Weida Wang

Yanchao Xu, Wenbo Shao, Jun Li et al.

Intersection is one of the most challenging scenarios for autonomous driving tasks. Due to the complexity and stochasticity, essential applications (e.g., behavior modeling, motion prediction, safety validation, etc.) at intersections rely heavily on data-driven techniques. Thus, there is an intense demand for trajectory datasets of traffic participants (TPs) in intersections. Currently, most intersections in urban areas are equipped with traffic lights. However, there is not yet a large-scale, high-quality, publicly available trajectory dataset for signalized intersections. Therefore, in this paper, a typical two-phase signalized intersection is selected in Tianjin, China. Besides, a pipeline is designed to construct a Signalized INtersection Dataset (SIND), which contains 7 hours of recording including over 13,000 TPs with 7 types. Then, the behaviors of traffic light violations in SIND are recorded. Furthermore, the SIND is also compared with other similar works. The features of the SIND can be summarized as follows: 1) SIND provides more comprehensive information, including traffic light states, motion parameters, High Definition (HD) map, etc. 2) The category of TPs is diverse and characteristic, where the proportion of vulnerable road users (VRUs) is up to 62.6% 3) Multiple traffic light violations of non-motor vehicles are shown. We believe that SIND would be an effective supplement to existing datasets and can promote related research on autonomous driving.The dataset is available online via: https://github.com/SOTIF-AVLab/SinD

Wenbo Shao, Yanchao Xu, Jun Li et al.

An accurate trajectory prediction is crucial for safe and efficient autonomous driving in complex traffic environments. In recent years, artificial intelligence has shown strong capabilities in improving prediction accuracy. However, its characteristics of inexplicability and uncertainty make it challenging to determine the traffic environmental effect on prediction explicitly, posing significant challenges to safety-critical decision-making. To address these challenges, this study proposes a trajectory prediction framework with the epistemic uncertainty estimation ability that outputs high uncertainty when confronting unforeseeable or unknown scenarios. The proposed framework is used to analyze the environmental effect on the prediction algorithm performance. In the analysis, the traffic environment is considered in terms of scenario features and shifts, respectively, where features are divided into kinematic features of a target agent, features of its surrounding traffic participants, and other features. In addition, feature correlation and importance analyses are performed to study the above features' influence on the prediction error and epistemic uncertainty. Further, a cross-dataset case study is conducted using multiple intersection datasets to investigate the impact of unavoidable distributional shifts in the real world on trajectory prediction. The results indicate that the deep ensemble-based method has advantages in improving prediction robustness and estimating epistemic uncertainty. The consistent conclusions are obtained by the feature correlation and importance analyses, including the conclusion that kinematic features of the target agent have relatively strong effects on the prediction error and epistemic uncertainty. Furthermore, the prediction failure caused by distributional shifts and the potential of the deep ensemble-based method are analyzed.

Chao Xu, Yang Liu, Jiazheng Xing et al. · cmu, uw

In this paper, we abstract the process of people hearing speech, extracting meaningful cues, and creating various dynamically audio-consistent talking faces, termed Listening and Imagining, into the task of high-fidelity diverse talking faces generation from a single audio. Specifically, it involves two critical challenges: one is to effectively decouple identity, content, and emotion from entangled audio, and the other is to maintain intra-video diversity and inter-video consistency. To tackle the issues, we first dig out the intricate relationships among facial factors and simplify the decoupling process, tailoring a Progressive Audio Disentanglement for accurate facial geometry and semantics learning, where each stage incorporates a customized training module responsible for a specific factor. Secondly, to achieve visually diverse and audio-synchronized animation solely from input audio within a single model, we introduce the Controllable Coherent Frame generation, which involves the flexible integration of three trainable adapters with frozen Latent Diffusion Models (LDMs) to focus on maintaining facial geometry and semantics, as well as texture and temporal coherence between frames. In this way, we inherit high-quality diverse generation from LDMs while significantly improving their controllability at a low training cost. Extensive experiments demonstrate the flexibility and effectiveness of our method in handling this paradigm. The codes will be released at https://github.com/modelscope/facechain.

