arXiv subjects

Yuntao Li

Publications and source records attributed to Yuntao Li.

At least 19 recordsLinked to original sources

arXiv2025

Entropy Ratio Clipping as a Soft Global Constraint for Stable Reinforcement Learning

Large language model post-training relies on reinforcement learning to improve model capability and alignment quality. However, the off-policy training paradigm introduces distribution shift, which often pushes the policy beyond the trust region, leading to training instabilities manifested as fluctuations in policy entropy and unstable gradients. Although PPO-Clip mitigates this issue through importance clipping, it still overlooks the global distributional shift of actions. To address these challenges, we propose using the entropy ratio between the current and previous policies as a new global metric that effectively quantifies the relative change in policy exploration throughout updates. Building on this metric, we introduce an \textbf{Entropy Ratio Clipping} (ERC) mechanism that imposes bidirectional constraints on the entropy ratio. This stabilizes policy updates at the global distribution level and compensates for the inability of PPO-clip to regulate probability shifts of un-sampled actions. We integrate ERC into both DAPO and GPPO reinforcement learning algorithms. Experiments across multiple benchmarks show that ERC consistently improves performance.

cs.LG↗

arXiv2025

Structural Reward Model: Enhancing Interpretability, Efficiency, and Scalability in Reward Modeling

Reward Models (RMs) are key components for evaluating and guiding language model outputs. However, traditional scalar RMs often struggle with incorporating contextual and background information during inference, leading to incomplete evaluations. Generative RMs (GRMs) attempt to address these limitations by generating intermediate reasoning steps. Yet, their uncontrolled black-box nature and inefficiency due to sequential decoding hinder their industrial deployment. Industrial scenarios, such as search and recommendation systems, often involve single-domain tasks requiring evaluation along specific dimensions. In such contexts, diagnosing "bad cases" necessitates structured feedback to identify and optimize dimension-specific issues. In this paper, we propose the Structural Reward Model (SRM), a modular and interpretable framework integrating side-branch models as auxiliary feature generators. By introducing fine-grained dimensions, SRMs enable interpretable and efficient evaluation, facilitating targeted diagnostics and optimization. This structured approach ensures adaptability and scalability for industrial applications. Through comprehensive experiments, we demonstrate that SRMs outperform scalar RMs and GRMs in robustness and alignment with human preferences. The modular design further supports efficient optimization for practical scenarios, allowing SRM to provide a practical reward modeling solution for industry.

cs.AI↗

arXiv2025

CE-GPPO: Coordinating Entropy via Gradient-Preserving Clipping Policy Optimization in Reinforcement Learning

Reinforcement learning (RL) has become a powerful paradigm for optimizing large language models (LLMs) to handle complex reasoning tasks. A core challenge in this process lies in managing policy entropy, which reflects the balance between exploration and exploitation during training. Existing methods, such as proximal policy optimization (PPO) and its variants, discard valuable gradient signals from low-probability tokens due to the clipping mechanism. We systematically analyze the entropy dynamics and reveal that these clipped tokens play a critical yet overlooked role in regulating entropy evolution. We propose \textbf{C}oordinating \textbf{E}ntropy via \textbf{G}radient-\textbf{P}reserving \textbf{P}olicy \textbf{O}ptimization (CE-GPPO), a novel algorithm that reintroduces gradients from clipped tokens in native PPO in a gentle and bounded manner. By controlling the magnitude of gradients from tokens outside the clipping interval, CE-GPPO is able to achieve an exploration-exploitation trade-off. We provide theoretical justification and empirical evidence showing that CE-GPPO effectively mitigates entropy instability. Extensive experiments on mathematical reasoning benchmarks show that CE-GPPO consistently outperforms strong baselines across different model scales.

cs.LG↗

arXiv2025

AdaCoT: Pareto-Optimal Adaptive Chain-of-Thought Triggering via Reinforcement Learning

Large Language Models (LLMs) have demonstrated remarkable capabilities but often face challenges with tasks requiring sophisticated reasoning. While Chain-of-Thought (CoT) prompting significantly enhances reasoning, it indiscriminately generates lengthy reasoning steps for all queries, leading to substantial computational costs and inefficiency, especially for simpler inputs. To address this critical issue, we introduce AdaCoT (Adaptive Chain-of-Thought), a novel framework enabling LLMs to adaptively decide when to invoke CoT. AdaCoT framed adaptive reasoning as a Pareto optimization problem that seeks to balance model performance with the costs associated with CoT invocation (both frequency and computational overhead). We propose a reinforcement learning (RL) based method, specifically utilizing Proximal Policy Optimization (PPO), to dynamically control the CoT triggering decision boundary by adjusting penalty coefficients, thereby allowing the model to determine CoT necessity based on implicit query complexity. A key technical contribution is Selective Loss Masking (SLM), designed to counteract decision boundary collapse during multi-stage RL training, ensuring robust and stable adaptive triggering. Experimental results demonstrate that AdaCoT successfully navigates the Pareto frontier, achieving substantial reductions in CoT usage for queries not requiring elaborate reasoning. For instance, on our production traffic testset, AdaCoT reduced CoT triggering rates to as low as 3.18\% and decreased average response tokens by 69.06%, while maintaining high performance on complex tasks.

