"authorids": [
"sungyons@usc.edu",
"chuizhem@usc.edu",
"yanliu.cs@usc.edu"
],
"keywords": [
"physics-aware learning",
"spatial di... | Accept (Poster) | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10
]
},... | [
[
{
"role": "PC",
"data": {
"comment": "All reviewers agree that this research is novel and well carried out, so this is a clear accept. Please ensure that the final version reflect the reviewer comments and the new information provided during the rebuttal"
}
}
]
] | [
"This paper proposes a method to reduce numerical error when predicting sequences governed by physical dynamics. The idea is that most physics simulators use finite difference differential operators, and the paper adds trainable parameters to them (the derivative and Laplacian operators). The added parameters are c... | [
[
9
],
[
24
],
[
26
],
[
10
],
[
3
],
[
7
],
[
21
],
[
2
],
[
4
],
[
5
],
[
12
],
[
13
],
[
14
],
[
15
],
[
19
],
[
22
],
[
23
],
[
27
],
[
1
],
[
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",
"incorrect",
"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/ICLR2020_r1gelyrtwH.pdf | openreview | benchmark/MD/ICLR2020_r1gelyrtwH.md | ICLR 2020 |
SkxBUpEKwH | {
"title": "Vid2Game: Controllable Characters Extracted from Real-World Videos",
"authors": [
"Oran Gafni",
"Lior Wolf",
"Yaniv Taigman"
],
"authorids": [
"oran.gafni@gmail.com",
"wolf@fb.com",
"yaniv@fb.com"
],
"keywords": [],
"abstract": "We extract a controllable model from a vi... | Accept (Poster) | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10
]
},... | [
[
{
"role": "PC",
"data": {
"comment": "This paper proposes to extract a character from a video, manually control the character, and render into the background in real time. The rendered video can have arbitrary background and capture both the dynamics and appearance of the person. All thre... | [
"This paper proposes a method to address the interesting task, i.e. controllable human activity synthesis, by conditioning on the previous frames and the input control signal. To synthesis the next frame, a Pose2Pose network is proposed to first transfer the input information into the next frame body structure and ... | [
[
5
],
[
10
],
[
13
],
[
4
],
[
6
],
[
7
],
[
8
],
[
11
],
[
2
],
[
15
],
[
3
],
[
1
],
[
14
],
[
0
],
[
9
],
[
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",
"correct",
"correct",
"correct",
"correct"
] | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1
]
}
],
"category": [
"SIGN-SOT"
]
},
... | benchmark/PDF/ICLR2020_SkxBUpEKwH.pdf | openreview | benchmark/MD/ICLR2020_SkxBUpEKwH.md | ICLR 2020 |
SJleNCNtDH | {
"authorids": [
"ronuchit@mit.edu",
"shubhtuls@fb.com",
"saurabhg@illinois.edu",
"abhinavg@cs.cmu.edu"
],
"title": "Intrinsic Motivation for Encouraging Synergistic Behavior",
"authors": [
"Rohan Chitnis",
"Shubham Tulsiani",
"Saurabh Gupta",
"Abhinav Gupta"
],
"pdf": "/pdf/... | Accept (Poster) | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4
]
},
"scores": {
"Solid": null,
"Presentation": null,
"No... | [
[
{
"role": "PC",
"data": {
"comment": "The authors address the important issue of exploration in reinforcement learning. In this case, they propose to use reward shaping to encourage joint-actions whose outcomes deviate from the sequential counterpart. Although the proposed intrinsic reward... | [
"The paper focuses on using intrinsic motivation to improve the exploration process of reinforcement learning agents in tasks with sparse-reward and that require multi-agent to achieve. The authors proposed to encourage the agents toward the actions which changed the world in the ways that \"would not be achieved i... | [
[
10
],
[
2
],
[
3
],
[
5
],
[
6
],
[
8
],
[
9
],
[
11
],
[
1
],
[
4
],
[
0
],
[
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"
] | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1,
2
]
}
],
"category": [
"QUAL-MET"... | benchmark/PDF/ICLR2020_SJleNCNtDH.pdf | openreview | benchmark/MD/ICLR2020_SJleNCNtDH.md | ICLR 2020 |
SygW0TEFwH | {
"title": "Sign Bits Are All You Need for Black-Box Attacks",
"authors": [
"Abdullah Al-Dujaili",
"Una-May O'Reilly"
],
"authorids": [
"ash.aldujaili@gmail.com",
"unamay@csail.mit.edu"
],
"keywords": [
"Black-box adversarial attack models",
"Deep Nets",
"Adversarial Examples",
... | Accept (Poster) | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7
]
},
"scores": {
"Solid": null,
... | [
[
{
"role": "PC",
"data": {
"comment": "This paper presents a novel black-box adversarial attack algorithm, which exploits a sign-based rather than magnitude-based, gradient estimator for black-box optimization. It also adaptively constructs queries to estimate the gradient. The proposed app... | [
"I'm satisfied with the response. I'll keep my original rating towards acceptance.",
"This paper proposes a black-box adversarial attack method to improve query efficiency and attack success rate.",
"Instead of estimating the gradient of a black-box model, the proposed method estimates the sign of the gradient,... | [
[
3
],
[
1
],
[
0,
15
],
[
6,
8,
11,
14
],
[
16
],
[
17
],
[
2
],
[
5,
10,
13
],
[
7,
12
],
[
4
],
[
9
]
] | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"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
]
}
],
"category": [
"ORIG-MTH"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
4,
5
]
... | benchmark/PDF/ICLR2020_SygW0TEFwH.pdf | openreview | benchmark/MD/ICLR2020_SygW0TEFwH.md | ICLR 2020 |
rygeHgSFDH | {
"authorids": [
"peter.sorrenson@gmail.com",
"carsten.rother@iwr.uni-heidelberg.de",
"ullrich.koethe@iwr.uni-heidelberg.de"
],
"title": "Disentanglement by Nonlinear ICA with General Incompressible-flow Networks (GIN)",
"authors": [
"Peter Sorrenson",
"Carsten Rother",
"Ullrich Köthe"
... | Accept (Spotlight) | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10
]
},... | [
[
{
"role": "PC",
"data": {
"comment": "This paper builds on the recent theoretical work by Khemakhem et al. (2019) to propose a novel flow-based method for performing non-linear ICA. The paper is well written, includes theoretical justifications for the proposed approach and convincing expe... | [
"This paper builds upon the recent theoretical framework on nonlinear ICA, put forward in recent work Khemakhem et al. (2019) that draw a lot of attention. The latter work provides an extension of the basic nonlinear ICA that is closely related to a VAE with a conditional factorized prior, essentially introducing s... | [
[
29
],
[
35
],
[
4
],
[
6
],
[
8,
8,
13,
13
],
[
9,
9
],
[
10
],
[
15
],
[
19
],
[
21
],
[
33
],
[
7
],
[
30
],
[
2,
16
],
[
3
],
[
5,
11,
24,
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",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"c... | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0,
1
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
2
]
}
],
"category": [
"SIGN-BRD"... | benchmark/PDF/ICLR2020_rygeHgSFDH.pdf | openreview | benchmark/MD/ICLR2020_rygeHgSFDH.md | ICLR 2020 |
Skl4mRNYDr | {
"authorids": [
"nrhineha@cs.cmu.edu",
"rmcallister@berkeley.edu",
"svlevine@eecs.berkeley.edu"
],
"title": "Deep Imitative Models for Flexible Inference, Planning, and Control",
"authors": [
"Nicholas Rhinehart",
"Rowan McAllister",
"Sergey Levine"
],
"pdf": "/pdf/a0ea230e9b98f012c... | 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": "PC",
"data": {
"comment": "This paper proposes to build an 'imitative model' to improve the performance for imitation learning. The main idea is to combine the model-based RL type of work to the imitation learning approach. The model is trained using a probabilistic method and c... | [
"Summary:",
"- key problem: expert-like probabilistic online motion planning to reach arbitrary goals without reward shaping thanks to off-line learning from expert demonstrations;",
"- contributions: 1) an imitative planning procedure via gradient-based log-likelihood maximization leveraging \"imitative models... | [
[
16
],
[
22
],
[
17
],
[
21
],
[
25
],
[
28
],
[
32
],
[
33
],
[
2
],
[
27
],
[
4
],
[
20
],
[
9
],
[
29
],
[
7
],
[
8
],
[
10
],
[
11
],
[
13
],
[
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",
"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": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer",
"data": [
2,
3
]
}
],
"category": [
... | benchmark/PDF/ICLR2020_Skl4mRNYDr.pdf | openreview | benchmark/MD/ICLR2020_Skl4mRNYDr.md | ICLR 2020 |
SJxzFySKwH | {
"authorids": [
"bsriniv@purdue.edu",
"ribeiro@cs.purdue.edu"
],
"title": "On the Equivalence between Positional Node Embeddings and Structural Graph Representations",
"authors": [
"Balasubramaniam Srinivasan",
"Bruno Ribeiro"
],
"pdf": "/pdf/f09dce2089f54246bef56fb0cbd14717aa609772.pdf",
... | Accept (Poster) | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8
]
},
"scores": {
... | [
[
{
"role": "PC",
"data": {
"comment": "The paper shows the relationship between node embeddings and structural graph representations. By careful definition of what structural node representation means, and what node embedding means, using the permutation group, the authors show in Theorem 2... | [
"The authors present mostly theoretical analysis indicating the equivalence of embeddings and structural graph representations. The authors argue that while most of the earlier work consider these to be different, they are actually the same and give theory and empirical results to back up this claim.",
"This is n... | [
[
1
],
[
2
],
[
9
],
[
10
],
[
7
],
[
5
],
[
6
],
[
0
],
[
3
],
[
4
],
[
8
]
] | [
"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/ICLR2020_SJxzFySKwH.pdf | openreview | benchmark/MD/ICLR2020_SJxzFySKwH.md | ICLR 2020 |
Sye57xStvB | {
"title": "Never Give Up: Learning Directed Exploration Strategies",
"authors": [
"Adrià Puigdomènech Badia",
"Pablo Sprechmann",
"Alex Vitvitskyi",
"Daniel Guo",
"Bilal Piot",
"Steven Kapturowski",
"Olivier Tieleman",
"Martin Arjovsky",
"Alexander Pritzel",
"Andrew Bolt",
... | Accept (Poster) | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4
]
},
"scores": {
"Solid": null,
"Presentation": null,
"No... | [
[
{
"role": "PC",
"data": {
"comment": "This paper tackles hard-exploration RL problems. The idea is to learn separate exploration and exploitation strategies using the same network (representation). The exploration is driven by intrinsic rewards, which are generated using an episodic memory... | [
"The paper proposes a novel intrinsic reward/curiosity metric that combines both episodic and “life-long” novelty. Essentially two competing pressures that push agents to explore as many novel states in a single rollout as possible and to explore as many states as possible as evenly as possible. The primary contrib... | [
[
22
],
[
3
],
[
5
],
[
8
],
[
12
],
[
13
],
[
19
],
[
21
],
[
23
],
[
24
],
[
25
],
[
1
],
[
28
],
[
7,
26
],
[
9
],
[
10
],
[
15
],
[
17
],
[
18
],
[
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",
"correct",
"correct",
"correct",
"correct",
"cor... | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1
]
}
],
"category": [
"ORIG-MTH"
]
},
... | benchmark/PDF/ICLR2020_Sye57xStvB.pdf | openreview | benchmark/MD/ICLR2020_Sye57xStvB.md | ICLR 2020 |
r1g87C4KwB | {
"authorids": [
"staszek.jastrzebski@gmail.com",
"msz93@o2.pl",
"stanislav.fort@gmail.com",
"devansharpit@gmail.com",
"jcktbr@gmail.com",
"kyunghyun.cho@nyu.edu",
"k.j.geras@nyu.edu"
],
"title": "The Break-Even Point on Optimization Trajectories of Deep Neural Networks",
"authors": ... | Accept (Spotlight) | [
[
{
"role": "Author",
"data": {
"value": {
"comment": [
0
]
}
}
}
],
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
1,
2,
... | [
[
{
"role": "PC",
"data": {
"comment": "This is an interesting study analyzing learning trajectories and their dependence on hyperparameters, important for better understanding of learning in deep neural networks. All reviewers agree that the paper has a useful message to the ICLR community... | [
"We forgot to mention that ResNet-32 in Section 4.2 does not use batch normalization. In 4.3 we extend the analysis to batch normalized networks.",
"This work analyzes the optimization of deep neural networks from the point of view of the different learning trajectories obtained during different learning settings... | [
[
1
],
[
9
],
[
11
],
[
25
],
[
26
],
[
28
],
[
7
],
[
10
],
[
15
],
[
23
],
[
12
],
[
6
],
[
17
],
[
3
],
[
8
],
[
13
],
[
14
],
[
19
],
[
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",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"c... | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
1,
2,
3
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer",
"data": [
4
]
}
],
"category": [
... | benchmark/PDF/ICLR2020_r1g87C4KwB.pdf | openreview | benchmark/MD/ICLR2020_r1g87C4KwB.md | ICLR 2020 |
Hyx0slrFvH | {
"authorids": [
"stefan.uhlich@sony.com",
"lukas.mauch@sony.com",
"fabien.cardinaux@sony.com",
"kazuki.yoshiyama@sony.com",
"javier.alonso@sony.com",
"stephen.tiedemann@sony.com",
"thomas.kemp@sony.com",
"akira.b.nakamura@sony.com"
],
"title": "Mixed Precision DNNs: All you need i... | Accept (Poster) | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9
]
},
"scores... | [
[
{
"role": "PC",
"data": {
"comment": "The reviewers uniformly vote to accept this paper. Please take comments into account when revising for the camera ready. I was also very impressed by the authors' responsiveness to reviewer comments, putting in additional work after submission."
