Jie Zhu

Jie Zhu, Huaixia Dou, Shuo Jiang et al.

Existing emotional support conversation (ESC) systems mainly rely on end-to-end response generation or coarse strategy supervision, offering limited interpretability and little support for systematic skill improvement. We propose ESC-Skills, a skill-centric framework that discovers and self-evolves executable emotional support skills. We first model localized support interactions as Intervention Units (IUs), which capture state--action--outcome dynamics between seeker states, support interventions, and post-response emotional changes. Based on IUs extracted from both successful and failed ESC dialogues, we construct the ESC-Skills Bank, a repository of executable emotional support skills containing intervention guidance, applicability conditions, expected outcomes, and potential risks. To further improve robustness, we introduce a multi-profile self-evolutionary refinement framework in which an ESC agent interacts with diverse simulated seeker profiles under SAGE evaluation. The resulting interaction traces are analyzed to identify missing skills, unsafe interventions, and profile-specific failure patterns, which are then used to refine the Skills Bank through simulation-based verification. Experimental results demonstrate that ESC-Skills improves both response-level quality and dialogue-level emotional outcomes while providing more interpretable and controllable support behaviors. We will release the code, prompts, and ESC-Skills Bank at https://github.com/aliyun/qwen-dianjin.

Tianheng Xu, Runke Fan, Jie Zhu et al.

With the surging demand for ultra-reliable, low-latency, and ubiquitous connectivity in Sixth-Generation (6G) networks, Non-Terrestrial Networks (NTNs) emerge as a key complement to terrestrial networks by offering flexible access and global coverage. Despite the significant potential, NTNs still face critical challenges, including dynamic propagation environments, energy constraints, and dense interference. As a key 6G technology, Fluid Antennas (FAs) can reshape wireless channels by reconfiguring radiating elements within a limited space, such as their positions and rotations, to provide higher channel diversity and multiplexing gains. Compared to fixed-position antennas, FAs can present a promising integration path for NTNs to mitigate dynamic channel fading and optimize resource allocation. This paper provides a comprehensive review of FA-assisted NTNs. We begin with a brief overview of the classical structure and limitations of existing NTNs, the fundamentals and advantages of FAs, and the basic principles of FA-assisted NTNs. We then investigate the joint optimization solutions, detailing the adjustments of FA configurations, NTN platform motion modes, and resource allocations. We also discuss the combination with other emerging technologies and explore FA-assisted NTNs as a novel network architecture for intelligent function integrations. Furthermore, we delve into the physical layer security and covert communication in FA-assisted NTNs. Finally, we highlight the potential future directions to empower broader applications of FA-assisted NTNs.

Huaixia Dou, Jie Zhu, Minghao Wu et al.

As large language models (LLMs) are increasingly deployed in financial services, a single non-compliant interaction can expose institutions to regulatory penalties and direct consumer harm. Existing guard models are built around general harm taxonomies and overlook violations grounded in specific financial regulations. We address this gap with a regulation-driven pipeline that operates directly on regulatory documents, inducing a financial compliance risk taxonomy and synthesizing grounded training data without any predefined violation categories. Instantiating the pipeline on Chinese financial regulations, we release \textbf{FinGuard-Bench}, to our knowledge the first benchmark for financial regulatory compliance detection, with expert-annotated labels at both the query and response levels. We further train \textbf{FinGuard}, a financial compliance detection model built on Qwen3-8B and trained on the regulation-grounded data via supervised fine-tuning and self-play reinforcement learning. On FinGuard-Bench, FinGuard substantially outperforms all baselines, including dedicated guard models and much larger general-purpose LLMs such as Qwen3.5-397B-A17B and GPT-5.1. Furthermore, FinGuard also preserves general safety capabilities and adapts to unseen institution-specific policies using policy documents alone. We will publicly release the code, prompts, and resources used in this work on GitHub.

Chen Yang, Chufan Yu, Hanfu Chen et al.