cs.LG↗

arXiv2025

Seed1.5-Thinking: Advancing Superb Reasoning Models with Reinforcement Learning

We introduce Seed1.5-Thinking, capable of reasoning through thinking before responding, resulting in improved performance on a wide range of benchmarks. Seed1.5-Thinking achieves 86.7 on AIME 2024, 55.0 on Codeforces and 77.3 on GPQA, demonstrating excellent reasoning abilities in STEM and coding. Beyond reasoning tasks, the method demonstrates notable generalization across diverse domains. For instance, it surpasses DeepSeek R1 by 8% in win rate on non-reasoning tasks, indicating its broader applicability. Compared to other state-of-the-art reasoning models, Seed1.5-Thinking is a Mixture-of-Experts (MoE) model with a relatively small size, featuring 20B activated and 200B total parameters. As part of our effort to assess generalized reasoning, we develop two internal benchmarks, BeyondAIME and Codeforces, both of which will be publicly released to support future research. Model trial link: https://www.volcengine.com/experience/ark.

cs.CL↗

arXiv2025

A Visual Reinforcement Learning-Based Separate Primitive Policy for Peg-in-Hole Tasks

For peg-in-hole tasks, humans rely on binocular visual perception to locate the peg above the hole surface and then proceed with insertion. This paper draws insights from this behavior to enable agents to learn efficient assembly strategies through visual reinforcement learning. Hence, we propose a Separate Primitive Policy (S2P) to learn how to derive location and insertion actions simultaneously. S2P is compatible with model-free reinforcement learning algorithms. Ten insertion tasks featuring different polygons are developed as benchmarks for evaluations. Simulation experiments show that S2P can boost the sample efficiency and success rate even with force constraints. Real-world experiments are also performed to verify the feasibility of S2P. Ablations are finally given to discuss the generalizability of S2P and some factors that affect its performance.

cs.RO↗

arXiv2024

Open-Source Reinforcement Learning Environments Implemented in MuJoCo with Franka Manipulator

This paper presents three open-source reinforcement learning environments developed on the MuJoCo physics engine with the Franka Emika Panda arm in MuJoCo Menagerie. Three representative tasks, push, slide, and pick-and-place, are implemented through the Gymnasium Robotics API, which inherits from the core of Gymnasium. Both the sparse binary and dense rewards are supported, and the observation space contains the keys of desired and achieved goals to follow the Multi-Goal Reinforcement Learning framework. Three different off-policy algorithms are used to validate the simulation attributes to ensure the fidelity of all tasks, and benchmark results are also given. Each environment and task are defined in a clean way, and the main parameters for modifying the environment are preserved to reflect the main difference. The repository, including all environments, is available at https://github.com/zichunxx/panda_mujoco_gym.

cs.RO↗

arXiv2024

Transformer-Enhanced Motion Planner: Attention-Guided Sampling for State-Specific Decision Making

Sampling-based motion planning (SBMP) algorithms are renowned for their robust global search capabilities. However, the inherent randomness in their sampling mechanisms often result in inconsistent path quality and limited search efficiency. In response to these challenges, this work proposes a novel deep learning-based motion planning framework, named Transformer-Enhanced Motion Planner (TEMP), which synergizes an Environmental Information Semantic Encoder (EISE) with a Motion Planning Transformer (MPT). EISE converts environmental data into semantic environmental information (SEI), providing MPT with an enriched environmental comprehension. MPT leverages an attention mechanism to dynamically recalibrate its focus on SEI, task objectives, and historical planning data, refining the sampling node generation. To demonstrate the capabilities of TEMP, we train our model using a dataset comprised of planning results produced by the RRT*. EISE and MPT are collaboratively trained, enabling EISE to autonomously learn and extract patterns from environmental data, thereby forming semantic representations that MPT could more effectively interpret and utilize for motion planning. Subsequently, we conducted a systematic evaluation of TEMP's efficacy across diverse task dimensions, which demonstrates that TEMP achieves exceptional performance metrics and a heightened degree of generalizability compared to state-of-the-art SBMPs.

