paper_id stringlengths 10 10 | items listlengths 4 82 | gt float64 1 9.5 |
|---|---|---|
DL9txImSzm | [
"strength: The research problem for this work is good. Imitation learning is a reasonable method for many control tasks where reward is difficult to specify and expert data is available. Aiming to improve the imitation policy from samples is an interesting a relevant problem.",
"strength: The novelty of this work... | 6 |
5KojubHBr8 | [
"strength: Enhanced Multi-modal Understanding: MMICL's approach to handling complex prompts with multiple images and text could significantly improve VLMs' performance on downstream tasks.",
"strength: State-of-the-Art Performance: The paper reports new benchmarks in zero-shot performance on vision-language tasks... | 5.6 |
59r0ntInvF | [
"strength: The paper proposes a novel training approach that addresses the complexity of learning pattern distributions in IR by breaking it down into simpler stages, which is a creative solution to a known challenge in the field.",
"strength: The method is evaluated extensively on benchmark datasets, demonstrati... | 4.666667 |
WNkW0cOwiz | [
"strength: This paper highlights a unique and previously unexplored challenge with DDPM: the instability encountered when learning $\\epsilon_{\\theta} = \\sigma_{t} \\cdot \\nabla \\log q_{t}(x)$ during the time steps where $\\sigma_{t}$ is minimal. One might naturally question why DDPM doesn't directly learn $\\n... | 7.5 |
vkakKdznFS | [
"strength: The authors have conducted extensive experiments and ablation studies to demonstrate the effectiveness of their proposed method.",
"strength: model pixel label as semantic text token by generalized VLLM model",
"strength: compress token length with row-wise run-length encoding",
"strength: comprehe... | 6.333333 |
21rSeWJHPF | [
"strength: The paper aims to promote balancedness in nodes’ centrality ranking, using community detection as a concrete application scenario. I find this focus interesting.",
"strength: The paper proposes a multi-core-periphery structure with communities (MCPC) to quantify unbalancedness in centrality measures.",... | 5.75 |
awHTL3Hpto | [
"strength: It provides the first in-depth, systematic study on the expressive power of ReLU networks under a wide range of convex relaxations commonly used in neural network certification.",
"strength: The analysis covers univariate and multivariate functions and simple as well as more complex function classes li... | 6.333333 |
WYsLU5TEEo | [
"strength: The authors make an interesting use case of the image to image translation framework where they convert an image from one class to another; because of the nature of the dataset, where one class has some artifacts (damage) which the other class does not, the resulting generator appears to be only introduc... | 2.5 |
w5h443GIGo | [
"strength: Timely and important problem especially due to the rise of IoT applications and the need for unsupervised data exploration",
"strength: Simply and intuitive ideas",
"strength: Results support the overall claims in the paper",
"strength: The paper addresses the relevant problem of automatically dete... | 2.333333 |
xQVxo9dSID | [
"strength: The approach is very efficent as they showed. It could be used to greatly mprove efficiency and performance of CMs at a large scale.",
"strength: The primary strength of this work lies in its empirical results, achieving state-of-the-art performance on standard image generation benchmarks.",
"strengt... | 6.75 |
aAI92OHA4t | [
"strength: Even though the proposed work lacks significant novelty, I do believe that there is some promise behind it. The, albeit limited results, show signs of the method working reasonably and with some further/major refinement could offer a nice solution.",
"strength: Figures throughout are useful and relevan... | 2.333333 |
VpCqrMMGVm | [
"strength: a timely topic is treated, how models that are used in practice perform mathematical addition and subtraction",
"strength: a large number of figures that show how attention heads are activated on concrete examples help to make the paper readable",
"strength: The language of the paper is concise and c... | 5.25 |
BmYzoPppij | [
"strength: Addresses the critical issue of carbon footprint estimation for LLM inferences",
"strength: This paper separate modeling of prefill/decode phases and provides more accurate estimate",
"strength: Use GNN to predict the carbon footprint and shows promising accuracy.",
"strength: The authors found a r... | 3.333333 |
7LZjuA4AB2 | [
"strength: The paper is easy to follow and for the most part well‑written.",
"strength: While the paper does not propose novel methods for robustness, it leads an important discussion on the interplay of pre‑trained networks with training methods for group‑robustness.",
"strength: It provides a novel characteri... | 3 |