Ziyang Wu, Tianjiao Ding, Yifu Lu et al.

The attention operator is arguably the key distinguishing factor of transformer architectures, which have demonstrated state-of-the-art performance on a variety of tasks. However, transformer attention operators often impose a significant computational burden, with the computational complexity scaling quadratically with the number of tokens. In this work, we propose a novel transformer attention operator whose computational complexity scales linearly with the number of tokens. We derive our network architecture by extending prior work which has shown that a transformer style architecture naturally arises by "white-box" architecture design, where each layer of the network is designed to implement an incremental optimization step of a maximal coding rate reduction objective (MCR$^2$). Specifically, we derive a novel variational form of the MCR$^2$ objective and show that the architecture that results from unrolled gradient descent of this variational objective leads to a new attention module called Token Statistics Self-Attention (TSSA). TSSA has linear computational and memory complexity and radically departs from the typical attention architecture that computes pairwise similarities between tokens. Experiments on vision, language, and long sequence tasks show that simply swapping TSSA for standard self-attention, which we refer to as the Token Statistics Transformer (ToST), achieves competitive performance with conventional transformers while being significantly more computationally efficient and interpretable. Our results also somewhat call into question the conventional wisdom that pairwise similarity style attention mechanisms are critical to the success of transformer architectures. Code will be available at https://github.com/RobinWu218/ToST.

Di Zhang, Junxian Li, Jingdi Lei et al.

Vision-language models (VLMs) have shown remarkable advancements in multimodal reasoning tasks. However, they still often generate inaccurate or irrelevant responses due to issues like hallucinated image understandings or unrefined reasoning paths. To address these challenges, we introduce Critic-V, a novel framework inspired by the Actor-Critic paradigm to boost the reasoning capability of VLMs. This framework decouples the reasoning process and critic process by integrating two independent components: the Reasoner, which generates reasoning paths based on visual and textual inputs, and the Critic, which provides constructive critique to refine these paths. In this approach, the Reasoner generates reasoning responses according to text prompts, which can evolve iteratively as a policy based on feedback from the Critic. This interaction process was theoretically driven by a reinforcement learning framework where the Critic offers natural language critiques instead of scalar rewards, enabling more nuanced feedback to boost the Reasoner's capability on complex reasoning tasks. The Critic model is trained using Direct Preference Optimization (DPO), leveraging a preference dataset of critiques ranked by Rule-based Reward~(RBR) to enhance its critic capabilities. Evaluation results show that the Critic-V framework significantly outperforms existing methods, including GPT-4V, on 5 out of 8 benchmarks, especially regarding reasoning accuracy and efficiency. Combining a dynamic text-based policy for the Reasoner and constructive feedback from the preference-optimized Critic enables a more reliable and context-sensitive multimodal reasoning process. Our approach provides a promising solution to enhance the reliability of VLMs, improving their performance in real-world reasoning-heavy multimodal applications such as autonomous driving and embodied intelligence.

Yanru Wu, Jianning Wang, Weida Wang et al.

Multi-source transfer learning is an effective solution to data scarcity by utilizing multiple source tasks for the learning of the target task. However, access to source data and model details is limited in the era of commercial models, giving rise to the setting of multi-source-free (MSF) transfer learning that aims to leverage source domain knowledge without such access. As a newly defined problem paradigm, MSF transfer learning remains largely underexplored and not clearly formulated. In this work, we adopt an information theoretic perspective on it and propose a framework named H-ensemble, which dynamically learns the optimal linear combination, or ensemble, of source models for the target task, using a generalization of maximal correlation regression. The ensemble weights are optimized by maximizing an information theoretic metric for transferability. Compared to previous works, H-ensemble is characterized by: 1) its adaptability to a novel and realistic MSF setting for few-shot target tasks, 2) theoretical reliability, 3) a lightweight structure easy to interpret and adapt. Our method is empirically validated by ablation studies, along with extensive comparative analysis with other task ensemble and transfer learning methods. We show that the H-ensemble can successfully learn the optimal task ensemble, as well as outperform prior arts.

