论文检索

Deep Learning · Large Language Models

Lunjun Zhang, Arian Hosseini, Hritik Bansal, Seyed Mehran Kazemi, Aviral Kumar, Rishabh Agarwal

Verifiers or reward models are often used to enhance the reasoning performance of large language models (LLMs). A common approach is the Best-of-N method, where N candidate solutions generated by the LLM are ranked by a verifier, and the best one is selected. While LLM-based verifiers are typically trained as discriminative classifiers to score solutions, they do not utilize the text generation capabilities of pretrained LLMs. To overcome this limitation, we instead propose training verifiers using the ubiquitous next-token prediction objective, jointly on verification and solution generation. Compared to standard verifiers, such generative verifiers (GenRM) can benefit from several advantages of LLMs: they integrate seamlessly with instruction tuning, enable chain-of-thought reasoning, and can utilize additional test-time compute via majority voting for better verification. We demonstrate that GenRM outperforms discriminative, DPO verifiers, and LLM-as-a-Judge, resulting in large performance gains with Best-of-N, namely 5% → 45.3% on algorithmic tasks, 73% → 93.4% on GSM8K, and 28% →44.6% on easy-to-hard generalization on MATH. Furthermore, we find that training GenRM with synthetic verification rationales is sufficient to pick out subtle errors on math problems. Finally, we demonstrate that generative verifiers scale favorably with model size and inference-time compute.

Deep Learning · Large Language Models

Hanyang Zhao, Genta Winata, Anirban Das, Shi-Xiong Zhang, David Yao, Wenpin Tang, Sambit Sahu

Recently, numerous preference optimization algorithms have been introduced as extensions to the Direct Preference Optimization (DPO) family. While these methods have successfully aligned models with human preferences, there is a lack of understanding regarding the contributions of their additional components. Moreover, fair and consistent comparisons are scarce, making it difficult to discern which components genuinely enhance downstream performance. In this work, we propose RainbowPO, a unified framework that demystifies the effectiveness of existing DPO methods by categorizing their key components into seven broad directions. We integrate these components into a single cohesive objective, enhancing the performance of each individual element. Through extensive experiments, we demonstrate that RainbowPO outperforms existing DPO variants. Additionally, we provide insights to guide researchers in developing new DPO methods and assist practitioners in their implementations.

Deep Learning · Other Representation Learning

Shihao Zhang, Yuguang Yan, Angela Yao

For deep regression, preserving the ordinality of the targets with respect to the feature representation improves performance across various tasks. However, a theoretical explanation for the benefits of ordinality is still lacking. This work reveals that preserving ordinality reduces the conditional entropy $H(Z|Y)$ of representation $Z$ conditional on the target $Y$. However, our findings reveal that typical regression losses fail to sufficiently reduce $H(Z|Y)$, despite its crucial role in generalization performance. With this motivation, we introduce an optimal transport-based regularizer to preserve the similarity relationships of targets in the feature space to reduce $H(Z|Y)$. Additionally, we introduce a simple yet efficient strategy of duplicating the regressor targets, also with the aim of reducing $H(Z|Y)$. Experiments on three real-world regression tasks verify the effectiveness of our strategies to improve deep regression. Code: https://github.com/needylove/Regression_tightness

Applications · Robotics

Huan Ren, Wenfei Yang, Xiang Liu, Shifeng Zhang, Tianzhu Zhang

Category-level object pose estimation aims to determine the pose and size of novel objects in specific categories. Existing correspondence-based approaches typically adopt point-based representations to establish the correspondences between primitive observed points and normalized object coordinates. However, due to the inherent shape-dependence of canonical coordinates, these methods suffer from semantic incoherence across diverse object shapes. To resolve this issue, we innovatively leverage the sphere as a shared proxy shape of objects to learn shape-independent transformation via spherical representations. Based on this insight, we introduce a novel architecture called SpherePose, which yields precise correspondence prediction through three core designs. Firstly, We endow the point-wise feature extraction with SO(3)-invariance, which facilitates robust mapping between camera coordinate space and object coordinate space regardless of rotation transformation. Secondly, the spherical attention mechanism is designed to propagate and integrate features among spherical anchors from a comprehensive perspective, thus mitigating the interference of noise and incomplete point cloud. Lastly, a hyperbolic correspondence loss function is designed to distinguish subtle distinctions, which can promote the precision of correspondence prediction. Experimental results on CAMERA25, REAL275 and HouseCat6D benchmarks demonstrate the superior performance of our method, verifying the effectiveness of spherical representations and architectural innovations.