... | [
"This paper considers the problem of training mixed-precision models.",
"Since quantization involves non-differentiable operations, this paper discusses how to use the straight-through estimator to estimate the gradients, and how different parameterizations of the quantized DNN affect the optimization process. T... | [
[
2
],
[
7,
10
],
[
3
],
[
4
],
[
8
],
[
1
],
[
5
],
[
9
],
[
11
],
[
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"
] | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0,
1,
2
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
3
]
},
{
"role": "Au... | benchmark/PDF/ICLR2020_Hyx0slrFvH.pdf | openreview | benchmark/MD/ICLR2020_Hyx0slrFvH.md | ICLR 2020 |
ryeYpJSKwr | {
"authorids": [
"mvolpp89@googlemail.com",
"lukas.froehlich@de.bosch.com",
"k.fischer-lotte@online.de",
"andreas.doerr3@de.bosch.com",
"stefan.falkner@de.bosch.com",
"fh@cs.uni-freiburg.de",
"christian.daniel@de.bosch.com"
],
"title": "Meta-Learning Acquisition Functions for Transfer ... | Accept (Spotlight) | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
... | [
[
{
"role": "PC",
"data": {
"comment": "This paper explores the idea of using meta-learning for acquisition functions. It is an interesting and novel research direction with promising results. \n\nThe paper could be strengthened by adding more insights about the new acquisition function and ... | [
"The authors present MetaBO, which uses reinforcement learning to meta-learn the acquisition function (AF) for Bayesian Optimization (BO) instead of using a standard constant AF. The authors shows that MetaBO enables transferring knowledge between tasks and increasing sample efficiency on new tasks. The paper is mo... | [
[
16
],
[
28
],
[
31
],
[
32
],
[
41
],
[
11,
12
],
[
7
],
[
13
],
[
34
],
[
42
],
[
1,
2
],
[
5
],
[
6
],
[
9
],
[
10
],
[
14
],
[
24
],
[
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",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"i... | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1
]
},
{
"role": "Author",
"data": [
... | benchmark/PDF/ICLR2020_ryeYpJSKwr.pdf | openreview | benchmark/MD/ICLR2020_ryeYpJSKwr.md | ICLR 2020 |
rklr9kHFDB | {
"authorids": [
"ivan.ustyuzhaninov@bethgelab.org",
"santiago.cadena@bethgelab.org",
"froudara@bcm.edu",
"paul.fahey@bcm.edu",
"eywalker@bcm.edu",
"ecobos@bcm.edu",
"reimer@bcm.edu",
"fabian.sinz@bcm.edu",
"astolias@bcm.edu",
"matthias@bethgelab.org",
"alexander.ecker@uni-... | Accept (Talk) | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9
]
},
"scores... | [
[
{
"role": "PC",
"data": {
"comment": "This paper is enthusiastically supported by all three reviewers. Thus an accept is recommended."
}
}
]
] | [
"The authors present a rotation-invariant representation of a CNN modeling the V1 neurons and a pipeline to cluster these neurons to find cell types that are rotation-invariant. Experimental validation is performed on a 6K neuron dataset with promising results.",
"The paper is well postulated.",
"Below are comm... | [
[
1
],
[
4
],
[
5
],
[
24
],
[
17
],
[
18
],
[
26
],
[
25
],
[
3
],
[
6
],
[
19
],
[
21
],
[
22
],
[
23
],
[
27
],
[
29
],
[
2
],
[
20
],
[
30
],
[
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",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"incorrect",
"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/ICLR2020_rklr9kHFDB.pdf | openreview | benchmark/MD/ICLR2020_rklr9kHFDB.md | ICLR 2020 |
SJxSOJStPr | {
"authorids": [
"soochan.lee@vision.snu.ac.kr",
"junsooha@hanyang.ac.kr",
"96lives@snu.ac.kr",
"gunhee@snu.ac.kr"
],
"title": "A Neural Dirichlet Process Mixture Model for Task-Free Continual Learning",
"authors": [
"Soochan Lee",
"Junsoo Ha",
"Dongsu Zhang",
"Gunhee Kim"
],
... | Accept (Poster) | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9
]
},
"scores... | [
[
{
"role": "PC",
"data": {
"comment": "This paper proposes an expansion-based approach for task-free continual learning, using a Bayesian nonparametric framework (a Dirichlet process mixture model).\n\nIt was well-reviewed, with reviewers agreeing that the paper is well-written, the experim... | [
"This paper proposes Continual Neural Dirichlet Process Mixture Model (CN-DPM) to solve task-free continual learning. The core idea is to employ Dirichlet process mixture model to create novel experts in online fashion when task distributions change. The proposed method is validated on various tasks and demonstrat... | [
[
1
],
[
16
],
[
22
],
[
9
],
[
18
],
[
19
],
[
21
],
[
2,
8,
10
],
[
11
],
[
12
],
[
15
],
[
6
],
[
13
],
[
14
],
[
20
],
[
24
],
[
3
],
[
4
],
[
5
],
[
... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"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",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
... | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1
]
}
],
"category": [
"CLAR-WRT",
"QUAL... | benchmark/PDF/ICLR2020_SJxSOJStPr.pdf | openreview | benchmark/MD/ICLR2020_SJxSOJStPr.md | ICLR 2020 |
SkxxtgHKPS | {
"authorids": [
"ljiian83@mail.tsinghua.edu.cn",
"luo-xy19@mails.tsinghua.edu.cn",
"mqiao@stanford.edu"
],
"title": "On Generalization Error Bounds of Noisy Gradient Methods for Non-Convex Learning",
"authors": [
"Jian Li",
"Xuanyuan Luo",
"Mingda Qiao"
],
"pdf": "/pdf/9020c8430bc5d... | Accept (Poster) | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7
]
},
"scores": {
"Solid": null,
... | [
[
{
"role": "PC",
"data": {
"comment": "The authors provide bounds on the expected generalization error for noisy gradient methods (such as SGLD). They do so using the information theoretic framework initiated by Russo and Zou, where the expected generalization error is controlled by the mut... | [
"This paper studies the generalization error bounds of stochastic gradient Langevin dynamics. The convexity of the loss function is not assumed. The author proposed \"Bayes-stability\" to derive generalization bound while taking the randomness of the algorithm into account. The generalization bound proposed in this... | [
[
2
],
[
4
],
[
1
],
[
0
],
[
10
],
[
11
],
[
6
],
[
7,
9
],
[
12
],
[
3
],
[
5
],
[
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"
] | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0,
1,
2,
3
]
}
],
"category": [
"CLAR-WRT"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
4,
5
]
... | benchmark/PDF/ICLR2020_SkxxtgHKPS.pdf | openreview | benchmark/MD/ICLR2020_SkxxtgHKPS.md | ICLR 2020 |
rke-f6NKvS | {
"authorids": [
"yupingl@cs.princeton.edu",
"huazhe_xu@eecs.berkeley.edu",
"tengyuma@stanford.edu"
],
"title": "Learning Self-Correctable Policies and Value Functions from Demonstrations with Negative Sampling",
"authors": [
"Yuping Luo",
"Huazhe Xu",
"Tengyu Ma"
],
"pdf": "/pdf/ffd... | 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": "PC",
"data": {
"comment": "The paper introduces Value Iteration with Negative Sampling (VINS) algorithm as a method to accelerate RL using expert demonstrations. VINS learns an initial value function that has a smaller value at states not encounter during the demonstrations.\n\n... | [
"This work tried to address the covariate shift problem in imitation learning, which is due to the mismatch between training and test state distribution and may cause compounding errors.",
"The authors proposed the algorithm called value iteration with negative sampling (VINS) of which the main ideas can be summa... | [
[
19
],
[
10
],
[
11
],
[
12
],
[
23
],
[
26
],
[
27
],
[
28
],
[
29
],
[
2,
3
],
[
25
],
[
31
],
[
1
],
[
4
],
[
7
],
[
13
],
[
14
],
[
18
],
[
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",
"cor... | [
"correct",
"incorrect",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct"... | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0,
1,
2,
3,
4,
5
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
6
... | benchmark/PDF/ICLR2020_rke-f6NKvS.pdf | openreview | benchmark/MD/ICLR2020_rke-f6NKvS.md | ICLR 2020 |
r1xGP6VYwH | {
"title": "Optimistic Exploration even with a Pessimistic Initialisation",
"authors": [
"Tabish Rashid",
"Bei Peng",
"Wendelin Boehmer",
"Shimon Whiteson"
],
"authorids": [
"tabish.rashid@cs.ox.ac.uk",
"bei.peng@cs.ox.ac.uk",
"wendelin.boehmer@cs.ox.ac.uk",
"shimon.whiteson@cs.o... | 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": "PC",
"data": {
"comment": "The paper propose a scheme to enable optimistic initialization in the deep RL setting, and shows that it's helpful.\n\nThe reviewers agreed that the paper is well-motivated and executed, but had some minor reservations (e.g. about the proposal scaling ... | [
"Optimistic Exploration even with a Pessimistic Initialisation",
"This paper presents an exploration algorithm based on \"optimism in the face of uncertainty\" via count-based bonus.",
"The authors observe that typical neural net initializations close to zero can be pessimistic, but show that augmenting a count... | [
[
3
],
[
4
],
[
9
],
[
18
],
[
26
],
[
21
],
[
6,
12,
15
],
[
8
],
[
17
],
[
10,
19
],
[
20
],
[
23
],
[
27,
28
],
[
2
],
[
7,
16
],
[
24
],
[
1
],
[
0
],... | [
"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",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"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/ICLR2020_r1xGP6VYwH.pdf | openreview | benchmark/MD/ICLR2020_r1xGP6VYwH.md | ICLR 2020 |
rklHqRVKvH | {
"title": "Harnessing Structures for Value-Based Planning and Reinforcement Learning",
"authors": [
"Yuzhe Yang",
"Guo Zhang",
"Zhi Xu",
"Dina Katabi"
],
"authorids": [
"yuzhe@mit.edu",
"guozhang@mit.edu",
"zhixu@mit.edu",
"dina@csail.mit.edu"
],
"keywords": [
"Deep rein... | Accept (Talk) | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
... | [
[
{
"role": "PC",
"data": {
"comment": "The paper shows empirical evidence that the the optimal action-value function Q* often has a low-rank structure. It uses ideas from the matrix estimation/completion literature to provide a modification of value iteration that benefits from such a low-r... | [
"Summary:\nThis paper develops a method for taking advantage of structure in the value function to facilitate faster planning and learning. The key insight is that MDPs with low rank Q^* matrices can be solved more expediently using matrix estimation methods, both for classical dynamic programming methods (value it... | [
[
5
],
[
40
],
[
6
],
[
7
],
[
14
],
[
15
],
[
16
],
[
17
],
[
18
],
[
20
],
[
21
],
[
22
],
[
23
],
[
24
],
[
25
],
[
26
],
[
27
],
[
36
],
[
1
],
[
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",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"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": [
"ORIG-ANL"
]
},
... | benchmark/PDF/ICLR2020_rklHqRVKvH.pdf | openreview | benchmark/MD/ICLR2020_rklHqRVKvH.md | ICLR 2020 |
Syg-ET4FPS | {
"title": "Posterior sampling for multi-agent reinforcement learning: solving extensive games with imperfect information",
"authors": [
"Yichi Zhou",
"Jialian Li",
"Jun Zhu"
],
"authorids": [
"vofhqn@gmail.com",
"lijialia16@mails.tsinghua.edu.cn",
"dcszj@mail.tsinghua.edu.cn"
],
"ke... | Accept (Talk) | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
... | [
[
{
"role": "PC",
"data": {
"comment": "The paper extends posterior sampling to the multi-agent RL setting, and develops a novel algorithm with convergence guarantees to a Nash Equilibrium strategy in two-player zero sum games. Reviewers raised several questions, many of which were well addr... | [
"Posterior sampling for multi-agent reinforcement learning: solving extensive games with imperfect information",
"This paper investigates the use of Thompson sampling in multi-agent reinforcement learning.",
"They present a natural extension of the PSRL algorithm paired with counterfactual regret minimization, ... | [
[
27,
32
],
[
34
],
[
19
],
[
28
],
[
30
],
[
31
],
[
3
],
[
6
],
[
12
],
[
13
],
[
18
],
[
20
],
[
29
],
[
35
],
[
1
],
[
2
],
[
7,
10
],
[
14
],
[
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",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"incorrect",
"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/ICLR2020_Syg-ET4FPS.pdf | openreview | benchmark/MD/ICLR2020_Syg-ET4FPS.md | ICLR 2020 |
rkgyS0VFvr | {
"authorids": [
"chulinxie@zju.edu.cn",
"nick_cooper@sjtu.edu.cn",
"pin-yu.chen@ibm.com",
"lbo@illinois.edu"
],
"title": "DBA: Distributed Backdoor Attacks against Federated Learning",
"authors": [
"Chulin Xie",
"Keli Huang",
"Pin-Yu Chen",
"Bo Li"
],
"pdf": "/pdf/61dc789b9f... | 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": "PC",
"data": {
"comment": "Thanks for the discussion, all. This paper proposes an attack strategy against federated learning. Reviewers put this in the top tier, and the authors responded appropriately to their criticisms. "
}
}
]
] | [
"This paper studies backdoor attacks under federated learning setting. To inject a certain backdoor pattern, existing work generate poisoning samples by blending the same pattern with different input samples. Even for federated learning where the adversary can control multiple parties, such as [1], all parties stil... | [
[
14
],
[
0,
12
],
[
10
],
[
15
],
[
2
],
[
3
],
[
5
],
[
6
],
[
7
],
[
9
],
[
8
],
[
1
],
[
4
],
[
11
],
[
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"
] | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0,
1
]
}
],
"category": [
"ORIG-MTH"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
2
]
}
],
"category": [
"SIGN... | benchmark/PDF/ICLR2020_rkgyS0VFvr.pdf | openreview | benchmark/MD/ICLR2020_rkgyS0VFvr.md | ICLR 2020 |
rkgpv2VFvr | {
"abstract": "We study the benefit of sharing representations among tasks to enable the effective use of deep neural networks in Multi-Task Reinforcement Learning. We leverage the assumption that learning from different tasks, sharing common properties, is helpful to generalize the knowledge of them resulting in a m... | Accept (Poster) | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4
]
},
"scores": {
"Solid": null,
"Presentation": null,
"No... | [
[
{
"role": "PC",
"data": {
"comment": "This paper considers the benefits of deep multi-task RL with shared representations, by deriving bounds for multi-task approximate value and policy iteration bounds. This shows both theoretically and empirically that shared representations across multi... | [
"The submission derives bounds for approximate value and policy iteration for the multitask case in reinforcement learning.",
"In addition, two common RL algorithms are adapted to demonstrate benefits of multitask RL given related tasks.",
"The paper is mostly well written but sometimes introduces potentially u... | [
[
1
],
[
11
],
[
12
],
[
2
],
[
8
],
[
3
],
[
5
],
[
6
],
[
7
],
[
9
],
[
10
],
[
0
],
[
4
]
] | [
"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",
"data": [
0,
1
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
2
]
},
{
"role": "Author",
... | benchmark/PDF/ICLR2020_rkgpv2VFvr.pdf | openreview | benchmark/MD/ICLR2020_rkgpv2VFvr.md | ICLR 2020 |
SylUiREKvB | {
"title": "Variational Hyper RNN for Sequence Modeling",
"authors": [
"Ruizhi Deng",
"Yanshuai Cao",
"Bo Chang",
"Leonid Sigal",
"Greg Mori",
"Marcus Brubaker"
],
"authorids": [
"ruizhid@sfu.ca",
"yanshuaicao@gmail.com",
"bchang@stat.ubc.ca",
"lsigal@cs.ubc.ca",
"mor... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10
]
},... | [
[
{
"role": "PC",
"data": {
"comment": "The paper proposes a neural network architecture that uses a hypernetwork (RNN or feedforward) to generate weights for a network (variational RNN), that models sequential data. An empirical comparison of a large number of configurations on synthetic an... | [
"This paper proposes the variational hyper RNN (VHRNN), which extends the previous variational RNN (VRNN) by learning the parameters of RNN using a hyper RNN. VRHNN is tested and compared with VRNN on synthetic and real datasets. The authors report superior performance parameter efficiency over VRNN.",
"The perfo... | [
[
7,
24
],
[
8
],
[
16
],
[
11
],
[
17
],
[
23
],
[
10
],
[
2,
15
],
[
12
],
[
3
],
[
5
],
[
6
],
[
13
],
[
18
],
[
4
],
[
19
],
[
20
],
[
21
],
[
22
],
[
... | [
"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... | [
"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",
"data": [
0
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1
]
},
{
"role": "Author",
"data": [
... | benchmark/PDF/ICLR2020_SylUiREKvB.pdf | openreview | benchmark/MD/ICLR2020_SylUiREKvB.md | ICLR 2020 |
r1laNeBYPB | {
"authorids": [
"amirhosein.khasahmadi@mail.utoronto.ca",
"kaveh.hassani@autodesk.com",
"parsa.moradi73@gmail.com",
"ljlee@psi.toronto.edu",
"quaid.morris@utoronto.ca"
],
"title": "Memory-Based Graph Networks",
"authors": [
"Amir Hosein Khasahmadi",
"Kaveh Hassani",
"Parsa Morad... | 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": "PC",
"data": {
"comment": "Four reviewers have assessed this paper and they have scored it as 6/6/6/6 after rebuttal. Nonetheless, the reviewers have raised a number of criticisms and the authors are encouraged to resolve them for the camera-ready submission."