MOSS-Audio is a unified audio-language model for speech, environmental sound, and music understanding, supporting audio captioning, time-aware question answering, timestamped transcription, and audio-grounded reasoning. MOSS-Audio couples a dedicated audio encoder with a modality adapter and a large language model: the encoder produces 12.5 Hz temporal representations, the adapter projects them into the decoder space, and the decoder generates autoregressive text outputs. Two design choices are central to the system: \textbf{DeepStack cross-layer feature injection}, which exposes the decoder to acoustic information from multiple encoder depths, and \textbf{time markers}, which provide explicit temporal cues by inserting timestamp markers into the audio-token stream. At the data level, we design an event-preserving audio annotation pipeline that segments raw audio at coherent event boundaries, applies branch-specific annotation to speech, music, and general audio, and merges the results into unified captions for pretraining. The intermediate branch-specific captions are further retained to support the construction of task-oriented SFT data. The model is pretrained on large-scale audio-language data, with time-aware objectives incorporated to support temporal grounding, and then undergoes multi-stage post-training to enhance instruction following and audio-grounded reasoning. We release 4B and 8B variants in both Instruct and Thinking configurations. MOSS-Audio achieves strong performance across general audio understanding, speech captioning, ASR, and timestamped ASR, positioning it as a promising understanding foundation for future voice agents.

Jie Zhu, Jiyang Qi, Mingyu Ding et al.

In this paper, we are interested in understanding self-supervised pretraining through studying the capability that self-supervised representation pretraining methods learn part-aware representations. The study is mainly motivated by that random views, used in contrastive learning, and random masked (visible) patches, used in masked image modeling, are often about object parts. We explain that contrastive learning is a part-to-whole task: the projection layer hallucinates the whole object representation from the object part representation learned from the encoder, and that masked image modeling is a part-to-part task: the masked patches of the object are hallucinated from the visible patches. The explanation suggests that the self-supervised pretrained encoder is required to understand the object part. We empirically compare the off-the-shelf encoders pretrained with several representative methods on object-level recognition and part-level recognition. The results show that the fully-supervised model outperforms self-supervised models for object-level recognition, and most self-supervised contrastive learning and masked image modeling methods outperform the fully-supervised method for part-level recognition. It is observed that the combination of contrastive learning and masked image modeling further improves the performance.

Kexin Huang, Liwei Fan, Botian Jiang et al.

Voice design from natural language aims to generate speaker timbres directly from free-form textual descriptions, allowing users to create voices tailored to specific roles, personalities, and emotions. Such controllable voice creation benefits a wide range of downstream applications-including storytelling, game dubbing, role-play agents, and conversational assistants, making it a significant task for modern Text-to-Speech models. However, existing models are largely trained on carefully recorded studio data, which produces speech that is clean and well-articulated, yet lacks the lived-in qualities of real human voices. To address these limitations, we present MOSS-VoiceGenerator, an open-source instruction-driven voice generation model that creates new timbres directly from natural language prompts. Motivated by the hypothesis that exposure to real-world acoustic variation produces more perceptually natural voices, we train on large-scale expressive speech data sourced from cinematic content. Subjective preference studies demonstrate its superiority in overall performance, instruction-following, and naturalness compared to other voice design models.

Haoran Ding, Wenlin Zhao, Yuchen Jiang et al.

Large recommendation models have demonstrated substantial potential gains under scaling laws, yet these gains are difficult to realize in industrial recommendation systems because real-world deployment requires lightweight models with strict serving efficiency and latency guarantees. This creates a fundamental gap between offline model scaling and online deployment. In this work, we present Rec-Distill, an industrial distillation pipeline that transfers the performance gains of large-scale recommendation modeling to efficient serving models. Rec-Distill combines large-teacher scaling with student-side transfer optimization through decoupled training, black-box distillation, debiasing mechanism, and a hybrid batch-streaming pipeline for dynamic recommendation environments. Across multiple recommendation and advertising scenarios on real-world platforms, our framework scales teacher models up to 24B dense parameters and 20K behavior sequence length, while enabling lightweight students to recover a substantial portion of teacher gains, with distillation transferability exceeding 60% in the best setting. Extensive offline and online experiments further show that these transferred gains consistently translate into measurable business improvements under industrial constraints. These results demonstrate that Rec-Distill provides a practical framework for distilling large-scale recommendation models into deployable, cost-efficient serving systems, while also establishing a reliable path toward scaling recommendation models to even larger regimes in the future.