Applications · Computer Vision

Fei Wang, XINGYU FU, James Y. Huang, Zekun Li, Qin Liu, Xiaogeng Liu, Mingyu Derek Ma, Nan Xu, Wenxuan Zhou, Kai Zhang 等

We introduce MuirBench, a comprehensive benchmark that focuses on robust multi-image understanding capabilities of multimodal LLMs. MuirBench consists of 12 diverse multi-image tasks (e.g., scene understanding, ordering) that involve 10 categories of multi-image relations (e.g., multiview, temporal relations). Comprising 11,264 images and 2,600 multiple-choice questions, MuirBench is created in a pairwise manner, where each standard instance is paired with an unanswerable variant that has minimal semantic differences, in order for a reliable assessment. Evaluated upon 20 recent multi-modal LLMs, our results reveal that even the best-performing models like GPT-4o and Gemini Pro find it challenging to solve MuirBench, achieving 68.0% and 49.3% in accuracy. Open-source multimodal LLMs trained on single images can hardly generalize to multi-image questions, hovering below 33.3% in accuracy. These results highlight the importance of MuirBench in encouraging the community to develop multimodal LLMs that can look beyond a single image, suggesting potential pathways for future improvements.

Applications · Chemistry and Drug Discovery

Chenbin Zhang, Zhiqiang Hu, Jiang Chuchu, Wen Chen, JIE XU, Shaoting Zhang

Drug-target binding affinity prediction is a fundamental task for drug discovery. It has been extensively explored in literature and promising results are reported. However, in this paper, we demonstrate that the results may be misleading and cannot be well generalized to real practice. The core observation is that the canonical randomized split of a test set in conventional evaluation leaves the test set dominated by samples with high similarity to the training set. The performance of models is severely degraded on samples with lower similarity to the training set but the drawback is highly overlooked in current evaluation. As a result, the performance can hardly be trusted when the model meets low-similarity samples in real practice. To address this problem, we propose a framework of similarity aware evaluation in which a novel split methodology is proposed to adapt to any desired distribution. This is achieved by a formulation of optimization problems which are approximately and efficiently solved by gradient descent. We perform extensive experiments across five representative methods in four datasets for two typical target evaluations and compare them with various counterpart methods. Results demonstrate that the proposed split methodology can significantly better fit desired distributions and guide the development of models.

Applications · Chemistry and Drug Discovery

Chenbin Zhang, Zhiqiang Hu, Jiang Chuchu, Wen Chen, JIE XU, Shaoting Zhang

Drug-target binding affinity prediction is a fundamental task for drug discovery. It has been extensively explored in literature and promising results are reported. However, in this paper, we demonstrate that the results may be misleading and cannot be well generalized to real practice. The core observation is that the canonical randomized split of a test set in conventional evaluation leaves the test set dominated by samples with high similarity to the training set. The performance of models is severely degraded on samples with lower similarity to the training set but the drawback is highly overlooked in current evaluation. As a result, the performance can hardly be trusted when the model meets low-similarity samples in real practice. To address this problem, we propose a framework of similarity aware evaluation in which a novel split methodology is proposed to adapt to any desired distribution. This is achieved by a formulation of optimization problems which are approximately and efficiently solved by gradient descent. We perform extensive experiments across five representative methods in four datasets for two typical target evaluations and compare them with various counterpart methods. Results demonstrate that the proposed split methodology can significantly better fit desired distributions and guide the development of models.