E2CR6hmV1I | [
"strength: The method effectively tackles the challenges of sparse rewards and rigid role assignments in multi-agent learning.",
"strength: The two-stage learning process not only enhances adaptability across various tasks and environments but also allows for automated data generation, reducing the need for manua... | 3 |
fRaK0cG9L8 | [
"strength: The issue of modeling 2AFC experimental data is an interesting topic.",
"strength: The proposed method enjoys simplicity.",
"strength: The paper introduces a novel method for evaluating perceptual distance models using binomial distributions to model decision-making in two-alternative forced choice (... | 3.666667 |
OZ3NXrF3gQ | [
"strength: Originality: The approach of using backward world models to derive policies in a reward-free setting is a creative departure from traditional imitation and reinforcement learning. By focusing solely on goal states rather than rewards, RFPO aligns well with goal-directed tasks.",
"strength: Quality: Met... | 2.5 |
SnDmPkOJ0T | [
"strength: The proposed approach REEF is training-free, simple and efficient.",
"strength: The proposed approach REEF does not impair the model’s general capabilities.",
"strength: The proposed approach REEF is robust to sequential fine-tuning, pruning, model merging, and permutations.",
"strength: The propos... | 8 |
M992mjgKzI | [
"strength: Recently, there has been an extensive push towards the development of benchmarks that focus on different aspects of reinforcement learning (e.g., for offline RL [1], [2]). However, as noted by the authors, there existed no centralized evaluation suite for algorithms in offline GCRL and the community ofte... | 7 |
WDxa9hnz4p | [
"strength: Quality: auto-demo prompt outperforms standard batch prompting on 7/10 tasks across batch sizes.",
"strength: Significance: improving inference-time efficiency in terms of prompting techniques is an important aspect of using LLMs on large amounts of data.",
"strength: The proposed method is interesti... | 2.333333 |
GXXQfSpJNI | [
"strength: Ensuring fairness in generative model is critical.",
"strength: The writing is easy to follow, for the most part.",
"strength: The proposed method is fairly simple.",
"strength: The proposed method only manipulates the latent space, thus taking less computational resources.",
"strength: A critica... | 2.333333 |
2GJm8yT2jN | [
"strength: The authors ground their approach in a concrete example that is returned to throughout thereby helping the reader to build a stronger understanding and intuition for the work.",
"strength: Method is demonstrated for multiple modalities suggesting its broader utility. The method seems to possess impactf... | 5.666667 |
qZwtPEw2qN | [
"strength: the paper is well written",
"strength: the tackled problem is very interesting",
"strength: experiments are exhaustive and convincing (Figure 1 shows the regime of interest for using noisy data)",
"strength: The studied problem is interesting and important.",
"strength: The authors provide theore... | 6.8 |
t7P5BUKcYv | [
"strength: With the proper hyper-parameters, MoE++ introduces zero-computation experts who can reduce the computational load by bypassing or simplifying processing for certain tokens, leading to efficient resource use.",
"strength: These heterogeneous experts with suitable routing designs can improve the model's ... | 8 |
MV5j4Qpq7N | [
"strength: The use of system prompt attention for detection is a novel contribution, addressing the limitations of existing methods that rely heavily on embedding-based classifiers.",
"strength: AttentionDefense is designed to be less computationally intensive than traditional LLM detection methods, making it mor... | 2.333333 |
Ze49bGd4ON | [
"strength: The introduction of a constrained tree memory structure is a novel contribution that effectively addresses the limitations of the existing SAM2 model, particularly in handling long video segmentation tasks.",
"strength: The method's ability to prevent error accumulation by maintaining multiple segmenta... | 5.25 |
Ij9ilPh36h | [
"strength: The paper is well-written and easy to follow",
"strength: The phenomenon is novel and quite surprising, especially its implication in reducing the repetition issues. The paper opens up a new perspective in understanding repetition in text generation.",
"strength: I enjoyed reading the experiment resu... | 6.25 |
ujX2l7mNX6 | [
"strength: Paper is very well-written even for a general reviewer. Furthermore, the settings of the model and experiments are accurately described as to make it easier for reproducibility.",
"strength: The novelty of using semantic features in the middle of their end-to-end model is quite intuitive. Plus this mes... | 5.75 |
LSp4KBhAom | [
"strength: The paper is very well written and easy to understand.",