Jie Zhu, Yuan Zong, Jingang Shi et al.

This paper focuses on the research of micro-expression recognition (MER) and proposes a flexible and reliable deep learning method called learning to rank onset-occurring-offset representations (LTR3O). The LTR3O method introduces a dynamic and reduced-size sequence structure known as 3O, which consists of onset, occurring, and offset frames, for representing micro-expressions (MEs). This structure facilitates the subsequent learning of ME-discriminative features. A noteworthy advantage of the 3O structure is its flexibility, as the occurring frame is randomly extracted from the original ME sequence without the need for accurate frame spotting methods. Based on the 3O structures, LTR3O generates multiple 3O representation candidates for each ME sample and incorporates well-designed modules to measure and calibrate their emotional expressiveness. This calibration process ensures that the distribution of these candidates aligns with that of macro-expressions (MaMs) over time. Consequently, the visibility of MEs can be implicitly enhanced, facilitating the reliable learning of more discriminative features for MER. Extensive experiments were conducted to evaluate the performance of LTR3O using three widely-used ME databases: CASME II, SMIC, and SAMM. The experimental results demonstrate the effectiveness and superior performance of LTR3O, particularly in terms of its flexibility and reliability, when compared to recent state-of-the-art MER methods.

Jie Zhu, Huaixia Dou, Junhui Li et al.

Emotional Support Conversation (ESC) aims to assist individuals experiencing distress by generating empathetic and supportive dialogue. While prior work typically assumes that each supporter turn corresponds to a single strategy, real-world supportive communication often involves multiple strategies within a single utterance. In this paper, we revisit the ESC task by formulating it as multi-strategy utterance generation, where each utterance may contain one or more strategy-response pairs. We propose two generation methods: All-in-One, which predicts all strategy-response pairs in a single decoding step, and One-by-One, which iteratively generates strategy-response pairs until completion. Both methods are further enhanced with cognitive reasoning guided by reinforcement learning to improve strategy selection and response composition. We evaluate our models on the ESConv dataset under both utterance-level and dialogue-level settings. Experimental results show that our methods effectively model multi-strategy utterances and lead to improved supportive quality and dialogue success. To our knowledge, this work provides the first systematic empirical evidence that allowing multiple support strategies within a single utterance is both feasible and beneficial for emotional support conversations. All code and data will be publicly available at https://github.com/aliyun/qwen-dianjin.

Yanbing Zeng, Jia Wang, Hanghang Ma et al.

Integrating image generation and understanding into a single framework has become a pivotal goal in the multimodal domain. However, how understanding can effectively assist generation has not been fully explored. Unlike previous works that focus on leveraging reasoning abilities and world knowledge from understanding models, this paper introduces a novel perspective: leveraging understanding to enhance the fidelity and detail richness of generated images. To this end, we propose Forge-and-Quench, a new unified framework that puts this principle into practice. In the generation process of our framework, an MLLM first reasons over the entire conversational context, including text instructions, to produce an enhanced text instruction. This refined instruction is then mapped to a virtual visual representation, termed the Bridge Feature, via a novel Bridge Adapter. This feature acts as a crucial link, forging insights from the understanding model to quench and refine the generation process. It is subsequently injected into the T2I backbone as a visual guidance signal, alongside the enhanced text instruction that replaces the original input. To validate this paradigm, we conduct comprehensive studies on the design of the Bridge Feature and Bridge Adapter. Our framework demonstrates exceptional extensibility and flexibility, enabling efficient migration across different MLLM and T2I models with significant savings in training overhead, all without compromising the MLLM's inherent multimodal understanding capabilities. Experiments show that Forge-and-Quench significantly improves image fidelity and detail across multiple models, while also maintaining instruction-following accuracy and enhancing world knowledge application. Models and codes are available at https://github.com/YanbingZeng/Forge-and-Quench.