Applications · Language, Speech and Dialog

Qingkai Fang, Shoutao Guo, Yan Zhou, Zhengrui Ma, Shaolei Zhang, Yang Feng

Models like GPT-4o enable real-time interaction with large language models (LLMs) through speech, significantly enhancing user experience compared to traditional text-based interaction. However, there is still a lack of exploration on how to build speech interaction models based on open-source LLMs. To address this, we propose LLaMA-Omni, a novel model architecture designed for low-latency and high-quality speech interaction with LLMs. LLaMA-Omni integrates a pretrained speech encoder, a speech adaptor, an LLM, and a streaming speech decoder. It eliminates the need for speech transcription, and can simultaneously generate text and speech responses directly from speech instructions with extremely low latency. We build our model based on the latest Llama-3.1-8B-Instruct model. To align the model with speech interaction scenarios, we construct a dataset named InstructS2S-200K, which includes 200K speech instructions and corresponding speech responses. Experimental results show that compared to previous speech-language models, LLaMA-Omni provides better responses in both content and style, with a response latency as low as 226ms. Additionally, training LLaMA-Omni takes less than 3 days on just 4 GPUs, paving the way for the efficient development of speech-language models in the future.

Applications · Computer Vision

Shaolei Zhang, Qingkai Fang, Yang, Yang Feng

The advent of real-time large multimodal models (LMMs) like GPT-4o has sparked considerable interest in efficient LMMs. LMM frameworks typically encode visual inputs into vision tokens (continuous representations) and integrate them and textual instructions into the context of large language models (LLMs), where large-scale parameters and numerous context tokens (predominantly vision tokens) result in substantial computational overhead. Previous efforts towards efficient LMMs always focus on replacing the LLM backbone with smaller models, while neglecting the crucial issue of token quantity. In this paper, we introduce LLaVA-Mini, an efficient LMM with minimal vision tokens. To achieve a high compression ratio of vision tokens while preserving visual information, we first analyze how LMMs understand vision tokens and find that most vision tokens only play a crucial role in the early layers of LLM backbone, where they mainly fuse visual information into text tokens. Building on this finding, LLaVA-Mini introduces modality pre-fusion to fuse visual information into text tokens in advance, thereby facilitating the extreme compression of vision tokens fed to LLM backbone into one token. LLaVA-Mini is a unified large multimodal model that can support the understanding of images, high-resolution images, and videos in an efficient manner. Experiments across 11 image-based and 7 video-based benchmarks demonstrate that LLaVA-Mini outperforms LLaVA-v1.5 with just 1 vision token instead of 576. Efficiency analyses reveal that LLaVA-Mini can reduce FLOPs by 77%, deliver low-latency responses within 40 milliseconds, and process over 10,000 frames of video on the GPU hardware with 24GB of memory.

Deep Learning · Large Language Models

Gaurav Sahu, Abhay Puri, Juan A. Rodriguez, Amirhossein Abaskohi, Mohammad Chegini, Alexandre Drouin, Perouz Taslakian, Valentina Zantedeschi, Alexandre Lacoste, David Vazquez 等

Data analytics is essential for extracting valuable insights from data that can assist organizations in making effective decisions. We introduce InsightBench, a benchmark dataset with three key features. First, it consists of 100 datasets representing diverse business use cases such as finance and incident management, each accompanied by a carefully curated set of insights planted in the datasets. Second, unlike existing benchmarks focusing on answering single queries, InsightBench evaluates agents based on their ability to perform end-to-end data analytics, including formulating questions, interpreting answers, and generating a summary of insights and actionable steps. Third, we conducted comprehensive quality assurance to ensure that each dataset in the benchmark had clear goals and included relevant and meaningful questions and analysis. Furthermore, we implement a two-way evaluation mechanism using LLaMA-3 as an effective, open-source evaluator to assess agents’ ability to extract insights. We also propose AgentPoirot, our baseline data analysis agent capable of performing end-to-end data analytics. Our evaluation on InsightBench shows that AgentPoirot outperforms existing approaches (such as Pandas Agent) that focus on resolving single queries. We also compare the performance of open- and closed-source LLMs and various evaluation strategies. Overall, this benchmark serves as a testbed to motivate further development in comprehensive automated data analytics and can be accessed here: https://github.com/ServiceNow/insight-bench.

