Yifei Zhou

Yuda Song, Yifei Zhou, Ayush Sekhari et al. · cmu

We consider a hybrid reinforcement learning setting (Hybrid RL), in which an agent has access to an offline dataset and the ability to collect experience via real-world online interaction. The framework mitigates the challenges that arise in both pure offline and online RL settings, allowing for the design of simple and highly effective algorithms, in both theory and practice. We demonstrate these advantages by adapting the classical Q learning/iteration algorithm to the hybrid setting, which we call Hybrid Q-Learning or Hy-Q. In our theoretical results, we prove that the algorithm is both computationally and statistically efficient whenever the offline dataset supports a high-quality policy and the environment has bounded bilinear rank. Notably, we require no assumptions on the coverage provided by the initial distribution, in contrast with guarantees for policy gradient/iteration methods. In our experimental results, we show that Hy-Q with neural network function approximation outperforms state-of-the-art online, offline, and hybrid RL baselines on challenging benchmarks, including Montezuma's Revenge.

Yifei Zhou, Ayush Sekhari, Yuda Song et al. · cmu

Hybrid RL is the setting where an RL agent has access to both offline data and online data by interacting with the real-world environment. In this work, we propose a new hybrid RL algorithm that combines an on-policy actor-critic method with offline data. On-policy methods such as policy gradient and natural policy gradient (NPG) have shown to be more robust to model misspecification, though sometimes it may not be as sample efficient as methods that rely on off-policy learning. On the other hand, offline methods that depend on off-policy training often require strong assumptions in theory and are less stable to train in practice. Our new approach integrates a procedure of off-policy training on the offline data into an on-policy NPG framework. We show that our approach, in theory, can obtain a best-of-both-worlds type of result -- it achieves the state-of-art theoretical guarantees of offline RL when offline RL-specific assumptions hold, while at the same time maintaining the theoretical guarantees of on-policy NPG regardless of the offline RL assumptions' validity. Experimentally, in challenging rich-observation environments, we show that our approach outperforms a state-of-the-art hybrid RL baseline which only relies on off-policy policy optimization, demonstrating the empirical benefit of combining on-policy and off-policy learning. Our code is publicly available at https://github.com/YifeiZhou02/HNPG.

Yifei Zhou, Zilu Li, Abhinav Shrivastava et al.

Modern retrieval system often requires recomputing the representation of every piece of data in the gallery when updating to a better representation model. This process is known as backfilling and can be especially costly in the real world where the gallery often contains billions of samples. Recently, researchers have proposed the idea of Backward Compatible Training (BCT) where the new representation model can be trained with an auxiliary loss to make it backward compatible with the old representation. In this way, the new representation can be directly compared with the old representation, in principle avoiding the need for any backfilling. However, followup work shows that there is an inherent tradeoff where a backward compatible representation model cannot simultaneously maintain the performance of the new model itself. This paper reports our ``not-so-surprising'' finding that adding extra dimensions to the representation can help here. However, we also found that naively increasing the dimension of the representation did not work. To deal with this, we propose Backward-compatible Training with a novel Basis Transformation ($BT^2$). A basis transformation (BT) is basically a learnable set of parameters that applies an orthonormal transformation. Such a transformation possesses an important property whereby the original information contained in its input is retained in its output. We show in this paper how a BT can be utilized to add only the necessary amount of additional dimensions. We empirically verify the advantage of $BT^2$ over other state-of-the-art methods in a wide range of settings. We then further extend $BT^2$ to other challenging yet more practical settings, including significant change in model architecture (CNN to Transformers), modality change, and even a series of updates in the model architecture mimicking the evolution of deep learning models.

Yifei Zhou, Juntao Ren, Fengyu Li et al.