"strength: I very much appreciate the core calibration framework text section and underlying method, in that is it technically principled, produces strong state-of-the-art improvements, and is presented in an intuitive manor. This is a great exam... | 7.5 |
v4PnwdA056 | [
"strength: The paper studies an interesting problem of the tradeoff between privacy and fairness in fine-tuning.",
"strength: As a mitigation, the paper further proposes a training-free method.",
"strength: The paper investigates an interesting phenomenon in supervised fine-tuning methods, where their approach ... | 4 |
YpWV7XRmFB | [
"strength: Efficient and simple method for in-context editing of knowledge in LLMs that does not require access to the model parameters.",
"strength: The writing in this article is well-organized, and the proposed approach is clear and easy to understand.",
"strength: This paper finds that previous in-context e... | 4 |
oX4FcNA4UC | [
"strength: The use of the Girsanov theorem for gradient estimation is theoretically sound, particularly in reducing the computational burden traditionally associated with Neural SDEs. This efficiency makes the approach suitable for high-dimensional SOC problems and applications that involve large neural networks.",... | 4.25 |
cb4etlGvOY | [
"strength: The proposed SALA method effectively combines reasoning, acting, and adaptation within a single language model, reducing model complexity while achieving autonomy.",
"strength: The integration of a correction mechanism enhances decision-making capabilities, demonstrating significant improvements over p... | 2.5 |
ylhKbwJrjC | [
"strength: The paper is well written and the theoretical results appear to be correct.",
"strength: The paper improves the previous results in Osogami [2023].",
"strength: The paper proposes numerical experiments to show the advantages of their designs.",
"strength: The automated mechanism design is an intere... | 4.666667 |
PVHoELf5UN | [
"strength: The paper combined the Retinex model and data-driven methods since the first effort of RetinexNet.",
"strength: Experiments are sufficient.",
"strength: Frequency domain decomposition is well-used.",
"strength: The unsupervised low-light enhancement method is a promising direction.",
"strength: T... | 6.4 |
E1EHO0imOb | [
"strength: The paper demonstrates opportunities and challenges in scaling FP8 training to trillion-token scale using LLaMA2 architecture. In particular, the authors identify that with SwiGLU (which is an important to contemporary LLMs), weight alignment issues can induce large-magnitude outliers and result in train... | 7.5 |
oZdaEiDBpF | [
"strength: The exploration of class-specific risk in MI-PLL settings is a novel angel.",
"strength: The theory is sound and robust.",
"strength: The proposed algorithms lead to empirical improvements.",
"strength: They derive class-specific error bounds that depend on the MI-PLL risk under minimal assumptions... | 5 |
csukJcpYDe | [
"strength: The authors did a good job presenting the necessary background of TT and its associated operations (such as decomposition, rounding, TT-Cross, and TT-go). I think the authors have fairly discussed the limitation of the method.",
"strength: The paper introduces TTPI, an Approximate Dynamic Programming (... | 7.5 |
ZPCBcR7Drg | [
"strength: The paper addresses an important problem. The authors propose a full pipeline in terms of data, algorithm, and evaluation for the problem.",
"strength: The paper goes beyond just English-speaking locales, which is important if we want to deploy AD/ADAS systems in the real world.",
"strength: The data... | 5 |
NPNUHgHF2w | [
"strength: The utilization of a criss-cross transformer to independently model spatial and temporal dependencies is a significant advancement. By partitioning attention mechanisms into Spatial-Attention (S-Attention) and TemporalAttention (T-Attention), CBraMod effectively captures the heterogeneous dependencies in... | 6.75 |
FDnZFpHmU4 | [
"strength: The paper identifies compatibility challenges in LLM ensembling and focuses on top-k tokens, aligning this strategy with empirical evidence that vocabulary alignment often introduces computational inefficiencies.",
"strength: The authors conduct extensive experiments on multiple models and benchmarks, ... | 7.5 |
506Sxc0Adp | [
"strength: The motivation behind building data quality metrics is an important direction for building better language models.",
"strength: The paper is well written and easy to follow along.",
"strength: This paper is well-written and rigorous.",
"strength: I find the idea of improving the performance of pre-... | 4 |
dDpB23VbVa | [
"strength: Reducing LLM training costs is an important research question. The problem formulation, motivation, and improvement objective of the paper are clear.",