Reinforcement Learning · Online

Shuze Liu, Claire Chen, Shangtong Zhang

Policy evaluation estimates the performance of a policy by (1) collecting data from the environment and (2) processing raw data into a meaningful estimate. Due to the sequential nature of reinforcement learning, any improper data-collecting policy or data-processing method substantially deteriorates the variance of evaluation results over long time steps. Thus, policy evaluation often suffers from large variance and requires massive data to achieve the desired accuracy. In this work, we design an optimal combination of data-collecting policy and data-processing baseline. Theoretically, we prove our doubly optimal policy evaluation method is unbiased and guaranteed to have lower variance than previously best-performing methods. Empirically, compared with previous works, we show our method reduces variance substantially and achieves superior empirical performance.

Deep Learning · Self-Supervised Learning

Jingyang Li, Jiachun Pan, Vincent Tan, Kim-chuan Toh, Pan Zhou

Semi-supervised learning (SSL), exemplified by FixMatch (Sohn et al., 2020), has shown significant generalization advantages over supervised learning (SL), particularly in the context of deep neural networks (DNNs). However, it is still unclear, from a theoretical standpoint, why FixMatch-like SSL algorithms generalize better than SL on DNNs. In this work, we present the first theoretical justification for the enhanced test accuracy observed in FixMatch-like SSL applied to DNNs by taking convolutional neural networks (CNNs) on classification tasks as an example. Our theoretical analysis reveals that the semantic feature learning processes in FixMatch and SL are rather different. In particular, FixMatch learns all the discriminative features of each semantic class, while SL only randomly captures a subset of features due to the well-known lottery ticket hypothesis. Furthermore, we show that our analysis framework can be applied to other FixMatch-like SSL methods, e.g., FlexMatch, FreeMatch, Dash, and SoftMatch. Inspired by our theoretical analysis, we develop an improved variant of FixMatch, termed Semantic-Aware FixMatch (SA-FixMatch). Experimental results corroborate our theoretical findings and the enhanced generalization capability of SA-FixMatch.

Deep Learning · Self-Supervised Learning

Jingyang Li, Jiachun Pan, Vincent Tan, Kim-chuan Toh, Pan Zhou

Semi-supervised learning (SSL), exemplified by FixMatch (Sohn et al., 2020), has shown significant generalization advantages over supervised learning (SL), particularly in the context of deep neural networks (DNNs). However, it is still unclear, from a theoretical standpoint, why FixMatch-like SSL algorithms generalize better than SL on DNNs. In this work, we present the first theoretical justification for the enhanced test accuracy observed in FixMatch-like SSL applied to DNNs by taking convolutional neural networks (CNNs) on classification tasks as an example. Our theoretical analysis reveals that the semantic feature learning processes in FixMatch and SL are rather different. In particular, FixMatch learns all the discriminative features of each semantic class, while SL only randomly captures a subset of features due to the well-known lottery ticket hypothesis. Furthermore, we show that our analysis framework can be applied to other FixMatch-like SSL methods, e.g., FlexMatch, FreeMatch, Dash, and SoftMatch. Inspired by our theoretical analysis, we develop an improved variant of FixMatch, termed Semantic-Aware FixMatch (SA-FixMatch). Experimental results corroborate our theoretical findings and the enhanced generalization capability of SA-FixMatch.