}
}
]
... | [
"The paper presents \"memory layer\" to simultaneously do graph representation learning and pooling in a hierarchical way. It shares the same spirit with the previous models (DiffPool and Mincut pooling) which cluster nodes and learn representation of the coarsened graph. In DiffPool, Graph convolutional Neural Net... | [
[
5
],
[
12
],
[
14
],
[
15
],
[
8
],
[
25
],
[
9
],
[
11
],
[
16,
21
],
[
26
],
[
19
],
[
28
],
[
6
],
[
7
],
[
23
],
[
13
],
[
1
],
[
3
],
[
4
],
[
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",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
... | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0,
1
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer",
"data": [
2
]
}
],
"category": [
"QUAL-REP"
... | benchmark/PDF/ICLR2020_r1laNeBYPB.pdf | openreview | benchmark/MD/ICLR2020_r1laNeBYPB.md | ICLR 2020 |
rklk_ySYPB | {
"authorids": [
"francesco91.croce@gmail.com",
"matthias.hein@uni-tuebingen.de"
],
"title": "Provable robustness against all adversarial $l_p$-perturbations for $p\\geq 1$",
"authors": [
"Francesco Croce",
"Matthias Hein"
],
"pdf": "/pdf/d268e4313086ad7910f2635de8c10331842748f4.pdf",
"TL;... | Accept (Poster) | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8
]
},
"scores": {
... | [
[
{
"role": "PC",
"data": {
"comment": "This paper extends the degree to which ReLU networks can be provably resistant to a broader class of adversarial attacks using a MMR-Universal regularization scheme. In particular, the first provably robust model in terms of lp norm perturbations is d... | [
"Overview:\nThe paper is dedicated to developing a regularization scheme for the provably robust model. The author proposes the MMR-Universal regularizer for ReLU based networks. It enforces l1 and l infinity robustness and leads to be provably robust with any lp norm attack for p larger than and equal to one.",
... | [
[
12
],
[
18
],
[
11
],
[
14
],
[
7
],
[
7
],
[
9
],
[
3
],
[
10
],
[
4
],
[
5
],
[
15
],
[
17
],
[
1
],
[
2
],
[
8
],
[
16
],
[
0
],
[
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",
"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": [
"QUAL-MET"... | benchmark/PDF/ICLR2020_rklk_ySYPB.pdf | openreview | benchmark/MD/ICLR2020_rklk_ySYPB.md | ICLR 2020 |
ryxB2lBtvH | {
"authorids": [
"lee504@usc.edu",
"jingyuny@usc.edu",
"limjj@usc.edu"
],
"title": "Learning to Coordinate Manipulation Skills via Skill Behavior Diversification",
"authors": [
"Youngwoon Lee",
"Jingyun Yang",
"Joseph J. Lim"
],
"pdf": "/pdf/b91b8720df5d00e5d36d1b9877ba04c392abb48a.p... | Accept (Poster) | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5
]
},
"scores": {
"Solid": null,
"Presentation": null... | [
[
{
"role": "PC",
"data": {
"comment": "This paper deals with multi-agent hierarchical reinforcement learning. A discrete set of pre-specified low-level skills are modulated by a conditioning vector and trained in a fashion reminiscent of Diversity Is All You Need, and then combined via a me... | [
"This paper aims to achieve multi-agent coordination by composing diverse skills learned by augmenting individual subtask objectives with DIAYN-style diversity bonuses. Once individual diverse skills are learned for the subtasks, the agents are combined by a meta-agent to coordinate multiple distinct robots to achi... | [
[
16
],
[
13
],
[
6
],
[
12
],
[
7
],
[
4
],
[
9
],
[
17
],
[
1
],
[
2
],
[
10
],
[
11
],
[
14
],
[
3
],
[
15
],
[
0
],
[
5
],
[
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"
] | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1,
3
]
},
{
"role": "Author",
... | benchmark/PDF/ICLR2020_ryxB2lBtvH.pdf | openreview | benchmark/MD/ICLR2020_ryxB2lBtvH.md | ICLR 2020 |
SkxJ8REYPH | {
"authorids": [
"jianyuw1@andrew.cmu.edu",
"tantia@fb.com",
"ballasn@fb.com",
"mikerabbat@fb.com"
],
"title": "SlowMo: Improving Communication-Efficient Distributed SGD with Slow Momentum",
"authors": [
"Jianyu Wang",
"Vinayak Tantia",
"Nicolas Ballas",
"Michael Rabbat"
],
"... | 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": "PC",
"data": {
"comment": "This paper presents a new approach, SlowMo, to improve communication-efficient distribution training with SGD. The main method is based on the BMUF approach and relies on workers to periodically synchronize and perform a momentum update. This works wel... | [
"The authors verify the effect of BMUF[1], which is called slow momentum in this paper, on computer vision and natural language processing tasks with different kinds of local optimizers. They also provided the theoretical convergence guarantee of BMUF.",
"The literature survey of this paper is quite good and the... | [
[
8
],
[
4
],
[
10
],
[
15
],
[
16
],
[
18
],
[
12
],
[
1,
13
],
[
2,
17
],
[
7
],
[
11
],
[
14
],
[
19
],
[
5
],
[
6
],
[
0
],
[
3
],
[
9
]
] | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"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": [
"QUAL-CMP",
"QUAL... | benchmark/PDF/ICLR2020_SkxJ8REYPH.pdf | openreview | benchmark/MD/ICLR2020_SkxJ8REYPH.md | ICLR 2020 |
S1eL4kBYwr | {
"title": "UNITER: Learning UNiversal Image-TExt Representations",
"authors": [
"Yen-Chun Chen",
"Linjie Li",
"Licheng Yu",
"Ahmed El Kholy",
"Faisal Ahmed",
"Zhe Gan",
"Yu Cheng",
"Jingjing Liu"
],
"authorids": [
"yen-chun.chen@microsoft.com",
"lindsey.li@microsoft.com"... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5
]
},
"scores": {
"Solid": null,
"Presentation": null... | [
[
{
"role": "PC",
"data": {
"comment": "This submission proposes an approach to pre-train general-purpose image and text representations that can be effective on target tasks requiring embeddings for both modes. The authors propose several pre-training tasks beyond masked language modelling ... | [
"This paper presents a novel method for image-text representations called UNITER. The proposed method has been subsequently tested in many downstream tasks. A detailed ablation study helps to understand the role of each pretrained task in the proposed model.",
"Although the empirical results are nice, performing ... | [
[
4
],
[
5
],
[
16
],
[
7
],
[
10
],
[
14,
15
],
[
17
],
[
23
],
[
1
],
[
11,
19
],
[
3,
13
],
[
20
],
[
0
],
[
8,
21,
22
],
[
9
],
[
18
],
[
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": [
"QUAL-EXP"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1
]
},
{
"role": "Author",
"data"... | benchmark/PDF/ICLR2020_S1eL4kBYwr.pdf | openreview | benchmark/MD/ICLR2020_S1eL4kBYwr.md | ICLR 2020 |
SyxS0T4tvS | {
"title": "RoBERTa: A Robustly Optimized BERT Pretraining Approach",
"authors": [
"Yinhan Liu",
"Myle Ott",
"Naman Goyal",
"Jingfei Du",
"Mandar Joshi",
"Danqi Chen",
"Omer Levy",
"Mike Lewis",
"Luke Zettlemoyer",
"Veselin Stoyanov"
],
"authorids": [
"yinhanliu@fb.co... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8
]
},
"scores": {
... | [
[
{
"role": "PC",
"data": {
"comment": "This paper conducts an extensive study of training BERT and shows that its performance can be improved significantly by choosing a better training setup (e.g., hyperparameters, objective functions). I think this paper clearly offers a better understand... | [
"This paper is a replication study of BERT for training large language models. Its main modification is simple: training longer with more data.",
"Significantly improvements have been reported, and the work achieves on-par or higher accuracy over a large set of downstream tasks compared to XLNet, which is a state... | [
[
7
],
[
9
],
[
10
],
[
3,
14
],
[
8
],
[
4
],
[
11
],
[
12
],
[
16
],
[
17
],
[
5
],
[
15
],
[
1,
2
],
[
0
],
[
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",
"correct",
"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
]
}
],
"category": [
"SIGN-SOT"... | benchmark/PDF/ICLR2020_SyxS0T4tvS.pdf | openreview | benchmark/MD/ICLR2020_SyxS0T4tvS.md | ICLR 2020 |
rkgMkCEtPB | {
"title": "Rapid Learning or Feature Reuse? Towards Understanding the Effectiveness of MAML",
"authors": [
"Aniruddh Raghu",
"Maithra Raghu",
"Samy Bengio",
"Oriol Vinyals"
],
"authorids": [
"aniruddhraghu@gmail.com",
"maithrar@gmail.com",
"bengio@google.com",
"vinyals@google.co... | Accept (Poster) | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7
]
},
"scores": {
"Solid": null,
... | [
[
{
"role": "PC",
"data": {
"comment": "Paper received mixed reviews: WR (R1), A (R2 and R3). AC has read reviews/rebuttal and examined paper. AC agrees that R1's concerns are misplaced and feels the paper should be accepted. \n"
}
}
]
] | [
"The paper claims to examine the reasons for the success of MAML---an influential meta-learning algorithm to tackle few-shot learning. It thoroughly investigated the importance of the two optimization loops, and found that feature reuse is the dominant factor for MAML’s success.",
"Moreover, the authors proposed ... | [
[
4
],
[
6
],
[
9
],
[
7
],
[
8
],
[
2
],
[
11
],
[
3
],
[
0
],
[
1
],
[
5
],
[
10
]
] | [
"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
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer",
"data": [
2
]
}
],
"category": [
"N/A"
]
... | benchmark/PDF/ICLR2020_rkgMkCEtPB.pdf | openreview | benchmark/MD/ICLR2020_rkgMkCEtPB.md | ICLR 2020 |
rkxawlHKDr | {
"authorids": [
"shiretzet@gmail.com",
"shaharabany@mail.tau.ac.il",
"wolf@fb.com"
],
"title": "End to End Trainable Active Contours via Differentiable Rendering",
"authors": [
"Shir Gur",
"Tal Shaharabany",
"Lior Wolf"
],
"pdf": "/pdf/8fd896acc91bd1f06a713396633ab37e62e6dd46.pdf",
... | 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": "PC",
"data": {
"comment": "The submission presents a differentiable take on classic active contour methods, which used to be popular in computer vision. The method is sensible and the results are strong. After the revision, all reviewers recommend accepting the paper."