"strength: The proposed method is described clearly, and there seem to be adequate details to reproduce the work. The cost reduction of the proposed tr... | 7.5 |
ljVCPV7jK3 | [
"strength: The paper targets an important problem. Given the increasingly stringent privacy constraint, the problem of studying fairness without full access to sensitive attributes is an important problem.",
"strength: The authors present a solution to a significant challenge that fairness-enhancing interventions... | 4 |
hqUznsPMLn | [
"strength: The approach is quite simple, and given access to an instruction-following LLM like ChatGPT, it would be straightforward to reproduce a comparable approach to ACES. The prompts included in the Appendix and the algorithm provided in the text aid in this reproducibility meaningfully. (The reliance on gpt-3... | 3.666667 |
mhgm0IXtHw | [
"strength: The presented concept is intriguing and efficient. It details an uncomplicated yet effective method of using noise map conditioning during real image inversion, which streamlines the reverse process and eradicates path divergence between the reconstruction path and inversion trajectory. This leads to a m... | 6.666667 |
yLhJYvkKA0 | [
"strength: The proposed algorithm is simpler and much more efficient than that in prior work.",
"strength: The multiplicative error is not too high for the graphs with edge weight at least 1, which is also a nice result and improvement over prior work.",
"strength: The reduction to balanced cut is an interestin... | 6.666667 |
prTI7MSt2X | [
"strength: The results show that the learned latent space captures the space of paths well. IO-LVM captures nuances such as big ships not passing through Oresund Straight even when it's the shortest path purely based on the data.",
"strength: The problem of learning generative models of optimal paths is interesti... | 4.5 |
Exkm5OReTY | [
"strength: Tackles the prevalent problem of missing features in tabular data.",
"strength: Shows slight improvement on TabReD benchmark datasets, though the results are somewhat questionable.",
"strength: Masking is an interesting approach to modeling missing values.",
"strength: The paper is well written.",
... | 3.25 |
8OrXrdPbef | [
"strength: The authors propose a promising and effective alternative to previous proposals for tackling the challenges of the clustered FL setting.",
"strength: The method FLAG proposed in this work is well motivated by the research gap in the literature and is the first to combine both data and gradient similari... | 4.25 |
aOPTDchLBz | [
"strength: The corpus has more than 10k hours of Hebrew speech, which is likely to be very valuable for future research on Hebrew ASR.",
"strength: The corpus also has CC-license, which makes it available to many use cases.",
"strength: Language-specific datasets play a crucial role in advancing ASR technology ... | 2.5 |
XCP0MOMLPo | [
"strength: The proposed method is the first gradient-based optimization of a parity-check matrix for error-correcting codes.",
"strength: In addition, a grid-search approach for finding the learning rate is proposed, which is an alternative to the line-search approach.",
"strength: Numerical results show that o... | 4.4 |
oY2jw2NLiM | [
"strength: The authors introduce a new problem, k-mean clustering of segments, in this paper, which may have its own interest in the future.",
"strength: They also proposed novel algorithm that generate coreset of small size, which transfer the problem into a transitional weighted k-means problem, which makes the... | 3 |
b2LklBgdcL | [
"strength: The study targets a timely and important area in addressing fairness concerns in Med (M)LLMs.",
"strength: Presenting a new benchmarking dataset for evaluating fairness is appreciated.",
"strength: The reported messages upon the empirical analysis are interesting, especially observing that medical-sp... | 3.5 |
x17qiTPDy5 | [
"strength: The paper tackles a well motivated and relevant topic: the links between GANs and score-based diffusion models, unifying both of them in a single framework, which is interesting as it opens up potentially fruitful areas of research.",
"strength: The unifying equation DiffFlow and its dual interpretatio... | 5 |
af2c8EaKl8 | [
"strength: The authors have demonstrated a potential risk and inefficiency of using the transformer with a long context length K when Markov property is strong.",
"strength: As Markov property can be interpreted as a locality (or local dependence) in the sequence of interactions between the agent and the environm... | 7 |
OwpLQrpdwE | [
"strength: The paper is very well written. The topic relies heavily on results from operator kernels, but I found that the introduction to the topic by the authors was adequate.",
"strength: The contributions of the authors are novel and have promising applications in data-driven discovery of latent dynamics.",
... | 7.4 |