Theory · Reinforcement Learning and Planning

Xiaochi Qian, Shangtong Zhang

Off-policy learning enables a reinforcement learning (RL) agent to reason counterfactually about policies that are not executed and is one of the most important ideas in RL. It, however, can lead to instability when combined with function approximation and bootstrapping, two arguably indispensable ingredients for large-scale reinforcement learning. This is the notorious deadly triad. The seminal work Sutton et al. (2008) pioneers Gradient Temporal Difference learning (GTD) as the first solution to the deadly triad, which has enjoyed massive success thereafter. During the derivation of GTD, some intermediate algorithm, called $A^\top$TD, was invented but soon deemed inferior. In this paper, we revisit this $A^\top$TD and prove that a variant of $A^\top$TD, called $A_t^\top$TD, is also an effective solution to the deadly triad. Furthermore, this $A_t^\top$TD only needs one set of parameters and one learning rate. By contrast, GTD has two sets of parameters and two learning rates, making it hard to tune in practice. We provide asymptotic analysis for $A^\top_t$TD and finite sample analysis for a variant of $A^\top_t$TD that additionally involves a projection operator. The convergence rate of this variant is on par with the canonical on-policy temporal difference learning.

Deep Learning · Large Language Models

Samy Jelassi, Clara Mohri, David Brandfonbrener, Alex Gu, Nikhil Vyas, Nikhil Anand, David Alvarez-Melis, Yuanzhi Li, Sham Kakade, Eran Malach

The Mixture-of-Experts (MoE) architecture enables a significant increase in the total number of model parameters with minimal computational overhead. However, it is not clear what performance tradeoffs, if any, exist between MoEs and standard dense transformers.In this paper, we show that as we increase the number of experts (while fixing the number of active parameters), the memorization performance consistently increases while the reasoning capabilities saturate. We begin by analyzing the theoretical limitations of MoEs at reasoning. We prove that there exist graph problems that cannot be solved by any number of experts of a certain width; however, the same task can be easily solved by a dense model with a slightly larger width. On the other hand, we find that on memory-intensive tasks, MoEs can effectively leverage a small number of active parameters with a large number of experts to memorize the data. We empirically validate these findings on synthetic graph problems and memory-intensive closed book retrieval tasks. Lastly, we pre-train a series of MoEs and dense transformers and evaluate them on commonly used benchmarks in math and natural language. We find that increasing the number of experts helps solve knowledge-intensive tasks, but fails to yield the same benefits for reasoning tasks.

Deep Learning · Large Language Models

Rosie Zhao, Depen Morwani, David Brandfonbrener, Nikhil Vyas, Sham Kakade

Training language models becomes increasingly expensive with scale, prompting numerous attempts to improve optimization efficiency. Despite these efforts, the Adam optimizer remains the most widely used, due to a prevailing view that it is the most effective approach. We aim to compare several optimization algorithms, including SGD, Adafactor, Adam, Lion, and Sophia in the context of autoregressive language modeling across a range of model sizes, hyperparameters, and architecture variants. Our findings indicate that, except for SGD, these algorithms all perform comparably both in their optimal performance and also in terms of how they fare across a wide range of hyperparameter choices. Our results suggest to practitioners that the choice of optimizer can be guided by practical considerations like memory constraints and ease of implementation, as no single algorithm emerged as a clear winner in terms of performance or stability to hyperparameter misspecification. Given our findings, we further dissect these approaches, examining two simplified versions of Adam: a) signed momentum (Signum) which we see recovers both the performance and hyperparameter stability of Adam and b) Adalayer, a layerwise variant of Adam which we introduce to study the impact on Adam's preconditioning for different layers of the network. Examining Adalayer leads us to the conclusion that, perhaps surprisingly, adaptivity on *both* the last layer and LayerNorm parameters in particular are necessary for retaining performance and stability to learning rate.

Optimization · Global Optimization

Depen Morwani, Itai Shapira, Nikhil Vyas, Eran Malach, Sham Kakade, Lucas Janson

Shampoo, a second-order optimization algorithm that uses a Kronecker product preconditioner, has recently received increasing attention from the machine learning community. Despite the increasing popularity of Shampoo, the theoretical foundations of its effectiveness are not well understood. The preconditioner used by Shampoo can be viewed as either an approximation of the Gauss--Newton component of the Hessian or the covariance matrix of the gradients maintained by Adagrad. Our key contribution is providing an explicit and novel connection between the optimal Kronecker product approximation of these matrices and the approximationmade by Shampoo. Our connection highlights a subtle but common misconception about Shampoo’s approximation. In particular, the square of the approximation used by the Shampoo optimizer is equivalent to a single step of the poweriteration algorithm for computing the aforementioned optimal Kronecker product approximation. Across a variety of datasets and architectures we empiricallydemonstrate that this is close to the optimal Kronecker product approximation. We also study the impact of batch gradients and empirical Fisher on the quality of Hessian approximation. Our findings not only advance the theoretical understanding of Shampoo but also illuminate potential pathways for enhancing its practical performance.