}
... | [
"EDIT: The rating changed from '1: Reject' to '6: Weak accept' after the rebuttal. See below for my reasoning.",
"The submission considers two-class image segmentation problems, where a closed-contour image region is to be specified as the 'object'/region of interest, vs. 'no-object'/background. The approach take... | [
[
8,
21
],
[
12
],
[
15
],
[
29
],
[
4
],
[
30
],
[
1,
5
],
[
3,
19
],
[
10
],
[
13
],
[
14
],
[
16,
18,
25
],
[
27
],
[
28
],
[
24
],
[
6
],
[
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",
"cor... | [
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"incorrect",
"correct",
"correct",
... | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0,
1
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
2
]
},
{
"role": "Author",
... | benchmark/PDF/ICLR2020_rkxawlHKDr.pdf | openreview | benchmark/MD/ICLR2020_rkxawlHKDr.md | ICLR 2020 |
S1xO4xHFvB | {
"title": "Atomic Compression Networks",
"authors": [
"Jonas Falkner",
"Josif Grabocka",
"Lars Schmidt-Thieme"
],
"authorids": [
"falkner@ismll.uni-hildesheim.de",
"josif@ismll.uni-hildesheim.de",
"schmidt-thieme@ismll.uni-hildesheim.de"
],
"keywords": [
"Network Compression"
... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10
]
},... | [
[
{
"role": "PC",
"data": {
"comment": "This paper proposed a very general idea called Atomic Compression Networks (ACNs) to construct neural networks. The idea looks simple and effective. However, the reason why it works is not well explained. The experiments are not sufficient enough to ... | [
"This paper explores the use of replicating neurons across and within layers to compress fully connected neural networks. The idea is simple, and is evaluated on a number of datasets and compared with fully connected, single layer, and several compression schemes.",
"Strengths: a lot of nice experiments with clea... | [
[
7
],
[
10
],
[
12
],
[
5
],
[
2
],
[
13
],
[
1
],
[
3
],
[
8
],
[
14
],
[
11
],
[
6
],
[
0
],
[
4
],
[
9
]
] | [
"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",
"correct",
"correct",
"incorrect"
] | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1
]
}
],
"category": [
"QUAL-EXP"
]
},
... | benchmark/PDF/ICLR2020_S1xO4xHFvB.pdf | openreview | benchmark/MD/ICLR2020_S1xO4xHFvB.md | ICLR 2020 |
ryl5CJSFPS | {
"authorids": [
"sametoymak@gmail.com",
"zfabian@usc.edu",
"mli176@ucr.edu",
"msoltoon@gmail.com"
],
"title": "GENERALIZATION GUARANTEES FOR NEURAL NETS VIA HARNESSING THE LOW-RANKNESS OF JACOBIAN",
"authors": [
"Samet Oymak",
"Zalan Fabian",
"Mingchen Li",
"Mahdi Soltanolkotabi... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9
]
},
"scores... | [
[
{
"role": "PC",
"data": {
"comment": "This submission investigates the properties of the Jacobian matrix in deep learning setup. Specifically, it splits the spectrum of the matrix into information (large singulars) and ``nuisance (small singulars) spaces. The paper shows that over the info... | [
"Note: The template used in this paper is of ICLR 2019, not ICLR 2020.",
"This paper identifies the information space and nuisance space by thresholding the singular values of the network's jacobian and shows that generally the residuals projected to the information space can be effectively optimized to zero, thu... | [
[
5
],
[
9
],
[
14
],
[
6
],
[
19,
22
],
[
2
],
[
3
],
[
4
],
[
8
],
[
7
],
[
0
],
[
17
],
[
20
],
[
11
],
[
12
],
[
13
],
[
15
],
[
18
],
[
24
],
[
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",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"c... | [
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
0
]
},
{
"role": "Author",
"data": [
11
]
}
],
"category": [
"POL-IMP"
]
},
{
"sentences": [
{
"role": "Reviewer",
"data"... | benchmark/PDF/ICLR2020_ryl5CJSFPS.pdf | openreview | benchmark/MD/ICLR2020_ryl5CJSFPS.md | ICLR 2020 |
SJgMK64Ywr | {
"title": "AssembleNet: Searching for Multi-Stream Neural Connectivity in Video Architectures",
"authors": [
"Michael S. Ryoo",
"AJ Piergiovanni",
"Mingxing Tan",
"Anelia Angelova"
],
"authorids": [
"mryoo@google.com",
"ajpiergi@indiana.edu",
"tanmingxing@google.com",
"anelia@go... | Accept (Poster) | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6
]
},
"scores": {
"Solid": null,
"Pres... | [
[
{
"role": "PC",
"data": {
"comment": "The submission applies architecture search to find effective architectures for video classification. The work is not terribly innovative, but the results are good. All reviewers recommend accepting the paper."
}
},
{
"role": "Author... | [
"This submission proposes a way to do multi-stream neural architecture search for video classification. I give an initial rating of accept because (1) there are not many work on video architecture search yet (2) the paper is well written (3) experiments are complete and results are strong. I have a few comments as ... | [
[
0
],
[
1
],
[
6,
13
],
[
9
],
[
14
],
[
7,
12
],
[
11
],
[
16
],
[
2
],
[
4
],
[
8
],
[
15
],
[
3
],
[
5
],
[
10
],
[
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",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct"
] | [
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
0
]
}
],
"category": [
"CLAR-WRT",
"QUAL-EXP",
"SIGN-BRD"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1
]
},
{
... | benchmark/PDF/ICLR2020_SJgMK64Ywr.pdf | openreview | benchmark/MD/ICLR2020_SJgMK64Ywr.md | ICLR 2020 |
SylOlp4FvH | {
"title": "V-MPO: On-Policy Maximum a Posteriori Policy Optimization for Discrete and Continuous Control",
"authors": [
"H. Francis Song",
"Abbas Abdolmaleki",
"Jost Tobias Springenberg",
"Aidan Clark",
"Hubert Soyer",
"Jack W. Rae",
"Seb Noury",
"Arun Ahuja",
"Siqi Liu",
"D... | 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": "PC",
"data": {
"comment": "This paper proposes an extension of MPO for on-policy reinforcement learning. The proposed method achieved promising results in a relatively hyper-parameter insensitive manner.\n\nOne concern of the reviewers is the lack of comparison with previous wor... | [
"Summary: This paper presents the V-MPO algorithm for on-policy reinforcement learning that can handle both continuous/discrete control, single/multi-task learning and use both low dimensional states and pixels.",
"V-MPO adapts MPO, a recent off-policy deep reinforcement learning algorithm, to the on-policy setti... | [
[
9
],
[
33
],
[
17
],
[
24
],
[
25
],
[
26
],
[
30
],
[
2
],
[
4
],
[
10
],
[
14
],
[
23
],
[
29
],
[
6
],
[
15
],
[
20
],
[
28
],
[
1
],
[
3
],
[
5,
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",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"c... | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0,
1,
2
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
3
]
},
{
"role": "Au... | benchmark/PDF/ICLR2020_SylOlp4FvH.pdf | openreview | benchmark/MD/ICLR2020_SylOlp4FvH.md | ICLR 2020 |
HyxG3p4twS | {
"title": "Quantifying the Cost of Reliable Photo Authentication via High-Performance Learned Lossy Representations",
"authors": [
"Pawel Korus",
"Nasir Memon"
],
"authorids": [
"pkorus@nyu.edu",
"memon@nyu.edu"
],
"keywords": [
"image forensics",
"photo manipulation detection",
... | Accept (Poster) | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4
]
},
"scores": {
"Solid": null,
"Presentation": null,
"No... | [
[
{
"role": "PC",
"data": {
"comment": "The paper introduces a new image compression approach that preserves the patterns indicating image manipulation. The reviewers appreciate the idea and the method. Please take into account the suggestions of Reviewer1, when preparing the final version."... | [
"Summary of the paper",
"- This work proposes a new deep-learning-based method to replace the lossy compression techniques of images., jpg.",
"- The work investigates the role of codec and shows that the proposed complex photo dissemination channels optimizes the codec related traits on images.",
"- The metho... | [
[
21,
22
],
[
3
],
[
18
],
[
20
],
[
6
],
[
15
],
[
19
],
[
23
],
[
14
],
[
24
],
[
8
],
[
12
],
[
13
],
[
16
],
[
10
],
[
2
],
[
4
],
[
9
],
[
5,
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",
"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": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer",
"data": [
2
]
}
],
"category": [
"N/A"
]
... | benchmark/PDF/ICLR2020_HyxG3p4twS.pdf | openreview | benchmark/MD/ICLR2020_HyxG3p4twS.md | ICLR 2020 |
rkg6FgrtPB | {
"authorids": [
"sruthi@comp.nus.edu.sg",
"anandl@iisc.ac.in",
"christos@columbia.edu",
"vempala@gatech.edu",
"y.naganand@gmail.com"
],
"title": "Biologically Plausible Neural Networks via Evolutionary Dynamics and Dopaminergic Plasticity",
"authors": [
"Sruthi Gorantla",
"Anand Lou... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
... | [
[
{
"role": "PC",
"data": {
"comment": "Unfortunately the paper is confusingly written, and there is only agreement by all reviewers on the rejection of the paper. Indeed, if all reviewers and the area chair do not interpret the paper well, the authors' best response would be to rewrite the... | [
"First of all, I must confess that my knowledge is quite limited to read this paper. Perhap the authors present something that I can not catch up at the present.",
"I conjecture the paper would like to bring the evolution in genetics and perhap brain cirecuits as well to define a novel neural net model, called NN... | [
[
9
],
[
10
],
[
11
],
[
12
],
[
13
],
[
15
],
[
16
],
[
17
],
[
2
],
[
3
],
[
4
],
[
5
],
[
6
],
[
7
],
[
8
],
[
14
],
[
25
],
[
22,
24
],
[
20
],
[
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",
"cor... | [
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"c... | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer",
"data": [
1
]
}
],
"category": [
"N/A"
]
},
{
... | benchmark/PDF/ICLR2020_rkg6FgrtPB.pdf | openreview | benchmark/MD/ICLR2020_rkg6FgrtPB.md | ICLR 2020 |
S1g7tpEYDS | {
"authorids": [
"partha.ghosh@tuebingen.mpg.de",
"msajjadi@tue.mpg.de",
"antonio.vergari@tuebingen.mpg.de",
"black@tue.mpg.de",
"bs@tue.mpg.de"
],
"title": "From Variational to Deterministic Autoencoders",
"authors": [
"Partha Ghosh",
"Mehdi S. M. Sajjadi",
"Antonio Vergari",
... | 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": "PC",
"data": {
"comment": "This paper proposes an extension to deterministic autoencoders, namely instead of noise injection in the encoders of VAEs to use deterministic autoencoders with an explicit regularization term on the latent representations. While the reviewers agree th... | [
"This paper propose an extension to deterministic autoencoders. Motivated from VAEs, the authors propose RAEs, which replace the noise injection in the encoders of VAEs with an explicit regularization term on the latent representations.",
"As a result, the model becomes a deterministic autoencoder with a L_2 regu... | [
[
2
],
[
1
],
[
4
],
[
5
],
[
6
],
[
18
],
[
3
],
[
7,
16
],
[
9
],
[
13
],
[
19
],
[
20
],
[
24
],
[
25
],
[
26
],
[
27
],
[
28
],
[
10,
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",
"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,
2
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
3
]
}
],
"category": [
... | benchmark/PDF/ICLR2020_S1g7tpEYDS.pdf | openreview | benchmark/MD/ICLR2020_S1g7tpEYDS.md | ICLR 2020 |
SylzhkBtDB | {
"authorids": [
"senwu@cs.stanford.edu",
"hongyang@cs.stanford.edu",
"chrismre@stanford.edu"
],
"title": "Understanding and Improving Information Transfer in Multi-Task Learning",
"authors": [
"Sen Wu",
"Hongyang R. Zhang",
"Christopher Ré"
],
"pdf": "/pdf/d5f06dad9eae7e9312f393d6bd... | Accept (Poster) | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6
]
},
"scores": {
"Solid": null,
"Pres... | [
[
{
"role": "PC",
"data": {
"comment": "Many existing approaches in multi-task learning rely on intuitions about how to transfer information. This paper, instead, tries to answer what does \"information transfer\" even mean in this context. Such ideas have already been presented in the past,... | [
"This paper studies how to improve the multi-task learning from both theoretical and experimental viewpoints. More specifically, they study an architecture where there is a shared model for all of the tasks and a separate module specific to each task. They show that data similarity of the tasks, measured by task co... | [
[
20
],
[
5
],
[
12
],
[
25
],
[
28
],
[
29
],
[
1
],
[
7
],
[
10
],
[
19
],
[
21
],
[
22
],
[
26
],
[
2
],
[
4
],
[
11
],
[
15
],
[
13
],
[
14
],
[
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",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"c... | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0,
1
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
2
]
}
],
"category": [
"CLAR-WRT"... | benchmark/PDF/ICLR2020_SylzhkBtDB.pdf | openreview | benchmark/MD/ICLR2020_SylzhkBtDB.md | ICLR 2020 |
rJgqMRVYvr | {
"authorids": [
"jwl3@andrew.cmu.edu",
"khodak@cs.cmu.edu",
"scaldas@cs.cmu.edu",
"talwalkar@cmu.edu"
],
"title": "Differentially Private Meta-Learning",
"authors": [
"Jeffrey Li",
"Mikhail Khodak",
"Sebastian Caldas",
"Ameet Talwalkar"
],
"pdf": "/pdf/d15549acf0d539a18c202c... | Accept (Poster) | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3
]
},
"scores": {
"Solid": null,
"Presentation": null,
"Novelty": null,
... | [
[
{
"role": "PC",
"data": {
