Shu Jiang

Yuanchen Zhou, Shuo Jiang, Jie Zhu et al.

Process Reward Models (PRMs) have emerged as a promising framework for supervising intermediate reasoning in large language models (LLMs), yet existing PRMs are primarily trained on general or Science, Technology, Engineering, and Mathematics (STEM) domains and fall short in domain-specific contexts such as finance, where reasoning is more structured, symbolic, and sensitive to factual and regulatory correctness. We introduce \textbf{Fin-PRM}, a domain-specialized, trajectory-aware PRM tailored to evaluate intermediate reasoning steps in financial tasks. Fin-PRM integrates step-level and trajectory-level reward supervision, enabling fine-grained evaluation of reasoning traces aligned with financial logic. We apply Fin-PRM in both offline and online reward learning settings, supporting three key applications: (i) selecting high-quality reasoning trajectories for distillation-based supervised fine-tuning, (ii) providing dense process-level rewards for reinforcement learning, and (iii) guiding reward-informed Best-of-N inference at test time. Experimental results on financial reasoning benchmarks, including CFLUE and FinQA, demonstrate that Fin-PRM consistently outperforms general-purpose PRMs and strong domain baselines in trajectory selection quality. Downstream models trained with Fin-PRM yield substantial improvements with baselines, with gains of 12.9\% in supervised learning, 5.2\% in reinforcement learning, and 5.1\% in test-time performance. These findings highlight the value of domain-specialized reward modeling for aligning LLMs with expert-level financial reasoning. Our project resources will be available at https://github.com/aliyun/qwen-dianjin.

Shenghao Zhu, Yifei Chen, Weihong Chen et al.

Accurate brain tumor segmentation is essential for preoperative evaluation and personalized treatment. Multi-modal MRI is widely used due to its ability to capture complementary tumor features across different sequences. However, in clinical practice, missing modalities are common, limiting the robustness and generalizability of existing deep learning methods that rely on complete inputs, especially under non-dominant modality combinations. To address this, we propose AdaMM, a multi-modal brain tumor segmentation framework tailored for missing-modality scenarios, centered on knowledge distillation and composed of three synergistic modules. The Graph-guided Adaptive Refinement Module explicitly models semantic associations between generalizable and modality-specific features, enhancing adaptability to modality absence. The Bi-Bottleneck Distillation Module transfers structural and textural knowledge from teacher to student models via global style matching and adversarial feature alignment. The Lesion-Presence-Guided Reliability Module predicts prior probabilities of lesion types through an auxiliary classification task, effectively suppressing false positives under incomplete inputs. Extensive experiments on the BraTS 2018 and 2024 datasets demonstrate that AdaMM consistently outperforms existing methods, exhibiting superior segmentation accuracy and robustness, particularly in single-modality and weak-modality configurations. In addition, we conduct a systematic evaluation of six categories of missing-modality strategies, confirming the superiority of knowledge distillation and offering practical guidance for method selection and future research. Our source code is available at https://github.com/Quanato607/AdaMM.

Jiajing Lin, Zhenzhong Wang, Dejun Xu et al.

4D content generation aims to create dynamically evolving 3D content that responds to specific input objects such as images or 3D representations. Current approaches typically incorporate physical priors to animate 3D representations, but these methods suffer from significant limitations: they not only require users lacking physics expertise to manually specify material properties but also struggle to effectively handle the generation of multi-material composite objects. To address these challenges, we propose Phys4DGen, a novel 4D generation framework that integrates multi-material composition perception with physical simulation. The framework achieves automated, physically plausible 4D generation through three innovative modules: first, the 3D Material Grouping module partitions heterogeneous material regions on 3D representations' surfaces via semantic segmentation; second, the Internal Physical Structure Discovery module constructs the mechanical structure of object interiors; finally, we distill physical prior knowledge from multimodal large language models to enable rapid and automatic material properties identification for both objects' surfaces and interiors. Experiments on both synthetic and real-world datasets demonstrate that Phys4DGen can generate high-fidelity 4D content with physical realism in open-world scenarios, significantly outperforming state-of-the-art methods.

Jie Zhu, Yuanchen Zhou, Shuo Jiang et al.