NSefAqUM6U | [
"strength: Authors proposed federated training approach for SOMs",
"strength: Deep clustering is an important direction, of interest to many in the ICLR community.",
"strength: Practitioners of SOMs may appreciate new methods for their use in the deep clustering area.",
"strength: This paper extends tradition... | 3 |
EP09OGPRzk | [
"strength: The paper is easy to follow for non-theory parts and presents a concise overview of the literature in this domain.",
"strength: The choice of PDEs is diverse and incorporates diverse challenges observed while training PINNs for simulating PDEs.",
"strength: Sample trajectory plots help observe the me... | 6 |
1959usnw3Z | [
"strength: This paper is easy to understand.",
"strength: Experiments are conducted on three new datasets.",
"strength: The organization is good to follow.",
"strength: The authors find a new potential of training large-scale graph.",
"strength: The scalability of GNNs is an important research problem.",
... | 3 |
pzZjyYee6L | [
"strength: A very relevant problem being investigated.",
"strength: Intuitive idea being proposed.",
"strength: Promising experimental results shown.",
"strength: The method proposed makes sense and is simple.",
"strength: The analysis on dataset size and noise was nice to see, and not common in existing li... | 2.5 |
Pjkes5MdKI | [
"strength: The paper is very thorough and complex. There is a lot of detail and technical content. I have never seen a control system like that used for the neural network feedback control and it seems like a very novel way of synthesizing programs. While I don't fully understand it, the authors analysis of inner c... | 2.5 |
hMjUnF3aQ8 | [
"strength: The paper is relatively clear to understand.",
"strength: Although the proposed approach is very simple, the authors show it can lead to strong performance improvement when used properly",
"strength: overall good quality",
"strength: Simplicity: I appreciate the simplicity of the proposed approach.... | 2 |
OF5x1dzWSS | [
"strength: The paper is well-written and easy to follow.",
"strength: The motivation is clear and the equivalent compositional optimization problem is reasonable.",
"strength: The proposed CID method has convergence guarantee.",
"strength: The mathematical formulation of instance-reweighted bilevel optimizati... | 6.666667 |
j5JvZCaDM0 | [
"strength: Safety in RL is an important and relevant topic and this paper is a nice addition to the existing body of work.",
"strength: The approach to introduce safety via the three learning objectives is well motivated and substantially explained (although I did not follow all details).",
"strength: The usage... | 7.5 |
OhauMUNW8T | [
"strength: The paper is easy-to-follow.",
"strength: The length extrapolation problem is important for language models.",
"strength: The authors provide a solid theoretical foundation by drawing parallels between RoPE and wavelet transforms, and by extending this analogy to propose their method.",
"strength: ... | 5.25 |
BKGM8fyFIo | [
"strength: The proposed method employs a hierarchical weighted directed acyclic graph instead of a tree structure, utilizing multi-path dynamic retrieval and hierarchical summary nodes.",
"strength: It is an interesting idea to assign weights to the edges based on attention, which allows GARLIC to adjust the retr... | 5 |
6PcJEFKvBD | [
"strength: The paper proposes a useful Python package for the off-policy evaluation methods in the RL domain, which can be helpful for an easy-to-use toolbox if one wants to implement an evaluation method quickly.",
"strength: The paper discussed some technique details of existing work, making readers out of the ... | 2.333333 |
tErHYBGlWc | [
"strength: Novel Perspective on Representation Specialization: This paper provides a fresh and thorough analysis of actor and critic representation specialization using information-theoretic metrics.",
"strength: Convincing Empirical Evidence: The authors provide substantial empirical support for their claims, wi... | 6.8 |
TPZRq4FALB | [
"strength: This paper has a good motivation. The authors focus on multi-modal test-time adaption and reveal a NEW task-specific challenge (i.e., cross-modal reliability bias) for the first time. This paper first empirically proves that the existing test-time adaption methods cannot tackle the cross-modal reliabilit... | 8 |
RVPZJpmyGU | [
"strength: The idea of performing routing by clustering is intuitive and promising.",
"strength: The proposed method is simple and seemingly easy to implement, which shows the potential of the proposed method for replacing existing MoE-based methods.",
"strength: The experiments are extensively conducted in lan... | 4.6 |
8vzMLo8LDN | [
"strength: Writing is very clear and the methodology is well explained. This allows readers to understand the differences between this method and previous ones.",