"comment": "Thanks to the authors for the submission. This paper studies differentially private meta-learning, where the algorithm needs to use information across several learning tasks to protect the privacy of the data set from each task. The reviewers agree tha... | [
"This paper proposes the notions of different privacy levels for different attack models, namely global and local meta-level and within-task level privacy for meta-learning. It proposes an algorithm for global within-task privacy. It provides privacy and utility guarantee of the proposed algorithm and experimental ... | [
[
1
],
[
14
],
[
13
],
[
5,
9
],
[
2
],
[
3,
12
],
[
7
],
[
8
],
[
6
],
[
10
],
[
0
],
[
4
],
[
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"
] | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1
]
},
{
"role": "Author",
"data": [
... | benchmark/PDF/ICLR2020_rJgqMRVYvr.pdf | openreview | benchmark/MD/ICLR2020_rJgqMRVYvr.md | ICLR 2020 |
Syxi6grFwH | {
"authorids": [
"dc18393@bristol.ac.uk",
"rui.costa@bristol.ac.uk"
],
"title": "HIPPOCAMPAL NEURONAL REPRESENTATIONS IN CONTINUAL LEARNING",
"authors": [
"Samia Mohinta",
"Rui Ponte Costa",
"Stephane Ciocchi"
],
"abstract": "The hippocampus has long been associated with spatial memory 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": "PC",
"data": {
"comment": "This paper analyzes neural recording data taken from rodents performing a continual learning task using demixed principal component analysis, and aims to find representations for behaviorally relevant variables. They compare these features with those o... | [
"This paper analyses a dataset of representations in the CA1 region of the hippocampus of a rat conducting a spatial plus maze task that switches between allocentric and egocentric versions. In the allocentric version of the task, the rat must always go from north or south arms to the west arm to receive a reward. ... | [
[
13
],
[
19
],
[
20
],
[
15
],
[
16
],
[
17
],
[
21
],
[
28
],
[
6
],
[
35
],
[
3,
33
],
[
5
],
[
9,
37
],
[
22
],
[
23
],
[
24
],
[
25
],
[
36
],
[
38
],
... | [
"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": [
"SIGN-BRD"... | benchmark/PDF/ICLR2020_Syxi6grFwH.pdf | openreview | benchmark/MD/ICLR2020_Syxi6grFwH.md | ICLR 2020 |
r1lGO0EKDH | {
"authorids": [
"cd574@cornell.edu",
"qzzhao@mtu.edu",
"yongyuw@mtu.edu",
"zhiruz@cornell.edu",
"zfeng12@stevens.edu"
],
"title": "GraphZoom: A Multi-level Spectral Approach for Accurate and Scalable Graph Embedding",
"authors": [
"Chenhui Deng",
"Zhiqiang Zhao",
"Yongyu Wang",
... | Accept (Talk) | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8
]
},
"scores": {
... | [
[
{
"role": "PC",
"data": {
"comment": "The authors present an approach for learning graph embeddings by first fusing the graph to generate a new graph with encodes structural information as well as node attribution information. They then iteratively merge nodes based spectral similarities t... | [
"Summary: The authors propose a way to fuse information on nodes of a graph with the topology of the graph in the large scale setting. The proposed approach is done in four phases where (i) the covariates in the nodes of the graph is first mapped in the graph space for fusion and fused using linear combination of t... | [
[
2
],
[
14
],
[
6,
13
],
[
1
],
[
7
],
[
8,
15
],
[
16
],
[
27
],
[
12
],
[
22
],
[
4,
25
],
[
5
],
[
9
],
[
20
],
[
21
],
[
23
],
[
26,
33
],
[
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",
"incorrect",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"incorrect",
"incorrect",
"correct",
"correct",
"correct",
"incorrect",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"co... | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1
]
},
{
"role": "Author",
"data": [
... | benchmark/PDF/ICLR2020_r1lGO0EKDH.pdf | openreview | benchmark/MD/ICLR2020_r1lGO0EKDH.md | ICLR 2020 |
SJeQi1HKDH | {
"authorids": [
"sh018@ie.cuhk.edu.hk",
"doubledaibo@gmail.com",
"sunjiankai@sensetime.com",
"pengzh@ie.cuhk.edu.hk",
"xg018@ie.cuhk.edu.hk",
"dhlin@ie.cuhk.edu.hk",
"bzhou@ie.cuhk.edu.hk"
],
"title": "Learning with Social Influence through Interior Policy Differentiation",
"author... | Reject | [
[
{
"role": "Author",
"data": {
"value": {
"comment": [
0,
1
]
}
}
}
],
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
2,
... | [
[
{
"role": "PC",
"data": {
"comment": "The paper proposes a mechanism for obtaining diverse policies for solving a task by posing it as a multi-agent problem, and incentivizing the agents to be different from each other via maximizing total variation.\n\nThe reviewers agreed that this is an... | [
"Dear reviewers and general audience,",
"Please download the demo video following this dropbox link: https://www.dropbox.com/s/rrs08dicidcim2l/ICLR20.mp4",
"This paper proposes a new method for learning diverse policies in RL environments, with the ultimate goal of increasing reward. The paper develops a novel ... | [
[
26
],
[
37
],
[
1
],
[
3
],
[
10
],
[
14
],
[
17
],
[
23
],
[
24
],
[
27
],
[
28
],
[
29
],
[
33
],
[
38
],
[
31
],
[
4
],
[
12
],
[
21
],
[
6
],
[
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",
"incorrect",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
... | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
2
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 2",
"data": [
3
]
}
],
"category": [
"CLAR-WRT",
"QUAL... | benchmark/PDF/ICLR2020_SJeQi1HKDH.pdf | openreview | benchmark/MD/ICLR2020_SJeQi1HKDH.md | ICLR 2020 |
Sye_OgHFwH | {
"authorids": [
"bhattad2@illinois.edu",
"mchong6@illinois.edu",
"kl2@illinois.edu",
"lbo@illinois.edu",
"daf@illinois.edu"
],
"title": "Unrestricted Adversarial Examples via Semantic Manipulation",
"authors": [
"Anand Bhattad",
"Min Jin Chong",
"Kaizhao Liang",
"Bo Li",
... | Accept (Poster) | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4
]
},
"scores": {
"Solid": null,
"Presentation": null,
"No... | [
[
{
"role": "PC",
"data": {
"comment": "In this paper, the authors present adversarial attacks by semantic manipulations, i.e., manipulating specific detectors that result in imperceptible changes in the picture, such as changing texture and color, but without affecting their naturalness. Mo... | [
"The paper proposes cAdv and sAdv, two new unrestricted adversarial attack methods that manipulates either color or texture of an image. To these end, the paper employes another parametrized colorization techniques (and texture transfer method) and proposes optimization objectives for finding adversarial examples w... | [
[
4
],
[
6,
11
],
[
2
],
[
3
],
[
7
],
[
8
],
[
1
],
[
10
],
[
0
],
[
5
],
[
9
]
] | [
"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
]
}
],
"category": [
"QUAL-MET",
"QUAL... | benchmark/PDF/ICLR2020_Sye_OgHFwH.pdf | openreview | benchmark/MD/ICLR2020_Sye_OgHFwH.md | ICLR 2020 |
ryljMpNtwr | {
"title": "Benchmarking Robustness in Object Detection: Autonomous Driving when Winter is Coming",
"authors": [
"Claudio Michaelis",
"Benjamin Mitzkus",
"Robert Geirhos",
"Evgenia Rusak",
"Oliver Bringmann",
"Alexander S. Ecker",
"Matthias Bethge",
"Wieland Brendel"
],
"authorid... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
... | [
[
{
"role": "PC",
"data": {
"comment": "This paper proposes a benchmark for assessing the impact of image quality degradation (e.g. simulated fog, snow, frost) on the performance of object detection models. The authors introduce corrupted versions of popular object detection datasets, namely... | [
"Summary:",
"- key problem or question: assessing / improving the robustness of object detectors to image corruptions (simulated fog, frost, snow, dragonfire...);",
"- contributions: 1) a benchmark (obtained by adding image-level corruptions to PASCAL, COCO, and Cityscapes) and an experimental protocol to measu... | [
[
25
],
[
2,
5
],
[
3
],
[
4
],
[
18
],
[
10
],
[
26
],
[
6
],
[
7
],
[
8
],
[
9
],
[
12
],
[
13
],
[
14
],
[
15
],
[
16
],
[
21
],
[
22
],
[
23
],
[
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",
"incorrect",
"incorrect",
"incorrect",
"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
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer",
"data": [
2
]
}
],
"category": [
"N/A"
]
... | benchmark/PDF/ICLR2020_ryljMpNtwr.pdf | openreview | benchmark/MD/ICLR2020_ryljMpNtwr.md | ICLR 2020 |
SkgbmyHFDS | {
"authorids": [
"zeyu@umich.edu",
"junhyuk@google.com",
"mtthss@google.com",
"zhongwen@google.com",
"makro@google.com",
"hado@google.com",
"davidsilver@google.com",
"baveja@google.com"
],
"title": "What Can Learned Intrinsic Rewards Capture?",
"authors": [
"Zeyu Zheng",
... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
... | [
[
{
"role": "PC",
"data": {
"comment": "The authors present a metalearning-based approach to learning intrinsic rewards that improve RL performance across distributions of problems. This is essentially a more computationally efficient approach to approaches suggested by Singh (2009/10). Th... | [
"Summary\nThe paper evaluates the intrinsic reward as a way of storing information about episodes. It adopts the optimal intrinsic reward setting (Singh'09), and extends its recent policy gradient implementation, LIRPG, to lifetime settings. The task in the lifetime setting is to learn an intrinsic reward such that... | [
[
36
],
[
29
],
[
2
],
[
3
],
[
38
],
[
32
],
[
1,
5
],
[
8
],
[
20
],
[
25
],
[
9,
16,
43
],
[
11
],
[
12
],
[
17
],
[
34
],
[
4
],
[
6
],
[
10
],
[
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",
"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/ICLR2020_SkgbmyHFDS.pdf | openreview | benchmark/MD/ICLR2020_SkgbmyHFDS.md | ICLR 2020 |
SJeq9JBFvH | {
"title": "Deep probabilistic subsampling for task-adaptive compressed sensing",
"authors": [
"Iris A.M. Huijben",
"Bastiaan S. Veeling",
"Ruud J.G. van Sloun"
],
"authorids": [
"i.a.m.huijben@tue.nl",
"basveeling@gmail.com",
"r.j.g.v.sloun@tue.nl"
],
"keywords": [],
"abstract": "... | Accept (Poster) | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7
]
},
"scores": {
"Solid": null,
... | [
[
{
"role": "PC",
"data": {
"comment": "This paper introduces a probabilistic data subsampling scheme that can be optimized end-to-end. The experimental evaluation is a bit weak, focusing mostly on toy-scale problems, and I would have liked to see a discussion of bias in the Gumbel-max grad... | [
"The paper proposes a learning-based adaptive compressed sensing framework in which both the sampling and the task functions (e.g., classification) are learned jointly end-to-end. The main contribution includes using the Gumbel-softmax trick to relax categorical distributions and use back-propagation to estimate th... | [
[
13
],
[
14
],
[
15
],
[
17
],
[
19,
30
],
[
4
],
[
1
],
[
2
],
[
3
],
[
22
],
[
26
],
[
8
],
[
21
],
[
28
],
[
31
],
[
6
],
[
7
],
[
12
],
[
16,
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",
"incorrect",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"incorrect",
"correct"... | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0,
1,
2
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
3
]
},
{
"role": "Au... | benchmark/PDF/ICLR2020_SJeq9JBFvH.pdf | openreview | benchmark/MD/ICLR2020_SJeq9JBFvH.md | ICLR 2020 |
SJlNnhVYDr | {
"abstract": "We propose a model to tackle classification tasks in the presence of very little training data. To this aim, we introduce a novel matching mechanism to focus on elements of the input by using vectors that represent semantically meaningful concepts for the task at hand.\nBy leveraging highlighted portio... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
... | [
[
{
"role": "PC",
"data": {
"comment": "The authors focus on low-resource text classifications tasks augmented with \"rationales\". They propose a new technique that improves performance over existing approaches and that allows human inspection of the learned weights.\n\nAlthough the reviewe... | [
"After responses:",
"I read the authors response and decided to stick to my original score mostly because:",
"1 - I understand that interpretability is hard to define. I also agree with the authors response.",
"However, this is still not reflected in the paper in any way. I expect a discussion on what is the ... | [
[
1
],
[
6,
12
],
[
16
],
[
18
],
[
17
],
[
2
],
[
3
],
[
7
],
[
9
],
[
14
],
[
15
],
[
8
],
[
5
],
[
11
],
[
0
],
[
4
],
[
10
],
[
13
]
] | [
"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",
"incorrect",
"correct",
"incorrect",
"correct"
] | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0,
1
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
2,
3
]
},
{
"role": "Au... | benchmark/PDF/ICLR2020_SJlNnhVYDr.pdf | openreview | benchmark/MD/ICLR2020_SJlNnhVYDr.md | ICLR 2020 |
HyxnMyBKwB | {
"title": "The Gambler's Problem and Beyond",
"authors": [
"Baoxiang Wang",
"Shuai Li",
"Jiajin Li",
"Siu On Chan"
],
"authorids": [
"bxwang@cse.cuhk.edu.hk",
"shuaili8@sjtu.edu.cn",
"jjli@se.cuhk.edu.hk",
"siuon@cse.cuhk.edu.hk"
],
"keywords": [
"the gambler's problem",... | Accept (Poster) | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5
]
},
"scores": {
"Solid": null,
"Presentation": null... | [
[
{
"role": "PC",
"data": {
"comment": "This paper studies the optimal value function for the gambler's problem, and presents some interesting characterizations thereof. The paper is well written and should be accepted."