Emotional Support Conversation (ESC) plays a vital role in alleviating psychological stress and providing emotional value through dialogue. While recent studies have largely focused on data augmentation and synthetic corpus construction, they often overlook the deeper cognitive reasoning processes that underpin effective emotional support. To address this gap, we propose \textbf{CARE}, a novel framework that strengthens reasoning in ESC without relying on large-scale synthetic data. CARE leverages the original ESC training set to guide models in generating logically coherent and supportive responses, thereby explicitly enhancing cognitive reasoning. Building on this foundation, we further employ reinforcement learning to refine and reinforce the reasoning process. Experimental results demonstrate that CARE significantly improves both the logical soundness and supportive quality of responses, advancing the development of empathetic, cognitively robust, and human-like emotional support systems.

Jiajing Lin, Shu Jiang, Qingyuan Zeng et al.

The intrinsic dynamics of an object governs its physical behavior in the real world, playing a critical role in enabling physically plausible interactive simulation with 3D assets. Existing methods have attempted to infer the intrinsic dynamics of objects from visual observations, but generally face two major challenges: one line of work relies on manually defined constitutive priors, making it difficult to generalize to complex scenarios; the other models intrinsic dynamics using neural networks, resulting in limited interpretability and poor generalization. To address these challenges, we propose VisionLaw, a bilevel optimization framework that infers interpretable expressions of intrinsic dynamics from visual observations. At the upper level, we introduce an LLMs-driven decoupled constitutive evolution strategy, where LLMs are prompted as a knowledgeable physics expert to generate and revise constitutive laws, with a built-in decoupling mechanism that substantially reduces the search complexity of LLMs. At the lower level, we introduce a vision-guided constitutive evaluation mechanism, which utilizes visual simulation to evaluate the consistency between the generated constitutive law and the underlying intrinsic dynamics, thereby guiding the upper-level evolution. Experiments on both synthetic and real-world datasets demonstrate that VisionLaw can effectively infer interpretable intrinsic dynamics from visual observations. It significantly outperforms existing state-of-the-art methods and exhibits strong generalization for interactive simulation in novel scenarios.

Qingyuan Zeng, Shu Jiang, Jiajing Lin et al.

With the rise of 3D Gaussian Splatting (3DGS), a variety of digital watermarking techniques, embedding either 1D bitstreams or 2D images, are used for copyright protection. However, the robustness of these watermarking techniques against potential attacks remains underexplored. This paper introduces the first universal black-box attack framework, the Group-based Multi-objective Evolutionary Attack (GMEA), designed to challenge these watermarking systems. We formulate the attack as a large-scale multi-objective optimization problem, balancing watermark removal with visual quality. In a black-box setting, we introduce an indirect objective function that blinds the watermark detector by minimizing the standard deviation of features extracted by a convolutional network, thus rendering the feature maps uninformative. To manage the vast search space of 3DGS models, we employ a group-based optimization strategy to partition the model into multiple, independent sub-optimization problems. Experiments demonstrate that our framework effectively removes both 1D and 2D watermarks from mainstream 3DGS watermarking methods while maintaining high visual fidelity. This work reveals critical vulnerabilities in existing 3DGS copyright protection schemes and calls for the development of more robust watermarking systems.

Jinyun Zhou, Rui Wang, Xu Liu et al.

We present a learning-based planner that aims to robustly drive a vehicle by mimicking human drivers' driving behavior. We leverage a mid-to-mid approach that allows us to manipulate the input to our imitation learning network freely. With that in mind, we propose a novel feedback synthesizer for data augmentation. It allows our agent to gain more driving experience in various previously unseen environments that are likely to encounter, thus improving overall performance. This is in contrast to prior works that rely purely on random synthesizers. Furthermore, rather than completely commit to imitating, we introduce task losses that penalize undesirable behaviors, such as collision, off-road, and so on. Unlike prior works, this is done by introducing a differentiable vehicle rasterizer that directly converts the waypoints output by the network into images. This effectively avoids the usage of heavyweight ConvLSTM networks, therefore, yields a faster model inference time. About the network architecture, we exploit an attention mechanism that allows the network to reason critical objects in the scene and produce better interpretable attention heatmaps. To further enhance the safety and robustness of the network, we add an optional optimization-based post-processing planner improving the driving comfort. We comprehensively validate our method's effectiveness in different scenarios that are specifically created for evaluating self-driving vehicles. Results demonstrate that our learning-based planner achieves high intelligence and can handle complex situations. Detailed ablation and visualization analysis are included to further demonstrate each of our proposed modules' effectiveness in our method.