Optimization · Large Scale, Parallel and Distributed

Hanlin Zhang, Depen Morwani, Nikhil Vyas, Jingfeng Wu, Difan Zou, Udaya Ghai, Dean Foster, Sham Kakade

Training large-scale models under given resources requires careful design of parallelism strategies. In particular, the efficiency notion of critical batch size (CBS), concerning the compromise between time and compute, marks the threshold beyond which greater data parallelism leads to diminishing returns. To operationalize it, we propose a measure of CBS and pre-train a series of auto-regressive language models, ranging from 85 million to 1.2 billion parameters, on the C4 dataset. Through extensive hyper-parameter sweeps and careful control of factors such as batch size, momentum, and learning rate along with its scheduling, we systematically investigate the impact of scale on CBS. Then we fit scaling laws with respect to model and data sizes to decouple their effects. Overall, our results demonstrate that CBS scales primarily with data size rather than model size, a finding we justify theoretically through the analysis of infinite-width limits of neural networks and infinite-dimensional least squares regression. Of independent interest, we highlight the importance of common hyper-parameter choices and strategies for studying large-scale pre-training beyond fixed training durations.

Deep Learning · Generative Models and Autoencoders

Shifeng Xu, Yanzhu Liu, Adams Kong

Recent research pinpoints that different diffusion methods and architectures trained on the same dataset produce similar results for the same input noise. This property suggests that they have some preferable noises for a given sample. By visualizing the noise-sample pairs of rectified flow models and stable diffusion models in two-dimensional spaces, we observe that the preferable paths, connecting preferable noises to the corresponding samples, are better organized with significant fewer crossings comparing with the random paths, connecting random noises to training samples. In high-dimensional space, paths rarely intersect. The path crossings in two-dimensional spaces indicate the shorter inter-path distance in the corresponding high-dimensional spaces. Inspired by this observation, we propose the Distance-Aware Noise-Sample Matching (DANSM) method to lengthen the inter-path distance for speeding up the model training. DANSM is derived from rectified flow models, which allow using a closed-form formula to calculate the inter-path distance. To further simplify the optimization, we derive the relationship between inter-path distance and path length, and use the latter in the optimization surrogate. DANSM is evaluated on both image and latent spaces by rectified flow models and diffusion models. The experimental results show that DANSM can significantly improve the training speed by 30\% $\sim$ 40\%without sacrificing the generation quality.

Junyu Zhang, Daochang Liu, Eunbyung Park, Shichao Zhang, Chang Xu

Diffusion models (DMs) have achieved record-breaking performance in image generation tasks.Nevertheless, in practice, the training-sampling discrepancy, caused by score estimation error and discretization error, limits the modeling ability of DMs, a phenomenon known as exposure bias.To alleviate such exposure bias and further improve the generative performance, we put forward a prompt learning framework built upon a lightweight prompt prediction model.Concretely, our model learns an anti-bias prompt for the generated sample at each sampling step, aiming to compensate for the exposure bias that arises.Following this design philosophy, our framework rectifies the sampling trajectory to match the training trajectory, thereby reducing the divergence between the target data distribution and the modeling distribution.To train the prompt prediction model, we simulate exposure bias by constructing training data and introduce a time-dependent weighting function for optimization.Empirical results on various DMs demonstrate the superiority of our prompt learning framework across three benchmark datasets.Importantly, the optimized prompt prediction model effectively improves image quality with only a 5\% increase in sampling overhead, which remains negligible.