}
},
{
"role": "Unknown",
"data": {
"... | [
"The paper revisits the Gambler's problem. It studies a generalized formulation with continuous state and action space and shows that the optimal value function is self-similar, fractal and non-rectifiable.",
"That is, it cannot be described by any simple analytic formula. Based on this, it also deeply analysis t... | [
[
9
],
[
7
],
[
1
],
[
5
],
[
8
],
[
11
],
[
12
],
[
13
],
[
14
],
[
15,
16,
17,
18,
19,
20,
21,
22,
23,
24,
25,
27
],
[
26
],
[
30
],
[
29
],
[
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",
"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/ICLR2020_HyxnMyBKwB.pdf | openreview | benchmark/MD/ICLR2020_HyxnMyBKwB.md | ICLR 2020 |
S1xKd24twB | {
"title": "SQIL: Imitation Learning via Reinforcement Learning with Sparse Rewards",
"authors": [
"Siddharth Reddy",
"Anca D. Dragan",
"Sergey Levine"
],
"authorids": [
"sgr@berkeley.edu",
"anca@berkeley.edu",
"svlevine@eecs.berkeley.edu"
],
"keywords": [
"Imitation Learning",
... | 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": "PC",
"data": {
"comment": "The authors present a simple alternative to adversarial imitation learning methods like GAIL that is potentially less brittle, and can skip learning a reward function, instead learning an imitation policy directly. Their method has a close relationshi... | [
"Summary\nThe authors propose SQUIL, an off-policy imitation learning (IL) algorithm which attempts to overcome the classic drift problems of behavioral cloning (BC). The idea is to reduce IL to a standard RL problem with a reward that incentivizes the agent to take expert actions in states observed in the demonstr... | [
[
10
],
[
11
],
[
7
],
[
9
],
[
18
],
[
19
],
[
20
],
[
21
],
[
22
],
[
23
],
[
25
],
[
2
],
[
8
],
[
27
],
[
12
],
[
16
],
[
6
],
[
14
],
[
15
],
[
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",
"correct",
"correct",
"correct",
"incorrect",
"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
]
}
],
"category": [
"QUAL-MET",
"CLAR... | benchmark/PDF/ICLR2020_S1xKd24twB.pdf | openreview | benchmark/MD/ICLR2020_S1xKd24twB.md | ICLR 2020 |
Skl3SkSKDr | {
"title": "Generating valid Euclidean distance matrices",
"authors": [
"Moritz Hoffmann",
"Frank Noe"
],
"authorids": [
"moritz.hoffmann@fu-berlin.de",
"frank.noe@fu-berlin.de"
],
"keywords": [
"euclidean distance matrices",
"wgan",
"point clouds",
"molecular structures"
]... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0
]
},
"scores": {
"Solid": null,
"Presentation": null,
"Novelty": null,
"Overall": 8,
"Confidence"... | [
[
{
"role": "PC",
"data": {
"comment": "This paper proposes a parametrisation of Euclidean distance matrices amenable to be used within a differentiable generative model. The resulting model is used in a WGAN architecture and demonstrated empirically in the generation of molecular structures... | [
"The paper is extremely interesting, solid and very well written. The idea is simple but nonetheless developed in a smart and effective fashion. The underlying theory is solid, even if some choices should have been discussed more deeply (e.g. the chosen loss function). Introduction and references are adequate, and ... | [
[
4
],
[
25
],
[
0
],
[
8
],
[
10,
12,
14,
15,
16
],
[
22
],
[
23
],
[
24
],
[
26
],
[
27
],
[
29
],
[
2
],
[
5
],
[
17
],
[
19
],
[
28
],
[
3
],
[
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
]
},
{
"role": "Author",
"data": [
2,
3,
4
]
}
],
"category": [
"CLAR-WRT",
"QUAL-EXP",
"QUAL-CMP"
]
},
{
"se... | benchmark/PDF/ICLR2020_Skl3SkSKDr.pdf | openreview | benchmark/MD/ICLR2020_Skl3SkSKDr.md | ICLR 2020 |
rJxFpp4Fvr | {
"title": "Feature-Robustness, Flatness and Generalization Error for Deep Neural Networks",
"authors": [
"Henning Petzka",
"Linara Adilova",
"Michael Kamp",
"Cristian Sminchisescu"
],
"authorids": [
"henning.petzka@gmail.com",
"adylova.linara.r@gmail.com",
"info@michaelkamp.org",
... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
... | [
[
{
"role": "PC",
"data": {
"comment": "The authors propose a notion of feature robustness, provide a straightforward decomposition of risk in terms of this robustness measure, and then provide some empirical evidence for their perspective. Across the board, the reviewers raised issues with ... | [
"This paper proposes a flatness measure that is invariant to layer-wise reparametrizations in ReLU networks. The notion of feature robustness, which is a notion the paper proposes, connects the flatness measure to generalization error.",
"This paper should be rejected because it is not well-placed in the literatu... | [
[
9
],
[
6
],
[
16
],
[
18
],
[
0
],
[
1
],
[
2
],
[
4
],
[
17
],
[
20,
25,
25,
28,
28
],
[
8
],
[
10
],
[
13
],
[
22,
26,
30
],
[
23
],
[
34
],
[
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",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"c... | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0
]
}
],
"category": [
"ORIG-MTH"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1
]
},
{
"role": "Author",
"data"... | benchmark/PDF/ICLR2020_rJxFpp4Fvr.pdf | openreview | benchmark/MD/ICLR2020_rJxFpp4Fvr.md | ICLR 2020 |
rylXBkrYDS | {
"authorids": [
"guneetdhillon@utexas.edu",
"pratikac@seas.upenn.edu",
"avinash.a.ravichandran@gmail.com",
"soattos@amazon.com"
],
"title": "A Baseline for Few-Shot Image Classification",
"authors": [
"Guneet Singh Dhillon",
"Pratik Chaudhari",
"Avinash Ravichandran",
"Stefano S... | Accept (Poster) | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6
]
},
"scores": {
"Solid": null,
"Pres... | [
[
{
"role": "PC",
"data": {
"comment": "This paper introduces a simple baseline for few-shot image classification in the transductive setting, which includes a standard cross-entropy loss on the labeled support samples and a conditional entropy loss on the unlabeled query samples.\n\nBoth lo... | [
"This paper provided a baseline method for few-shot learning. It utilizes a simple but effective approach via a transductive fine-tuning. The experimental results on several benchmarks show the improvements over state-of-the-art approaches.",
"It is a comprehensive study of the methods and datasets in this domain... | [
[
17
],
[
2
],
[
9
],
[
15
],
[
3
],
[
10
],
[
4
],
[
11
],
[
18
],
[
7
],
[
14
],
[
1
],
[
6
],
[
8
],
[
12
],
[
13,
16
],
[
0
],
[
5
]
] | [
"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",
"incorrect",
"incorrect",
"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": [
"QUAL-EXP"
]
},
... | benchmark/PDF/ICLR2020_rylXBkrYDS.pdf | openreview | benchmark/MD/ICLR2020_rylXBkrYDS.md | ICLR 2020 |
rkeJRhNYDH | {
"title": "TabFact: A Large-scale Dataset for Table-based Fact Verification",
"authors": [
"Wenhu Chen",
"Hongmin Wang",
"Jianshu Chen",
"Yunkai Zhang",
"Hong Wang",
"Shiyang Li",
"Xiyou Zhou",
"William Yang Wang"
],
"authorids": [
"wenhuchen@ucsb.edu",
"hongmin@ucsb.edu... | 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": "PC",
"data": {
"comment": "This paper presents a new dataset for fact verification in text from tables. The task is to identify whether a given claim is supported by the information presented in the table. The authors have also presented two baseline models, one based on BERT an... | [
"This paper proposes a new dataset for table-based fact verification and introduces a couple of methods for the task. I think that the dataset would be a useful resource (see some comments nevertheless on its construction), however the methods proposed are not particularly interesting, and the contributions to ML a... | [
[
18
],
[
1
],
[
6
],
[
15
],
[
16
],
[
17
],
[
29
],
[
30
],
[
31
],
[
9
],
[
10
],
[
0
],
[
4
],
[
5
],
[
25
],
[
3,
24
],
[
11
],
[
12
],
[
13
],
[
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",
"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
]
},
{
"role": "Author",
"data": [
18,
19,
20,
21,
22,
23,
24,
25,
26,
27,
28,
... | benchmark/PDF/ICLR2020_rkeJRhNYDH.pdf | openreview | benchmark/MD/ICLR2020_rkeJRhNYDH.md | ICLR 2020 |
rJxBa1HFvS | {
"authorids": [
"aguez@google.com",
"fviola@google.com",
"theophane@google.com",
"lbuesing@google.com",
"skapturowski@google.com",
"doinap@google.com",
"davidsilver@google.com",
"heess@google.com"
],
"title": "Value-Driven Hindsight Modelling",
"authors": [
"Arthur Guez",
... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
... | [
[
{
"role": "PC",
"data": {
"comment": "This paper studies the problem of estimating the value function in an RL setting by learning a representation of the value function. While this topic is one of general interest to the ICLR community, the paper would benefit from a more careful revision... | [
"This paper presents a new model-based reinforcement learning method, termed hindsight modelling. The method works by training a value function which, in addition to depending on information available at the present time is conditioned on some learned embedding of a partial future trajectory. A model is then traine... | [
[
26
],
[
40
],
[
41
],
[
24
],
[
8,
13
],
[
15
],
[
21
],
[
22
],
[
25
],
[
27
],
[
28
],
[
30
],
[
32,
34
],
[
1
],
[
2
],
[
35
],
[
23
],
[
6,
14
],
[
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",
"incorrect",
"correct",
"incorrect",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"incorrect",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correc... | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0,
1
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
2,
3
]
}
],
"category": [
... | benchmark/PDF/ICLR2020_rJxBa1HFvS.pdf | openreview | benchmark/MD/ICLR2020_rJxBa1HFvS.md | ICLR 2020 |
r1lEd64YwH | {
"title": "Learning Semantically Meaningful Representations Through Embodiment",
"authors": [
"Viviane Clay",
"Peter König",
"Kai-Uwe Kühnberger",
"Gordon Pipa"
],
"authorids": [
"vkakerbeck@uos.de",
"pkoenig@uos.de",
"kkuehnbe@uos.de",
"gpipa@uos.de"
],
"keywords": [
"r... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9
]
},
"scores... | [
[
{
"role": "PC",
"data": {
"comment": "What is investigated is what kind of representations are formed by embodied agents; it is argued that these are different than from non-embodied arguments. This is an interesting question related to foundational AI and Alife questions, such as the symb... | [
"Paper Summary: The goal of the paper is to analyze what information is encoded in the representation learned using RL for a specific game. The paper shows that the activations are sparse and the activation patterns are distinct and shows that the conceptually similar images are clustered together in t-sne visualiz... | [
[
1
],
[
2
],
[
6
],
[
11
],
[
3
],
[
7
],
[
17
],
[
4
],
[
5
],
[
10
],
[
12,
18
],
[
15
],
[
16
],
[
19
],
[
13
],
[
8
],
[
0
],
[
9
],
[
14
],
[
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",
"incorrect",
"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": [
"CLAR-WRT"
]
},
... | benchmark/PDF/ICLR2020_r1lEd64YwH.pdf | openreview | benchmark/MD/ICLR2020_r1lEd64YwH.md | ICLR 2020 |
rklTmyBKPH | {
"authorids": [
"jaminfong@hust.edu.cn",
"yzsun@hust.edu.cn",
"kangjian.peng@horizon.ai",
"qian01.zhang@horizon.ai",
"yuan.li@horizon.ai",
"liuwy@hust.edu.cn",
"xgwang@hust.edu.cn"
],
"title": "Fast Neural Network Adaptation via Parameter Remapping and Architecture Search",
"authors... | 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": "PC",
"data": {
"comment": "Main content: Paper proposes a fast network adaptation (FNA) method, which takes a pre-trained image classification network, and produces a network for the task of object detection/semantic segmentation\n\nSummary of discussion:\nreviewer1: interesting... | [
"In this paper, the authors take a MobileNet v2 trained for ImageNet classification, and adapt it either (i) semantic segmentation on Cityscapes, or (ii) object detection on COCO. They do this by first expanding the network into a \"supernet\" and copy weights in an ad-hoc manner, then, they perform DARTS-style ar... | [
[
2
],
[
4
],
[
5
],
[
6
],
[
7
],
[
1
],
[
10
],
[
12
],
[
3
],
[
13
],
[
11
],
[
0
],
[
8
],
[
9
],
[
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"
] | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1
]
},
{
"role": "Author",
"data": [
... | benchmark/PDF/ICLR2020_rklTmyBKPH.pdf | openreview | benchmark/MD/ICLR2020_rklTmyBKPH.md | ICLR 2020 |
SJe9qT4YPr | {
"title": "RISE and DISE: Two Frameworks for Learning from Time Series with Missing Data",
"authors": [
"Alberto Garcia-Duran",
"Robert West"
],
"authorids": [
"alberto.duran@epfl.ch",
"robert.west@epfl.ch"
],
"keywords": [
"Time Series",
"Missing Data",
"RNN"
],
"TL;DR": "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": "PC",
"data": {
"comment": "The paper attacks the important problem of learning time series models with missing data and proposes two learning frameworks, RISE and DISE, for this problem. The reviewers had several concerns about the paper and experimental setup and agree that thi... | [
"The authors propose a new univariate time series analysis framework called RISE, which unifies the existing work on adapting RNNs to irregular time series. Building on top of the RISE, they propose a modification called DISE in which the algorithm skips the intervals without any observations. In that sense, DISE c... | [
[
1
],
[
12
],
[
17
],
[
10
],
[
5,
15
],
[
6
],
[
13
],
[
2,
7
],
[
3,
8
],
[
18
],
[
19
],
[
20
],
[
4
],
[
14
],
[
16
],
[
11
],
[
0
],
[
9
]
] | [
"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",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct"
] | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1
]
}
],
"category": [
"CLAR-FIG"
]
},
... | benchmark/PDF/ICLR2020_SJe9qT4YPr.pdf | openreview | benchmark/MD/ICLR2020_SJe9qT4YPr.md | ICLR 2020 |
S1lVhxSYPH | {
"authorids": [
"dibakar.gope@arm.com",
"jesse.beu@arm.com",
"urmish.thakker@arm.com",
"matthew.mattina@arm.com"
],
"title": "Ternary MobileNets via Per-Layer Hybrid Filter Banks",
"authors": [
"Dibakar Gope",
"Jesse G Beu",
"Urmish Thakker",
"Matthew Mattina"
],
"pdf": "/pd... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4
]
},
"scores": {
"Solid": null,
"Presentation": null,
"No... | [
[
{
"role": "PC",
"data": {
"comment": "The paper presents a quantization method that generates per-layer hybrid filter banks consisting of full-precision and ternary weight filters for MobileNets. The paper is well-written. However, it is incremental. Moreover, empirical results are not con... | [
"The authors focus on quantizing the MobileNets architecture to ternary values, resulting in less space and compute.",
"The space of making neural networks more energy efficient is vital towards their deployment in the real world.",
"I think the authors over-state their claims of no loss in accuracy, in Table 2... | [
[
12
],
[
10
],
[
2
],
[
5
],
[
8
],
[
11
],
[
1
],
[
6
],
[
9
],
[
3
],
[
0
],
[
4
],
[
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",
"correct",
"correct"
] | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0,
1
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
2
]
},
{
"role": "Author",
... | benchmark/PDF/ICLR2020_S1lVhxSYPH.pdf | openreview | benchmark/MD/ICLR2020_S1lVhxSYPH.md | ICLR 2020 |
S1lRg0VKDr | {
"title": "On summarized validation curves and generalization",
"authors": [
"Mohammad Hashir",
"Yoshua Bengio",
"Joseph Paul Cohen"
],
"authorids": [
"mohammad.hashir.khan@umontreal.ca",
"yoshua.bengio@mila.quebec",
"joseph@josephpcohen.com"
],
"keywords": [
"model selection",
... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10
]
},... | [
[
{
"role": "PC",
"data": {
"comment": "The reviewers reached a consensus that the paper is preliminary and has a very limited contribution. Therefore, I cannot recommend acceptance at this time."