Advances in the field of vision-language contrastive learning have made it possible for many downstream applications to be carried out efficiently and accurately by simply taking the dot product between image and text representations. One of the most representative approaches proposed recently known as CLIP has garnered widespread adoption due to its effectiveness. CLIP is trained with an InfoNCE loss that takes into account both positive and negative samples to help learn a much more robust representation space. This paper reveals that the common downstream practice of taking a dot product is only a zeroth-order approximation of the optimization goal, resulting in a loss of information during test-time. Intuitively, since the model has been optimized based on the InfoNCE loss, test-time procedures should also be in alignment. The question lies in how one can retrieve any semblance of negative samples information during inference in a computationally efficient way. To this end, we propose Distribution Normalization (DN), where we approximate the mean representation of a batch of test samples and use such a mean to represent what would be analogous to negative samples in the InfoNCE loss. DN requires no retraining or fine-tuning and can be effortlessly applied during inference. Extensive experiments on a wide variety of downstream tasks exhibit a clear advantage of DN over the dot product on top of other existing test-time augmentation methods.

Yifei Zhou, Yansong Feng

Recent works show that discourse analysis benefits from modeling intra- and inter-sentential levels separately, where proper representations for text units of different granularities are desired to capture both the meaning of text units and their relations to the context. In this paper, we propose to take advantage of transformers to encode contextualized representations of units of different levels to dynamically capture the information required for discourse dependency analysis on intra- and inter-sentential levels. Motivated by the observation of writing patterns commonly shared across articles, we propose a novel method that treats discourse relation identification as a sequence labelling task, which takes advantage of structural information from the context of extracted discourse trees, and substantially outperforms traditional direct-classification methods. Experiments show that our model achieves state-of-the-art results on both English and Chinese datasets.

Yifei Zhou, Renyu Li, Hayden Housen et al.

Paraphrase Identification is a fundamental task in Natural Language Processing. While much progress has been made in the field, the performance of many state-of-the-art models often suffer from distribution shift during inference time. We verify that a major source of this performance drop comes from biases introduced by negative examples. To overcome these biases, we propose in this paper to train two separate models, one that only utilizes the positive pairs and the other the negative pairs. This enables us the option of deciding how much to utilize the negative model, for which we introduce a perplexity based out-of-distribution metric that we show can effectively and automatically determine how much weight it should be given during inference. We support our findings with strong empirical results.

Yifei Zhou, Andrea Zanette, Jiayi Pan et al. · berkeley

A broad use case of large language models (LLMs) is in goal-directed decision-making tasks (or "agent" tasks), where an LLM needs to not just generate completions for a given prompt, but rather make intelligent decisions over a multi-turn interaction to accomplish a task (e.g., when interacting with the web, using tools, or providing customer support). Reinforcement learning (RL) provides a general paradigm to address such agent tasks, but current RL methods for LLMs largely focus on optimizing single-turn rewards. By construction, most single-turn RL methods cannot endow LLMs with the ability to intelligently seek information over multiple turns, perform credit assignment, or reason about their past actions -- all of which are critical in agent tasks. This raises the question: how can we design effective and efficient multi-turn RL algorithms for LLMs? In this paper, we develop a framework for building multi-turn RL algorithms for fine-tuning LLMs, that preserves the flexibility of existing single-turn RL methods for LLMs (e.g., proximal policy optimization), while accommodating multiple turns, long horizons, and delayed rewards effectively. To do this, our framework adopts a hierarchical RL approach and runs two RL algorithms in parallel: a high-level off-policy value-based RL algorithm to aggregate reward over utterances, and a low-level RL algorithm that utilizes this high-level value function to train a token policy within each utterance or turn. Our hierarchical framework, Actor-Critic Framework with a Hierarchical Structure (ArCHer), can also give rise to other RL methods. Empirically, we find that ArCHer significantly improves efficiency and performance on agent tasks, attaining a sample efficiency of about 100x over existing methods, while also improving with larger model capacity (upto the 7 billion scale that we tested on).