Jie Zhu, Leye Wang, Xiao Han

The size of deep learning models in artificial intelligence (AI) software is increasing rapidly, which hinders the large-scale deployment on resource-restricted devices (e.g., smartphones). To mitigate this issue, AI software compression plays a crucial role, which aims to compress model size while keeping high performance. However, the intrinsic defects in the big model may be inherited by the compressed one. Such defects may be easily leveraged by attackers, since the compressed models are usually deployed in a large number of devices without adequate protection. In this paper, we try to address the safe model compression problem from a safety-performance co-optimization perspective. Specifically, inspired by the test-driven development (TDD) paradigm in software engineering, we propose a test-driven sparse training framework called SafeCompress. By simulating the attack mechanism as the safety test, SafeCompress can automatically compress a big model to a small one following the dynamic sparse training paradigm. Further, considering a representative attack, i.e., membership inference attack (MIA), we develop a concrete safe model compression mechanism, called MIA-SafeCompress. Extensive experiments are conducted to evaluate MIA-SafeCompress on five datasets for both computer vision and natural language processing tasks. The results verify the effectiveness and generalization of our method. We also discuss how to adapt SafeCompress to other attacks besides MIA, demonstrating the flexibility of SafeCompress.

Jie Zhu, Girish Chandar Ganesan, Xiaoming Liu

Monocular metric depth estimation has achieved strong progress with large-scale training and universal-camera modeling, yet robust deployment across diverse camera settings, such as perspective, fisheye, and panoramic images, remains challenging. Existing methods typically rely on a single depth estimator, overlooking that different models encode different camera assumptions and perform best under different input domains. In this paper, we show that depth experts exhibit strong sample-wise complementarity: model preference is highly correlated with camera geometry, and multi-model fusion brings the largest gains on difficult samples where individual experts are unreliable. Motivated by these observations, we propose \textbf{\ours}, a vision-language agent for adaptive monocular depth estimation. DepthAgent treats existing depth models as frozen tools and learns to analyze scene and camera cues, invoke suitable experts through multi-turn tool utilization, and select or fuse their predictions for each input. To optimize such discrete decision-making toward dense geometric quality, we design a multi-reward reinforcement fine-tuning scheme that jointly encourages valid tool execution, camera/scene analysis, expert-selection quality, and inference efficiency. Extensive experiments across perspective, fisheye, and panoramic benchmarks show that \ours consistently outperforms individual experts, fixed model fusion, and different selection strategies, with strong improvements on challenging samples, highlighting the critical role of expert selection and fusion. The code and model will be released upon publication.

Jie Zhu, Xiao Guo, Yiyang Su et al.

Model fusion is a key strategy for robust recognition in unconstrained scenarios, as different models provide complementary strengths. This is especially important for whole-body human recognition, where biometric cues such as face, gait, and body shape vary across samples and are typically integrated via score-fusion. However, existing score-fusion strategies are usually static, invoking all models for every test sample regardless of sample quality or modality reliability. To overcome these limitations, we propose \textbf{FusionAgent}, a novel agentic framework that leverages a Multimodal Large Language Model (MLLM) to perform dynamic, sample-specific model selection. Each expert model is treated as a tool, and through Reinforcement Fine-Tuning (RFT) with a metric-based reward, the agent learns to adaptively determine the optimal model combination for each test input. To address the model score misalignment and embedding heterogeneity, we introduce Anchor-based Confidence Top-k (ACT) score-fusion, which anchors on the most confident model and integrates complementary predictions in a confidence-aware manner. Extensive experiments on multiple whole-body biometric benchmarks demonstrate that FusionAgent significantly outperforms SoTA methods while achieving higher efficiency through fewer model invocations, underscoring the critical role of dynamic, explainable, and robust model fusion in real-world recognition systems. Project page: \href{https://fusionagent.github.io/}{FusionAgent}.