"strength: Interesting use of context vectors through the 3-network model. Ablation studies in supplementary material show the need for such an archite... | 6.25 |
GkJOCga62u | [
"strength: The paper proposes a hierarchy which brings under one umbrella different GNN architectures.",
"strength: The paper provides a study of theoretical expressiveness (akin to the WL hierarchy)",
"strength: The paper shows empirical evidence on simple new datasets (proposed by the authors) to validate the... | 7 |
mXpNp8MMr5 | [
"strength: This study takes a leading role in exposing the vulnerability of adversarially trained models to 'two-faced attacks', which fraudulently overestimate the robustness of the model during the verification stage.",
"strength: Notable features include its comprehensive evaluation across a variety of model a... | 7.333333 |
QQ5eVDIMu4 | [
"strength: The problem this paper targets is a significant problem and the paper is well motivated",
"strength: The proposed model seems reasonable and interesting to my knowledge",
"strength: The experiment results are promising and the improvements are solid",
"strength: The paper aptly addresses OOD as a c... | 5 |
NkYCuGM7E2 | [
"strength: interesting application to showcase the power of LLMs.",
"strength: demonstration of a full pipeline on how to use an LLM for control from observations.",
"strength: Combining LLM and MPC is a smart way to leverage the high-level reasoning capability of LLM. Since LLM is not good at low-level control... | 3.75 |
iGV6Sg5bI0 | [
"strength: Overall I find this work to be exciting and interesting.",
"strength: The use of a multi-agent LLM method to generate a benchmark is both a timely and resourceful solution.",
"strength: Moreover, textual conditioning allows for more diverse and nuanced means to condition time-series generation; a val... | 5.2 |
5I39Zvlb3Y | [
"strength: Collu-Bench differs from previous benchmarks by focusing on finer-grained code hallucinations, providing a new benchmark that includes richer features such as log probabilities and execution feedback. It aims to deepen understanding and predict where hallucinations occur.",
"strength: The authors write... | 4.2 |
cHy00K3Och | [
"strength: The work is very simple and straight forward, particularly so for smaller datasets. Run the dataset through a model for a few epochs to stabilize the training, run every sample and measure the gradients of the last layer, measure a cross-similarity matrix and average it column-wise and sort.",
"strengt... | 2.5 |
d8w0pmvXbZ | [
"strength: Overall, I think the experimental work in this paper was well executed and carefully controlled.",
"strength: The authors successfully reproduce training instabilities on smaller transformers, by increasing the learning rate; they show that as model size increases (e.g. figure 1 and figure 6), training... | 8 |
B5iOSxM2I0 | [
"strength: Tokenization is a critical aspect of modern-day natural language processing, but its theoretical underpinnings are not yet fully understood; the formalisms introduced in the paper help close this gap and might become the basis for future work.",
"strength: The application of stochastic maps to tokeniza... | 6.5 |
vw0NurJ7UX | [
"strength: The authors showed the possibility that per-tensor static quantization can outperform per-token dynamic quantization.",
"strength: They measured the real time-to-first-token (pre-filling) speed-up.",
"strength: I like the fact that the paper reports wall-clock inference speed on various devices (rtx ... | 3 |
9JE3HogPCw | [
"strength: The paper is generally well-written.",
"strength: The proposed solution is simple to use and novel.",
"strength: The usefulness of the proposed method is shown to be more effective than standard DQN.",
"strength: Overall, I found the paper well-written and easy to parse through.",
"strength: Alre... | 4.75 |
MiPyle6Jef | [
"strength: The illustrations are thoughtfully crafted.",
"strength: The manuscript presents a clear structure, and maintains logical consistency.",
"strength: A detailed analysis of the weight distribution in SNNs provides the motivation for the re-scaling method.",
"strength: The experiments demonstrate the ... | 6.75 |
4h1apFjO99 | [
"strength: This idea and algorithm of the proposed method are generally well presented.",
"strength: It’s compared with multiple recent time series generation methods and simulation results outperform these in most cases.",
"strength: The model explores both unconditional and conditional generation using a sing... | 6.333333 |
71pur4y8gs | [
"strength: The paper proposes the first sampling-phase watermarking method for tabular diffusion models.",
"strength: To enhance the robustness, the paper proposes a valid bit mechanism.",
"strength: The paper shows theoretical guarantee for the proposed method.",