Yifei Zhou, Song Jiang, Yuandong Tian et al. · meta-ai

Large language model (LLM) agents need to perform multi-turn interactions in real-world tasks. However, existing multi-turn RL algorithms for optimizing LLM agents fail to perform effective credit assignment over multiple turns while leveraging the generalization capabilities of LLMs and it remains unclear how to develop such algorithms. To study this, we first introduce a new benchmark, ColBench, where an LLM agent interacts with a human collaborator over multiple turns to solve realistic tasks in backend programming and frontend design. Building on this benchmark, we propose a novel RL algorithm, SWEET-RL (RL with Step-WisE Evaluation from Training-time information), that uses a carefully designed optimization objective to train a critic model with access to additional training-time information. The critic provides step-level rewards for improving the policy model. Our experiments demonstrate that SWEET-RL achieves a 6% absolute improvement in success and win rates on ColBench compared to other state-of-the-art multi-turn RL algorithms, enabling Llama-3.1-8B to match or exceed the performance of GPT4-o in realistic collaborative content creation.

Jiayi Pan, Xiuyu Li, Long Lian et al. · berkeley

Scaling inference-time computation has substantially improved the reasoning capabilities of language models. However, existing methods have significant limitations: serialized chain-of-thought approaches generate overly long outputs, leading to increased latency and exhausted context windows, while parallel methods such as self-consistency suffer from insufficient coordination, resulting in redundant computations and limited performance gains. To address these shortcomings, we propose Adaptive Parallel Reasoning (APR), a novel reasoning framework that enables language models to orchestrate both serialized and parallel computations end-to-end. APR generalizes existing reasoning methods by enabling adaptive multi-threaded inference using spawn() and join() operations. A key innovation is our end-to-end reinforcement learning strategy, optimizing both parent and child inference threads to enhance task success rate without requiring predefined reasoning structures. Experiments on the Countdown reasoning task demonstrate significant benefits of APR: (1) higher performance within the same context window (83.4% vs. 60.0% at 4k context); (2) superior scalability with increased computation (80.1% vs. 66.6% at 20k total tokens); (3) improved accuracy at equivalent latency (75.2% vs. 57.3% at approximately 5,000ms). APR represents a step towards enabling language models to autonomously optimize their reasoning processes through adaptive allocation of computation.

Yifei Zhou, Sergey Levine, Jason Weston et al.

Large language models are quickly becoming the foundation for intelligent agents that are capable of using tools. However, training such agents is challenging because it requires human creation and annotation of a diverse set of tasks, tools, and evaluation criteria. In this paper, we propose the Self-Challenging framework for training an agent on high-quality tasks that are generated by itself. The agent first plays the role of challenger and generates a task after interacting with the given tools. The tasks take the form of a novel general class of problems termed Code-as-Task, which are defined by an instruction, a verification function and solution and failure cases which serve as tests, allowing to filter only for high-quality tasks. The agent then takes an executor role and trains on those tasks with reinforcement learning using the evaluation feedback as a reward. Evaluation on two existing multi-turn tool-use agent benchmarks, M3ToolEval and TauBench, shows the Self-Challenging framework achieves over a two-fold improvement in Llama-3.1-8B-Instruct, despite using only self-generated training data.

Chang Peng, Yifei Zhou, Huifeng Xi et al.

Marker-free human pose estimation (HPE) has found increasing applications in various fields. Current HPE suffers from occasional errors in keypoint recognition and random fluctuation in keypoint trajectories when analyzing kinematic human poses. The performance of existing deep learning-based models for HPE refinement is considerably limited by inaccurate training datasets in which the keypoints are manually annotated. This paper proposed a novel method to overcome the difficulty through joint angle-based modeling. The key techniques include: (i) A joint angle-based model of human pose, which is robust to describe kinematic human poses; (ii) Approximating temporal variation of joint angles through high order Fourier series to get reliable "ground truth"; (iii) A bidirectional recurrent network is designed as a post-processing module to refine the estimation of well-established HRNet. Trained with the high-quality dataset constructed using our method, the network demonstrates outstanding performance to correct wrongly recognized joints and smooth their spatiotemporal trajectories. Tests show that joint angle-based refinement (JAR) outperforms the state-of-the-art HPE refinement network in challenging cases like figure skating and breaking.