Danila Rozhevskii, Jie Zhu, Boyuan Zhao

In the last few years, automated recommendation systems have been a major focus in the music field, where companies such as Spotify, Amazon, and Apple are competing in the ability to generate the most personalized music suggestions for their users. One of the challenges developers still fail to tackle is taking into account the psychological and emotional aspects of the music. Our goal is to find a way to integrate users' personal traits and their current emotional state into a single music recommendation system with both collaborative and content-based filtering. We seek to relate the personality and the current emotional state of the listener to the audio features in order to build an emotion-aware MRS. We compare the results both quantitatively and qualitatively to the output of the traditional MRS based on the Spotify API data to understand if our advancements make a significant impact on the quality of music recommendations.

Jie Zhu, Junhui Li, Yalong Wen et al.

In light of recent breakthroughs in large language models (LLMs) that have revolutionized natural language processing (NLP), there is an urgent need for new benchmarks to keep pace with the fast development of LLMs. In this paper, we propose CFLUE, the Chinese Financial Language Understanding Evaluation benchmark, designed to assess the capability of LLMs across various dimensions. Specifically, CFLUE provides datasets tailored for both knowledge assessment and application assessment. In knowledge assessment, it consists of 38K+ multiple-choice questions with associated solution explanations. These questions serve dual purposes: answer prediction and question reasoning. In application assessment, CFLUE features 16K+ test instances across distinct groups of NLP tasks such as text classification, machine translation, relation extraction, reading comprehension, and text generation. Upon CFLUE, we conduct a thorough evaluation of representative LLMs. The results reveal that only GPT-4 and GPT-4-turbo achieve an accuracy exceeding 60\% in answer prediction for knowledge assessment, suggesting that there is still substantial room for improvement in current LLMs. In application assessment, although GPT-4 and GPT-4-turbo are the top two performers, their considerable advantage over lightweight LLMs is noticeably diminished. The datasets and scripts associated with CFLUE are openly accessible at https://github.com/aliyun/cflue.

Jie Zhu, Yiyang Su, Xiaoming Liu

Multi-modal large language models (MLLMs) exhibit strong general-purpose capabilities, yet still struggle on Fine-Grained Visual Classification (FGVC), a core perception task that requires subtle visual discrimination and is crucial for many real-world applications. A widely adopted strategy for boosting performance on challenging tasks such as math and coding is Chain-of-Thought (CoT) reasoning. However, several prior works have reported that CoT can actually harm performance on visual perception tasks. These studies, though, examine the issue from relatively narrow angles and leave open why CoT degrades perception-heavy performance. We systematically re-examine the role of CoT in FGVC through the lenses of zero-shot evaluation and multiple training paradigms. Across these settings, we uncover a central paradox: the degradation induced by CoT is largely driven by the reasoning length, in which longer textual reasoning consistently lowers classification accuracy. We term this phenomenon the ``Cost of Thinking''. Building on this finding, we make two key contributions: (1) MRN, a simple and general plug-and-play normalization method for multi-reward optimization that balances heterogeneous reward signals, and (2) ReFine-RFT, a framework that combines ensemble rewards with MRN to constrain reasoning length while providing dense accuracy-oriented feedback. Extensive experiments demonstrate the effectiveness of our findings and the proposed ReFine-RFT, achieving state-of-the-art performance across FGVC benchmarks. Project page: \href{https://refine-rft.github.io/}{ReFine-RFT}.

Yangjun Zeng, Yiwei Qiu, Xiaocong Sun et al.