}
}
]
] | [
"The paper examines the common practice of performing model selection by choosing the model that maximizes validation accuracy. In a setting where there are multiple tasks, the average validation error hides performance on individual tasks, which may be relevant. The paper casts multi-class image classification as ... | [
[
15
],
[
19
],
[
12
],
[
14
],
[
1
],
[
4
],
[
5
],
[
9
],
[
10
],
[
0
],
[
3,
6,
7
],
[
11
],
[
17
],
[
2
],
[
8
],
[
13
],
[
16
],
[
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"
] | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0,
1,
2,
3
]
}
],
"category": [
"QUAL-EXP"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
4
]
},
{
... | benchmark/PDF/ICLR2020_S1lRg0VKDr.pdf | openreview | benchmark/MD/ICLR2020_S1lRg0VKDr.md | ICLR 2020 |
SyxdC6NKwH | {
"authorids": [
"xxz190020@utdallas.edu",
"feng.chen@utdallas.edu",
"shu2@albany.edu",
"jicho@vt.edu"
],
"title": "Uncertainty-Aware Prediction for Graph Neural Networks",
"authors": [
"Xujiang Zhao",
"Feng Chen",
"Shu Hu",
"jin-Hee Cho"
],
"pdf": "/pdf/265d646e4a599fc10713f... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
... | [
[
{
"role": "PC",
"data": {
"comment": "The authors propose a way to produce uncertainty measures in graph neural networks. However, the reviewers find that the methods proposed lack novelty and are incremental additions to prior work."
}
}
]
] | [
"The authors proposed a Bayesian graph neural network framework for node classification. The proposed models outperformed the baselines in six node classification tasks. The main contribution is to evaluate various uncertainty measures for the uncertainty analysis of Bayesian graph neural networks. The authors show... | [
[
12
],
[
26
],
[
1,
13
],
[
2
],
[
3
],
[
5
],
[
18
],
[
15
],
[
7
],
[
9
],
[
20
],
[
21
],
[
4
],
[
8
],
[
10
],
[
11,
14
],
[
19
],
[
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",
"incorrect",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"incorrec... | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
1,
2
]
},
{
"role": "Author",
... | benchmark/PDF/ICLR2020_SyxdC6NKwH.pdf | openreview | benchmark/MD/ICLR2020_SyxdC6NKwH.md | ICLR 2020 |
ryxQ6T4YwB | {
"title": "GraphNVP: an Invertible Flow-based Model for Generating Molecular Graphs",
"authors": [
"Kaushalya Madhawa",
"Katsuhiko Ishiguro",
"Kosuke Nakago",
"Motoki Abe"
],
"authorids": [
"kaushalya@net.c.titech.ac.jp",
"k.ishiguro.jp@ieee.org",
"nakago@preferred.jp",
"motoki@... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
... | [
[
{
"role": "PC",
"data": {
"comment": "The authors propose an invertible flow-based model for molecular graph generation. The reviewers like the idea but have several concerns: in particular, overfitting in the model, need for more experiments and missing related work. It is important for a... | [
"Contributions:",
"1. This paper proposes an invertible flow-based method for the one-shot graph generation.",
"2. The paper demonstrates their method on a molecular graph generation task.",
"3. Empirical results show the effectiveness of the proposed method.",
"The merit of the proposed invertible flow met... | [
[
5
],
[
21
],
[
28
],
[
6
],
[
7
],
[
22
],
[
11
],
[
14
],
[
15
],
[
3,
18
],
[
8
],
[
19
],
[
20
],
[
26
],
[
10
],
[
12,
13
],
[
17
],
[
25
],
[
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",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"correct",
"correct",
... | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0,
1
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer",
"data": [
2
]
}
],
"category": [
"N/A"
]
... | benchmark/PDF/ICLR2020_ryxQ6T4YwB.pdf | openreview | benchmark/MD/ICLR2020_ryxQ6T4YwB.md | ICLR 2020 |
rkxNh1Stvr | {
"title": "Quantifying Point-Prediction Uncertainty in Neural Networks via Residual Estimation with an I/O Kernel",
"authors": [
"Xin Qiu",
"Elliot Meyerson",
"Risto Miikkulainen"
],
"authorids": [
"qiuxin.nju@gmail.com",
"elliot.meyerson@cognizant.com",
"risto@cognizant.com"
],
"ke... | 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": "PC",
"data": {
"comment": "This paper presents a method to model uncertainty in deep learning regressors by applying a post-hoc procedure. Specifically, the authors model the residuals of neural networks using Gaussian processes, which provide a principled Bayesian estimate of ... | [
"This paper proposes a new framework (RIO) to estimate uncertainty in pretrained neural networks. For this purpose, RIO employs Gaussian Processes whose kernels are calculated by kernel functions of input and output samples and the corresponding target values.",
"- The proposed approach is interesting and the ini... | [
[
25,
27
],
[
33
],
[
5
],
[
6
],
[
11
],
[
12
],
[
13
],
[
26,
37
],
[
35,
36
],
[
38
],
[
45
],
[
22
],
[
19
],
[
42
],
[
9
],
[
30
],
[
44
],
[
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",
"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
]
}
],
"category": [
"SIGN-BRD"
]
},
... | benchmark/PDF/ICLR2020_rkxNh1Stvr.pdf | openreview | benchmark/MD/ICLR2020_rkxNh1Stvr.md | ICLR 2020 |
rklOg6EFwS | {
"authorids": [
"eewangyisen@gmail.com",
"knowzou@ucla.edu",
"jinfengyi.ustc@gmail.com",
"baileyj@unimelb.edu.au",
"xingjun.ma@unimelb.edu.au",
"qgu@cs.ucla.edu"
],
"title": "Improving Adversarial Robustness Requires Revisiting Misclassified Examples",
"authors": [
"Yisen Wang",
... | Accept (Poster) | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10
]
},... | [
[
{
"role": "PC",
"data": {
"comment": "This paper presents modifications to the adversarial training loss that yield improvements in adversarial robustness. While some reviewers were concerned by the lack of mathematical elegance in the proposed method, there is consensus that the proposed... | [
"Summary:\nNeural Networks (NN) have been shown to be susceptible to various adversarial attacks i.e. if we perturb the \"x\" just a little, the output prediction changes.",
"So, there has been much research devoted to how we can make NNs robust to such attacks. Typically, the adversarial examples that are used t... | [
[
18
],
[
27
],
[
30
],
[
20
],
[
5
],
[
1
],
[
2
],
[
7
],
[
12
],
[
3,
16
],
[
19
],
[
21
],
[
26
],
[
28
],
[
14
],
[
17
],
[
22
],
[
23
],
[
29
],
[
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",
"incorrect",
"correct",
"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,
2,
3
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
4
]
}
],
"cat... | benchmark/PDF/ICLR2020_rklOg6EFwS.pdf | openreview | benchmark/MD/ICLR2020_rklOg6EFwS.md | ICLR 2020 |
S1ly10EKDS | {
"authorids": [
"xu.3260@osu.edu",
"wang.10982@osu.edu",
"yi.zhou@utah.edu",
"liang.889@osu.edu"
],
"title": "Reanalysis of Variance Reduced Temporal Difference Learning",
"authors": [
"Tengyu Xu",
"Zhe Wang",
"Yi Zhou",
"Yingbin Liang"
],
"pdf": "/pdf/43b86eef03848350e81352... | Accept (Poster) | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9
]
},
"scores... | [
[
{
"role": "PC",
"data": {
"comment": "The paper studies the variance reduced TD algorithm by Konda and Prashanth (2015). The original paper provided a convergence analysis that had some technical issues. This paper provides a new convergence analysis, and shows the advantage of VRTD to van... | [
"Summary:\nIn this paper, the authors study the variance reduced TD (VRTD) algorithm, by Korda and Prashanth (2015) (KP15), for policy evaluation in RL. They first highlight technical errors in the analysis of KP15, and then provide new convergence analysis for this algorithm. The new analysis is based on a new te... | [
[
50
],
[
1,
30
],
[
4,
42
],
[
5,
44
],
[
6
],
[
7
],
[
27
],
[
28
],
[
36
],
[
37,
46
],
[
43
],
[
45
],
[
47
],
[
49
],
[
51
],
[
10
],
[
15
],
[
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",
"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,
1,
2
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
3
]
},
{
"role": "Au... | benchmark/PDF/ICLR2020_S1ly10EKDS.pdf | openreview | benchmark/MD/ICLR2020_S1ly10EKDS.md | ICLR 2020 |
S1e__ANKvB | {
"authorids": [
"mkl18@mails.tsinghua.edu.cn",
"masonzhao@tencent.com",
"tingyangxu@tencent.com",
"yu.rong@hotmail.com",
"xiaox@sz.tsinghua.edu.cn",
"joehhuang@tencent.com"
],
"title": "Molecular Graph Enhanced Transformer for Retrosynthesis Prediction",
"authors": [
"Kelong Mao",
... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11
... | [
[
{
"role": "PC",
"data": {
"comment": "Several approaches can be used to feed structured data to a neural network, such as convolutions or recurrent network. This paper proposes to combine both roads, by presenting molecular structures to the network using both their graph structured and a ... | [
"This paper focuses on the retrosynthesis prediction problem which to my understanding is the factorization of a target molecule into simpler structures.",
"Previously, retrosynthesis prediction has been tackled as a language translation problem by using a Seq2Seq algorithm called \"Transformer\". This sequential... | [
[
16
],
[
17
],
[
18
],
[
19
],
[
23
],
[
22
],
[
1
],
[
3,
4
],
[
5
],
[
10
],
[
11
],
[
12
],
[
13
],
[
14
],
[
20
],
[
15
],
[
9
],
[
7
],
[
2
],
[
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",
"correct",
"correct"
] | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0,
1,
2,
3
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
4
]
},
{
... | benchmark/PDF/ICLR2020_S1e__ANKvB.pdf | openreview | benchmark/MD/ICLR2020_S1e__ANKvB.md | ICLR 2020 |
SJeFNlHtPS | {
"authorids": [
"davidscottkrueger@gmail.com",
"tegan.jrm@gmail.com",
"legg@google.com",
"leike@google.com"
],
"title": "Hidden incentives for self-induced distributional shift",
"authors": [
"David Scott Krueger",
"Tegan Maharaj",
"Shane Legg",
"Jan Leike"
],
"pdf": "/pdf/3... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10
]
},... | [
[
{
"role": "PC",
"data": {
"comment": "The paper shows how meta-learning contains hidden incentives for distributional shift and how a technique called context swapping can help deal with this. Overall, distributional shift is an important problem, but the contributions made by this paper t... | [
"The authors study the phenomena of self-introduced distributional shift. They define the term along with the term hidden incentives for distributional shift. The latter describes factors that motivate the learner to change the distribution in order to achieve a higher performance. The authors study both phenomena ... | [
[
8
],
[
13
],
[
1
],
[
2
],
[
4
],
[
7
],
[
9
],
[
10
],
[
11
],
[
12
],
[
15
],
[
16,
21
],
[
3
],
[
5
],
[
17,
24
],
[
18
],
[
23
],
[
19
],
[
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",
"incorrect",
"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
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
2
]
},
{
"role": "Author",
... | benchmark/PDF/ICLR2020_SJeFNlHtPS.pdf | openreview | benchmark/MD/ICLR2020_SJeFNlHtPS.md | ICLR 2020 |
S1gvg0NYvH | {
"title": "Mean Field Models for Neural Networks in Teacher-student Setting",
"authors": [
"Lexing Ying",
"Yuandong Tian"
],
"authorids": [
"lexing@stanford.edu",
"yuandong@fb.com"
],
"keywords": [
"mean field model",
"optimal transport",
"ResNet"
],
"TL;DR": "We discuss mea... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
... | [
[
{
"role": "PC",
"data": {
"comment": "This paper studies the evolution of the mean field dynamics of a two layer-fully connected and Resnet model. The focus is in a realizable or student/teacher setting where the labels are created according to a planted network. The authors study the stat... | [
"The paper studies the dynamics of neural network in the Teacher-Student setting, using the approach pionnered in the last few years.",
"Concerning the presentation and the review of other works, I am a bit surprise that the \"teacher-student\" setting appears out of the blue, without any references to any paper.... | [
[
25
],
[
27
],
[
11
],
[
12
],
[
13
],
[
15
],
[
19
],
[
22
],
[
26
],
[
28
],
[
29
],
[
30
],
[
6
],
[
8,
14
],
[
9
],
[
1,
2,
3,
4,
10,
18,
24,