cs.RO↗

arXiv2023

Heterophily-Based Graph Neural Network for Imbalanced Classification

Graph neural networks (GNNs) have shown promise in addressing graph-related problems, including node classification. However, conventional GNNs assume an even distribution of data across classes, which is often not the case in real-world scenarios, where certain classes are severely underrepresented. This leads to suboptimal performance of standard GNNs on imbalanced graphs. In this paper, we introduce a unique approach that tackles imbalanced classification on graphs by considering graph heterophily. We investigate the intricate relationship between class imbalance and graph heterophily, revealing that minority classes not only exhibit a scarcity of samples but also manifest lower levels of homophily, facilitating the propagation of erroneous information among neighboring nodes. Drawing upon this insight, we propose an efficient method, called Fast Im-GBK, which integrates an imbalance classification strategy with heterophily-aware GNNs to effectively address the class imbalance problem while significantly reducing training time. Our experiments on real-world graphs demonstrate our model's superiority in classification performance and efficiency for node classification tasks compared to existing baselines.

cs.LG↗

arXiv2023

Improving Semantic Matching through Dependency-Enhanced Pre-trained Model with Adaptive Fusion

Transformer-based pre-trained models like BERT have achieved great progress on Semantic Sentence Matching. Meanwhile, dependency prior knowledge has also shown general benefits in multiple NLP tasks. However, how to efficiently integrate dependency prior structure into pre-trained models to better model complex semantic matching relations is still unsettled. In this paper, we propose the \textbf{D}ependency-Enhanced \textbf{A}daptive \textbf{F}usion \textbf{A}ttention (\textbf{DAFA}), which explicitly introduces dependency structure into pre-trained models and adaptively fuses it with semantic information. Specifically, \textbf{\emph{(i)}} DAFA first proposes a structure-sensitive paradigm to construct a dependency matrix for calibrating attention weights. It adopts an adaptive fusion module to integrate the obtained dependency information and the original semantic signals. Moreover, DAFA reconstructs the attention calculation flow and provides better interpretability. By applying it on BERT, our method achieves state-of-the-art or competitive performance on 10 public datasets, demonstrating the benefits of adaptively fusing dependency structure in semantic matching task.

cs.CL↗

arXiv2023

T5-SR: A Unified Seq-to-Seq Decoding Strategy for Semantic Parsing

Translating natural language queries into SQLs in a seq2seq manner has attracted much attention recently. However, compared with abstract-syntactic-tree-based SQL generation, seq2seq semantic parsers face much more challenges, including poor quality on schematical information prediction and poor semantic coherence between natural language queries and SQLs. This paper analyses the above difficulties and proposes a seq2seq-oriented decoding strategy called SR, which includes a new intermediate representation SSQL and a reranking method with score re-estimator to solve the above obstacles respectively. Experimental results demonstrate the effectiveness of our proposed techniques and T5-SR-3b achieves new state-of-the-art results on the Spider dataset.

cs.CL↗

arXiv2023

DABERT: Dual Attention Enhanced BERT for Semantic Matching

Transformer-based pre-trained language models such as BERT have achieved remarkable results in Semantic Sentence Matching. However, existing models still suffer from insufficient ability to capture subtle differences. Minor noise like word addition, deletion, and modification of sentences may cause flipped predictions. To alleviate this problem, we propose a novel Dual Attention Enhanced BERT (DABERT) to enhance the ability of BERT to capture fine-grained differences in sentence pairs. DABERT comprises (1) Dual Attention module, which measures soft word matches by introducing a new dual channel alignment mechanism to model affinity and difference attention. (2) Adaptive Fusion module, this module uses attention to learn the aggregation of difference and affinity features, and generates a vector describing the matching details of sentence pairs. We conduct extensive experiments on well-studied semantic matching and robustness test datasets, and the experimental results show the effectiveness of our proposed method.

cs.CL↗

arXiv2022

PetLock:A Genderless and Standard Interface for the Future On-orbit Construction

Modular design is the foundation of on orbit construction technology of large space facilities in the future.Standard interface is the key technology of modular design of the future space robotic systems and space facilities.This paper presents the designed and tested of PetLock,a standard and genderless interface which can transfer mechanical loads,power and data between the future modular space robotic manipulator and spacecraft.PetLock adopts a completely genderless design,including connection face,locking mechanism,data and power interface.The connection surface provides a large translation and rotation misalignment tolerance,due to its 120-degree symmetrical and 3D shape design.The locking mechanism features the three locking pins retraction structure design,which is simple and reliable.POGO pin connectors in the center of the interface provides the power and data transfer capabilities.Due to the advantages of high locking force,large tolerance,high reliability and low cost,PetLock has the very big application potential in future on orbit construction missions.