Integrating renewable generation, hydrogen production, and renewable ammonia (RA) synthesis into power-to-ammonia (P2A) systems creates interactions across electricity and hydrogen markets. Limited operational flexibility, however, places RA at a disadvantage at the Nash equilibrium (NE). Recent advances in ammonia synthesis reactor design enable hot standby (HSB) operation, improving flexibility but introducing integer decision variables that complicate market equilibrium analysis. To address this challenge, we develop a potential game model and derive a convergent ε-approximate equilibrium via an iterative best-response approach. Case studies show that HSB reduces RA's reliance on hydrogen purchases and increases its profit by 20.14%. More importantly, HSB shifts the market equilibrium toward a more mutually beneficial outcome.

Jie Zhu, Yimin Tian, Boyang Li et al.

This paper introduces \textbf{FinMCP-Bench}, a novel benchmark for evaluating large language models (LLMs) in solving real-world financial problems through tool invocation of financial model context protocols. FinMCP-Bench contains 613 samples spanning 10 main scenarios and 33 sub-scenarios, featuring both real and synthetic user queries to ensure diversity and authenticity. It incorporates 65 real financial MCPs and three types of samples, single tool, multi-tool, and multi-turn, allowing evaluation of models across different levels of task complexity. Using this benchmark, we systematically assess a range of mainstream LLMs and propose metrics that explicitly measure tool invocation accuracy and reasoning capabilities. FinMCP-Bench provides a standardized, practical, and challenging testbed for advancing research on financial LLM agents.

Junhong Lin, Xinyue Zeng, Jie Zhu et al. · amazon-science

Large Language Models (LLMs) have achieved remarkable success in complex reasoning tasks, but their inference remains computationally inefficient. We observe a common failure mode in many prevalent LLMs, overthinking, where models generate verbose and tangential reasoning traces even for simple queries. Recent works have tried to mitigate this by enforcing fixed token budgets, however, this can lead to underthinking, especially on harder problems. Through empirical analysis, we identify that this inefficiency often stems from unclear problem-solving strategies. To formalize this, we develop a theoretical model, BBAM (Bayesian Budget Allocation Model), which models reasoning as a sequence of sub-questions with varying uncertainty, and introduce the $E^3$ metric to capture the trade-off between correctness and computation efficiency. Building on theoretical results from BBAM, we propose Plan-and-Budget, a model-agnostic, test-time framework that decomposes complex queries into sub-questions and allocates token budgets based on estimated complexity using adaptive scheduling. Plan-and-Budget improves reasoning efficiency across a range of tasks and models, achieving up to +70% accuracy gains, -39% token reduction, and +187.5% improvement in $E^3$. Notably, it elevates a smaller model (DS-Qwen-32B) to match the efficiency of a larger model (DS-LLaMA-70B)-demonstrating Plan-and-Budget's ability to close performance gaps without retraining. Our code is available at https://github.com/junhongmit/P-and-B.

Jie Zhu, Jirong Zha, Ding Li et al.

Self-supervised learning shows promise in harnessing extensive unlabeled data, but it also confronts significant privacy concerns, especially in vision. In this paper, we aim to perform membership inference on visual self-supervised models in a more realistic setting: self-supervised training method and details are unknown for an adversary when attacking as he usually faces a black-box system in practice. In this setting, considering that self-supervised model could be trained by completely different self-supervised paradigms, e.g., masked image modeling and contrastive learning, with complex training details, we propose a unified membership inference method called PartCrop. It is motivated by the shared part-aware capability among models and stronger part response on the training data. Specifically, PartCrop crops parts of objects in an image to query responses with the image in representation space. We conduct extensive attacks on self-supervised models with different training protocols and structures using three widely used image datasets. The results verify the effectiveness and generalization of PartCrop. Moreover, to defend against PartCrop, we evaluate two common approaches, i.e., early stop and differential privacy, and propose a tailored method called shrinking crop scale range. The defense experiments indicate that all of them are effective. Our code is available at https://github.com/JiePKU/PartCrop.

Donglei Yu, Yang Zhao, Jie Zhu et al.