"strength: Extensive experiments validate the... | 7.2 |
IB1HqbA2Pn | [
"strength: The significant contributions mostly lie in the data perspective, while the training algorithm and the model architecture are basically following the previous work.",
"strength: The authors create a new multimodal instruction-following tool using data, integrating lots of real-world tools (skills), lik... | 3.25 |
SuH5SdOXpe | [
"strength: Proposed idea of reprogramming is interesting as it would allow reusing the learned features by the original model. May be an important approach to explore in large models.",
"strength: Empirical results consider multiple datasets of various sizes and different perturbations",
"strength: There is a g... | 7.5 |
JzFLBOFMZ2 | [
"strength: The targeted problem is valid and important, timely Since the year 2023, there has been more evidence that LLM has a sort of commonsense causal knowledge, and it is very important to consider leveraging its power to enhance data-driven CSL for causal discovery.",
"strength: The idea of using the data-d... | 3.2 |
ErpRu7qMq1 | [
"strength: As far as I know this is the first application of discrete diffusion to symbolic music generation.",
"strength: The generated samples sound quite good!",
"strength: Evaluation seems good, with the caveat that I don't really trust any evaluation of generative music models :)",
"strength: The paper i... | 4 |
keA1Ea7v6p | [
"strength: It is interesting to utilize the power of foundational models to assist federated learning.",
"strength: The experiment and ablation study are detailed.",
"strength: Thorough experiments that investigate the algorithmic choices and how federated learning is affected under new generative data.",
"st... | 5.666667 |
YJxhZnGU1q | [
"strength: The proposed problem setting is timely and interesting and the high-level idea of recommending states to influence agents' beliefs sounds reasonable.",
"strength: The problem of inducing learning agents towards particular solutions is a very interesting direction on both the theoretical and applied sid... | 4.25 |
kRBQwlkFSP | [
"strength: Well written",
"strength: Easy to Understand",
"strength: The idea of projecting the gradient to the manifold of intermediate noise is novel and making sense to me. This method supposes to suppress artifacts that arises with hard optimization.",
"strength: The paper is written very well. The work i... | 6.75 |
gVbPYihQag | [
"strength: Innovative Integration of Diffusion in Sequential Modeling: The authors embed the diffusion process into each time step within the model, enhancing its ability to capture dynamic stochastic behaviors and improving its adaptability to high-variability time series data.",
"strength: Data-Driven Prior Kno... | 5 |
eOCvA8iwXH | [
"strength: The idea is quite natural and convincing. It does seem more suitable for time series and video data, due to the lack of availability of suitably transformed data in usual setting, but for a framework it seems reasonable.",
"strength: It relies on the natural idea of using an invariant kernel mapping (w... | 7 |
M9SAhECerP | [
"strength: Although fairness has been widely studied in the machine learning literature, it has been less considered in RL, to my knowledge. In that sense, the problem formulation is relatively \"original\".",
"strength: The approach to ensuring fairness through optimizing bisimulation metrics is \"original\", to... | 5.5 |
SYv9b4juom | [
"strength: The paper provides a fresh perspective on token selection by identifying a relationship between token importance and orthogonality to the sink token.",
"strength: By dynamically selecting only important tokens, OrthoRank effectively reduces the computational load without a significant accuracy drop.",
... | 5.25 |
SA19ijj44B | [
"strength: The methods under comparison are carefully selected to span a wide range of possible BNNs, and experiments are nicely designed to unveil specific insights about the relative strengths/weaknesses of different families of methods.",
"strength: I think some of the conclusions/insights from the empirical c... | 7.333333 |
2RNGX3iTr6 | [
"strength: The proposed method outperforms other methods on most datasets and metrics presented.",
"strength: Tabby achieves strong performance in benchmark evaluation. It generates high-quality synthethic tabular data in comparison with the baseline methods.",
"strength: The introduction of MoE shows effective... | 3 |
CRkvR8TJkk | [
"strength: New model for PFL",
"strength: The proposed algorithm outperforms existing works in the experiments.",
"strength: Modeling clients' objective functions as a composite of individual clients' objective functions is promising.",
"strength: The existence and uniqueness of a Nash equilibrium are provide... | 5 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.