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
]
},
{
"role": "Author",
"data": [
... | benchmark/PDF/ICLR2020_S1gvg0NYvH.pdf | openreview | benchmark/MD/ICLR2020_S1gvg0NYvH.md | ICLR 2020 |
ryxmrpNtvH | {
"authorids": [
"scottyugochang@gmail.com",
"quanlu.zhang@microsoft.com",
"jyjiang97@gmail.com",
"gdzejlin@gmail.com",
"yujing.wang@microsoft.com"
],
"title": "Deeper Insights into Weight Sharing in Neural Architecture Search",
"authors": [
"Yuge Zhang",
"Quanlu Zhang",
"Junyang... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
... | [
[
{
"role": "PC",
"data": {
"comment": "This paper provides a series of empirical evaluations on a small neural architecture search space with 64 architectures. The experiments are interesting, but limited in scope and limited to 64 architectures trained on CIFAR-10. It is unclear whether le... | [
"Many NAS methods rely on weight sharing.",
"Notably in ENAS, a single weight tensor is used for all candidate operations each edge of a cell. In this paper, the authors take a small NAS search space (64 possible networks) and train each network separately to obtain their individual rankings. They then examine ho... | [
[
11
],
[
14
],
[
3,
18
],
[
4
],
[
5
],
[
6
],
[
12
],
[
25
],
[
26
],
[
20
],
[
24
],
[
29
],
[
10
],
[
1,
17,
21,
28
],
[
8
],
[
13
],
[
15
],
[
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",
"incorrect",
"c... | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0,
1
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
2,
4
]
},
{
"role": "Au... | benchmark/PDF/ICLR2020_ryxmrpNtvH.pdf | openreview | benchmark/MD/ICLR2020_ryxmrpNtvH.md | ICLR 2020 |
SylkzaEYPS | {
"abstract": "We present a new approach to defining a sequence loss function to train a summarizer by using a secondary encoder-decoder as a loss function, alleviating a shortcoming of word level training for sequence outputs. The technique is based on the intuition that if a summary is a good one, it should contain... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
... | [
[
{
"role": "PC",
"data": {
"comment": "This paper presents an encoder-decoder based architecture to generate summaries. The real contribution of the paper is to use a recoder matrix which takes the output from an existing encoder-decoder network and tries to generate the reference summary ... | [
"This paper proposes to use an additional component to the commonly used encoder-decoder approach for summarization, which is referred to as the recoder, which is an RNN-syle component that takes the output of the decoder. The intuition offered in the paper is that a good summary should produce itself via the recod... | [
[
14
],
[
15
],
[
5
],
[
6
],
[
7
],
[
23
],
[
24
],
[
25
],
[
26
],
[
27
],
[
28
],
[
3
],
[
4
],
[
9
],
[
10,
18
],
[
11
],
[
2
],
[
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",
"incorrect",
"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",
"data": [
1
]
}
],
"category": [
"N/A"
]
},
{
... | benchmark/PDF/ICLR2020_SylkzaEYPS.pdf | openreview | benchmark/MD/ICLR2020_SylkzaEYPS.md | ICLR 2020 |
rJx9vaVtDS | {
"title": "Individualised Dose-Response Estimation using Generative Adversarial Nets",
"authors": [
"Ioana Bica",
"James Jordon",
"Mihaela van der Schaar"
],
"authorids": [
"ioana.bica@eng.ox.ac.uk",
"james.jordon@wolfson.ox.ac.uk",
"mschaar@turing.ac.uk"
],
"keywords": [
"indiv... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
... | [
[
{
"role": "PC",
"data": {
"comment": "This paper addresses the problem of estimating treatment responses involving a continuous dosage parameter. The basic idea is to learn a GAN model capable of generating synthetic dose-response curves for each training sample, which then facilitates th... | [
"The paper introduces Dose Response Generative Adversarial Network (DRGAN) that is aimed at generating entire dose-response curve from observational data with single dose treatments. This work is an extension of GANITE (Yoon et al., 2018) for the case of real-valued treatments (i.e., dosage). The proposed model con... | [
[
24
],
[
26
],
[
10
],
[
9
],
[
18
],
[
29,
30
],
[
23
],
[
31
],
[
13
],
[
19
],
[
22
],
[
25
],
[
27
],
[
28
],
[
5,
8
],
[
11
],
[
12
],
[
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",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"correct",
"incorrect",
"c... | [
{
"sentences": [
{
"role": "Reviewer",
"data": [
0,
1
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer",
"data": [
2,
3
]
}
],
"category": [
... | benchmark/PDF/ICLR2020_rJx9vaVtDS.pdf | openreview | benchmark/MD/ICLR2020_rJx9vaVtDS.md | ICLR 2020 |
SkgjKR4YwH | {
"title": "MixUp as Directional Adversarial Training",
"authors": [
"Guillaume Perrault-Archambault",
"Yongyi Mao",
"Hongyu Guo",
"Richong Zhang"
],
"authorids": [
"gperr050@uottawa.ca",
"yymao@eecs.uottawa.ca",
"hongyu.guo@nrc-cnrc.gc.ca",
"zhangrc@act.buaa.edu.cn"
],
"keyw... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3
]
},
"scores": {
"Solid": null,
"Presentation": null,
"Novelty": null,
... | [
[
{
"role": "PC",
"data": {
"comment": "This paper builds a connection between MixUp and adversarial training. It introduces untied MixUp (UMixUp), which generalizes the methods of MixUp. Then, it also shows that DAT and UMixUp use the same method of MixUp for generating samples but use diff... | [
"First of all, the concept of 'Directional Adversarial Training (DAT)' is not appropriate. Actually similar method has been proposed in Hiroshi Inoue (2018) as a data augmentation method.",
"What surprised me the most is that after trying to connect UMixUp with adversarial training in the whole paper, there is no... | [
[
2
],
[
21
],
[
14
],
[
0,
3
],
[
17
],
[
18
],
[
8
],
[
4
],
[
6
],
[
10,
11,
12,
13,
19
],
[
1
],
[
15
],
[
16
],
[
20
],
[
5
],
[
7
],
[
9
]
] | [
"correct",
"correct",
"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",
"correct",
"correct",
"incorrect",
"correct",
"correct",
"correct",
"incorrect",
"correct",
"incorrect",
"correct"
] | [
{
"sentences": [
{
"role": "Reviewer 1",
"data": [
0
]
},
{
"role": "Author",
"data": [
4,
5,
6,
7,
8
]
}
],
"category": [
"ORIG-MTH"
]
},
{
"sentences": ... | benchmark/PDF/ICLR2020_SkgjKR4YwH.pdf | openreview | benchmark/MD/ICLR2020_SkgjKR4YwH.md | ICLR 2020 |
S1ejj64YvS | {
"authorids": [
"fenghz@zju.edu.cn",
"kong@cs.umd.edu",
"zhangtianye1026@zju.edu.cn",
"3160104527@zju.edu.cn",
"chenwei@cad.zju.edu.cn"
],
"title": "Good Semi-supervised VAE Requires Tighter Evidence Lower Bound",
"authors": [
"Haozhe Feng",
"Kezhi Kong",
"Tianye Zhang",
"Si... | Reject | [
[
{
"role": "Author",
"data": {
"value": {
"comment": [
0,
1,
2,
3
]
}
}
}
],
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review... | [
[
{
"role": "PC",
"data": {
"comment": "The paper proposes to combine a VAE model with the Optimal Transport to approximate some components of the model. The authors evaluate their approach on semi-supervised problems and claim to obtain very competitive results compared to literature. Unfor... | [
"We have modified a symbol error and a display error in Figure 1: The schematic of OSPOT-VAE.",
"1. Change the original bitmap figure to the vector graph (from .png to .svg).",
"2. Change the KL-divergency symbol 'KL()' to 'D_{KL}()', matching the symbols in the paper.",
"The new figure is now available on th... | [
[
11
],
[
21
],
[
27
],
[
2
],
[
9
],
[
12
],
[
17
],
[
18,
29
],
[
22
],
[
26
],
[
15,
16
],
[
3
],
[
4,
5,
6,
7
],
[
8
],
[
10
],
[
19
],
[
25
],
[
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",
"incorrect",
"correct",
"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": [
4
]
}
],
"category": [
"N/A"
]
},
{
"sentences": [
{
"role": "Reviewer 2",
"data": [
5
]
},
{
"role": "Author",
"data": [
... | benchmark/PDF/ICLR2020_S1ejj64YvS.pdf | openreview | benchmark/MD/ICLR2020_S1ejj64YvS.md | ICLR 2020 |
Syl38yrFwr | {
"authorids": [
"james.lichao.sun@gmail.com",
"yingbo.zhou@salesforce.com",
"jia.li@salesforce.com",
"rsocher@salesforce.com",
"psyu@uic.edu",
"cxiong@salesforce.com"
],
"title": "Near-Zero-Cost Differentially Private Deep Learning with Teacher Ensembles",
"authors": [
"Lichao Sun",... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1
]
},
"scores": {
"Solid": null,
"Presentation": null,
"Novelty": null,
"Overall": 1,
... | [
[
{
"role": "PC",
"data": {
"comment": "This paper presents a differentially private mechanism, called Noisy ArgMax, for privately aggregating predictions from several teacher models. There is a consensus in the discussion that the technique of adding a large constant to the largest vote bre... | [
"The paper proposes an improvement on the PATE framework for achieving near-zero privacy cost, and showed privacy analyses and experimental evaluations of the proposed method.",
"The proposed method can be technically flawed. Adding a consent to the max will not guarantee privacy unless you account for the privac... | [
[
1,
3,
5
],
[
0
],
[
2
],
[
4
]
] | [
"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/ICLR2020_Syl38yrFwr.pdf | openreview | benchmark/MD/ICLR2020_Syl38yrFwr.md | ICLR 2020 |
r1lh6C4FDr | {
"title": "COMBINED FLEXIBLE ACTIVATION FUNCTIONS FOR DEEP NEURAL NETWORKS",
"authors": [
"Renlong Jie",
"Junbin Gao",
"Andrey Vasnev",
"Minh-Ngoc Tran"
],
"authorids": [
"renlong.jie@sydney.edu.au",
"junbin.gao@sydney.edu.au",
"andrey.vasnev@sydney.edu.au",
"minh-ngoc.tran@sydn... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
... | [
[
{
"role": "PC",
"data": {
"comment": "Main content: Proposes combining flexible activation functions \n\nDiscussion:\nreviewer 1: main issue is unfamiliar with stock dataset, and CIFAR dataset has a bad baseline.\nreviewer 2: main issue is around baselines and writing. \nreviewer 3: main i... | [
"This paper presents a family of parameterized composite activation functions, and a regularization technique for parameterized activation functions in general.",
"While the three sets of experiments show potentially promising results, they aren't able to disambiguate clearly between the effect of the activation ... | [
[
7
],
[
1
],
[
2
],
[
3
],
[
4
],
[
5
],
[
6
],
[
23
],
[
24
],
[
14
],
[
15,
18
],
[
19
],
[
21
],
[
0
],
[
9
],
[
10
],
[
22
],
[
11
],
[
12
],
[
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 1",
"data": [
0,
1
]
},
{
"role": "Author",
"data": [
23,
31,
32,
33,
81,
82
]
}
],
"category": [
"QUAL-EXP"
... | benchmark/PDF/ICLR2020_r1lh6C4FDr.pdf | openreview | benchmark/MD/ICLR2020_r1lh6C4FDr.md | ICLR 2020 |
SJgXs1HtwH | {
"authorids": [
"vinojjayasundara@gmail.com",
"dqnbui.2016@phdis.smu.edu.sg",
"lxjiang@smu.edu.sg",
"davidlo@smu.edu.sg"
],
"title": "TreeCaps: Tree-Structured Capsule Networks for Program Source Code Processing",
"authors": [
"Vinoj Jayasundara",
"Nghi Duy Quoc Bui",
"Lingxiao Jian... | Reject | [
[
{
"role": "Reviewer",
"data": {
"summary_of_the_paper": null,
"value": {
"review": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9
]
},
"scores... | [
[
{
"role": "PC",
"data": {
"comment": "This paper proposes an application of capsule networks to code modeling.\n\nI see the potential in this approach, but as the reviewers pointed out, in the current draft there are significant issues with respect to both clarity of motivating the work, a... | [
"This paper proposes a tree-structured capsule network for program source code processing (essentially a program classification task with three datasets).",
"The idea of incorporating tree structures into the design for capsule networks is not wrong.",
"However, I am not sure why a capsule network is even neede... | [
[
3
],
[
7
],
[
17
],
[
21
],
[
1
],
[
2
],
[
8
],
[
4
],
[
14
],
[
19
],
[
20
],
[
5
],
[
11
],
[
12
],
[
13
],
[
10
],
[
16
],
[
18
],
[
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",
"incorrect",
"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
]
}
],
"category": [
"ORIG-COM"
]
},
... | benchmark/PDF/ICLR2020_SJgXs1HtwH.pdf | openreview | benchmark/MD/ICLR2020_SJgXs1HtwH.md | ICLR 2020 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.