Zhihao Dou, Qinjian Zhao, Zhongwei Wan et al.

Large language models (LLMs) have demonstrated remarkable reasoning abilities in complex tasks, often relying on Chain-of-Thought (CoT) reasoning. However, due to their autoregressive token-level generation, the reasoning process is largely constrained to local decision-making and lacks global planning. This limitation frequently results in redundant, incoherent, or inaccurate reasoning, which significantly degrades overall performance. Existing approaches, such as tree-based algorithms and reinforcement learning (RL), attempt to address this issue but suffer from high computational costs and often fail to produce optimal reasoning trajectories. To tackle this challenge, we propose Plan-Then-Action Enhanced Reasoning with Group Relative Policy Optimization PTA-GRPO, a two-stage framework designed to improve both high-level planning and fine-grained CoT reasoning. In the first stage, we leverage advanced LLMs to distill CoT into compact high-level guidance, which is then used for supervised fine-tuning (SFT). In the second stage, we introduce a guidance-aware RL method that jointly optimizes the final output and the quality of high-level guidance, thereby enhancing reasoning effectiveness. We conduct extensive experiments on multiple mathematical reasoning benchmarks, including MATH, AIME2024, AIME2025, and AMC, across diverse base models such as Qwen2.5-7B-Instruct, Qwen3-8B, Qwen3-14B, and LLaMA3.2-3B. Experimental results demonstrate that PTA-GRPO consistently achieves stable and significant improvements across different models and tasks, validating its effectiveness and generalization.

Jiatong Li, Weida Wang, Qinggang Zhang et al.

Large language models (LLMs), especially Explicit Long Chain-of-Thought (CoT) reasoning models like DeepSeek-R1 and QWQ, have demonstrated powerful reasoning capabilities, achieving impressive performance in commonsense reasoning and mathematical inference. Despite their effectiveness, Long-CoT reasoning models are often criticized for their limited ability and low efficiency in knowledge-intensive domains such as molecule discovery. Success in this field requires a precise understanding of domain knowledge, including molecular structures and chemical principles, which is challenging due to the inherent complexity of molecular data and the scarcity of high-quality expert annotations. To bridge this gap, we introduce Mol-R1, a novel framework designed to improve explainability and reasoning performance of R1-like Explicit Long-CoT reasoning LLMs in text-based molecule generation. Our approach begins with a high-quality reasoning dataset curated through Prior Regulation via In-context Distillation (PRID), a dedicated distillation strategy to effectively generate paired reasoning traces guided by prior regulations. Building upon this, we introduce MoIA, Molecular Iterative Adaptation, a sophisticated training strategy that iteratively combines Supervised Fine-tuning (SFT) with Reinforced Policy Optimization (RPO), tailored to boost the reasoning performance of R1-like reasoning models for molecule discovery. Finally, we examine the performance of Mol-R1 in the text-based molecule reasoning generation task, showing superior performance against existing baselines.

Di Zhang, Weida Wang, Junxian Li et al.

This paper target in addressing the challenges of underthinking and overthinking in long chain-of-thought (CoT) reasoning for Large Reasoning Models (LRMs) by introducing Reasoning Control Fields (RCF)--a novel test-time approach that injects structured control signals to guide reasoning from a tree search perspective. RCF enables models to adjust reasoning effort according to given control conditions when solving complex tasks. Additionally, we present the Control-R-4K dataset, which consists of challenging problems annotated with detailed reasoning processes and corresponding control fields. To further enhance reasoning control, we propose a Conditional Distillation Finetuning (CDF) method, which trains model--particularly Control-R-32B--to effectively adjust reasoning effort during test time. Experimental results on benchmarks such as AIME2024 and MATH500 demonstrate that our approach achieves state-of-the-art performance at the 32B scale while enabling a controllable Long CoT reasoning process (L-CoT). Overall, this work introduces an effective paradigm for controllable test-time scaling reasoning.