cs.RO↗

arXiv2022

A Combined Inverse Kinematics Algorithm Using FABRIK with Optimization

Forward and backward reaching inverse kinematics (FABRIK) is a heuristic inverse kinematics solver that is gradually applied to manipulators with the advantages of fast convergence and generating more realistic configurations. However, under the high error constraint, FABRIK exhibits unstable convergence behavior, which is unsatisfactory for the real-time motion planning of manipulators. In this paper, a novel inverse kinematics algorithm that combines FABRIK and the sequential quadratic programming (SQP) algorithm is presented, in which the joint angles deduced by FABRIK will be taken as the initial seed of the SQP algorithm to avoid getting stuck in local minima. The combined algorithm is evaluated with experiments, in which our algorithm can achieve higher success rates and faster solution times than FABRIK under the high error constraint. Furthermore, the combined algorithm can generate continuous trajectories for the UR5 and KUKA LBR IIWA 14 R820 manipulators in path tracking with no pose error and permitted position error of the end-effector.

cs.RO↗

arXiv2022

Let Me Check the Examples: Enhancing Demonstration Learning via Explicit Imitation

Demonstration learning aims to guide the prompt prediction via providing answered demonstrations in the few shot settings. Despite achieving promising results, existing work only concatenates the answered examples as demonstrations to the prompt template (including the raw context) without any additional operation, neglecting the prompt-demonstration dependencies. Besides, prior research found that randomly replacing the labels of demonstrations marginally hurts performance, illustrating that the model could not properly learn the knowledge brought by the demonstrations. Inspired by the human learning process, in this paper, we introduce Imitation DEMOnstration Learning (Imitation-Demo) to strengthen demonstration learning via explicitly imitating human review behaviour, which includes: (1) contrastive learning mechanism to concentrate on the similar demonstrations. (2) demonstration-label re-prediction method to consolidate known knowledge. Experiment results show that our proposed method achieves state-of-the-art performance on 11 out of 14 classification corpora. Further studies also prove that Imitation-Demo strengthen the association between prompt and demonstrations, which could provide the basis for exploring how demonstration learning works.

cs.LG↗

arXiv2022

Pay More Attention to History: A Context Modelling Strategy for Conversational Text-to-SQL

Conversational text-to-SQL aims at converting multi-turn natural language queries into their corresponding SQL (Structured Query Language) representations. One of the most intractable problems of conversational text-to-SQL is modelling the semantics of multi-turn queries and gathering the proper information required for the current query. This paper shows that explicitly modelling the semantic changes by adding each turn and the summarization of the whole context can bring better performance on converting conversational queries into SQLs. In particular, we propose two conversational modelling tasks in both turn grain and conversation grain. These two tasks simply work as auxiliary training tasks to help with multi-turn conversational semantic parsing. We conducted empirical studies and achieved new state-of-the-art results on the large-scale open-domain conversational text-to-SQL dataset. The results demonstrate that the proposed mechanism significantly improves the performance of multi-turn semantic parsing.

cs.CL↗

arXiv2021

Small Changes Make Big Differences: Improving Multi-turn Response Selection in Dialogue Systems via Fine-Grained Contrastive Learning

Retrieve-based dialogue response selection aims to find a proper response from a candidate set given a multi-turn context. Pre-trained language models (PLMs) based methods have yielded significant improvements on this task. The sequence representation plays a key role in the learning of matching degree between the dialogue context and the response. However, we observe that different context-response pairs sharing the same context always have a greater similarity in the sequence representations calculated by PLMs, which makes it hard to distinguish positive responses from negative ones. Motivated by this, we propose a novel \textbf{F}ine-\textbf{G}rained \textbf{C}ontrastive (FGC) learning method for the response selection task based on PLMs. This FGC learning strategy helps PLMs to generate more distinguishable matching representations of each dialogue at fine grains, and further make better predictions on choosing positive responses. Empirical studies on two benchmark datasets demonstrate that the proposed FGC learning method can generally and significantly improve the model performance of existing PLM-based matching models.

cs.CL↗

arXiv2021

r-GAT: Relational Graph Attention Network for Multi-Relational Graphs

Graph Attention Network (GAT) focuses on modelling simple undirected and single relational graph data only. This limits its ability to deal with more general and complex multi-relational graphs that contain entities with directed links of different labels (e.g., knowledge graphs). Therefore, directly applying GAT on multi-relational graphs leads to sub-optimal solutions. To tackle this issue, we propose r-GAT, a relational graph attention network to learn multi-channel entity representations. Specifically, each channel corresponds to a latent semantic aspect of an entity. This enables us to aggregate neighborhood information for the current aspect using relation features. We further propose a query-aware attention mechanism for subsequent tasks to select useful aspects. Extensive experiments on link prediction and entity classification tasks show that our r-GAT can model multi-relational graphs effectively. Also, we show the interpretability of our approach by case study.

cs.AI↗