Simultaneous Machine Translation (SiMT) generates translations while receiving streaming source inputs. This requires the SiMT model to learn a read/write policy, deciding when to translate and when to wait for more source input. Numerous linguistic studies indicate that audiences in SiMT scenarios have distinct preferences, such as accurate translations, simpler syntax, and no unnecessary latency. Aligning SiMT models with these human preferences is crucial to improve their performances. However, this issue still remains unexplored. Additionally, preference optimization for SiMT task is also challenging. Existing methods focus solely on optimizing the generated responses, ignoring human preferences related to latency and the optimization of read/write policy during the preference optimization phase. To address these challenges, we propose Simultaneous Preference Learning (SimulPL), a preference learning framework tailored for the SiMT task. In the SimulPL framework, we categorize SiMT human preferences into five aspects: \textbf{translation quality preference}, \textbf{monotonicity preference}, \textbf{key point preference}, \textbf{simplicity preference}, and \textbf{latency preference}. By leveraging the first four preferences, we construct human preference prompts to efficiently guide GPT-4/4o in generating preference data for the SiMT task. In the preference optimization phase, SimulPL integrates \textbf{latency preference} into the optimization objective and enables SiMT models to improve the read/write policy, thereby aligning with human preferences more effectively. Experimental results indicate that SimulPL exhibits better alignment with human preferences across all latency levels in Zh$\rightarrow$En, De$\rightarrow$En and En$\rightarrow$Zh SiMT tasks. Our data and code will be available at https://github.com/EurekaForNLP/SimulPL.

Kangbo Sun, Jie Zhu

Pseudo-supervised learning methods have been shown to be effective for weakly supervised object localization tasks. However, the effectiveness depends on the powerful regularization ability of deep neural networks. Based on the assumption that the localization network should have similar location predictions on different versions of the same image, we propose a two-stage approach to learn more consistent localization. In the first stage, we propose a mask-based pseudo label generator algorithm, and use the pseudo-supervised learning method to initialize an object localization network. In the second stage, we propose a simple and effective method for evaluating the confidence of pseudo-labels based on classification discrimination, and by learning consistency from high-quality pseudo-labels, we further refine the localization network to get better localization performance. Experimental results show that our proposed approach achieves excellent performance in three benchmark datasets including CUB-200-2011, ImageNet-1k and Tiny-ImageNet, which demonstrates its effectiveness.

Xiao Guo, Jie Zhu, Anil Jain et al.

The rapid advancement of AI-generated content (AIGC) has escalated the threat of deepfakes, from facial manipulations to the synthesis of entire photorealistic human bodies. However, existing detection methods remain fragmented, specializing either in facial-region forgeries or full-body synthetic images, and consequently fail to generalize across the full spectrum of human image manipulations. We introduce HuForDet, a holistic framework for human image forgery detection, which features a dual-branch architecture comprising: (1) a face forgery detection branch that employs heterogeneous experts operating in both RGB and frequency domains, including an adaptive Laplacian-of-Gaussian (LoG) module designed to capture artifacts ranging from fine-grained blending boundaries to coarse-scale texture irregularities; and (2) a contextualized forgery detection branch that leverages a Multi-Modal Large Language Model (MLLM) to analyze full-body semantic consistency, enhanced with a confidence estimation mechanism that dynamically weights its contribution during feature fusion. We curate a human image forgery (HuFor) dataset that unifies existing face forgery data with a new corpus of full-body synthetic humans. Extensive experiments show that our HuForDet achieves state-of-the-art forgery detection performance and superior robustness across diverse human image forgeries.

Junwen Chen, Jie Zhu, Yu Kong

Despite significant progress in video question answering (VideoQA), existing methods fall short of questions that require causal/temporal reasoning across frames. This can be attributed to imprecise motion representations. We introduce Action Temporality Modeling (ATM) for temporality reasoning via three-fold uniqueness: (1) rethinking the optical flow and realizing that optical flow is effective in capturing the long horizon temporality reasoning; (2) training the visual-text embedding by contrastive learning in an action-centric manner, leading to better action representations in both vision and text modalities; and (3) preventing the model from answering the question given the shuffled video in the fine-tuning stage, to avoid spurious correlation between appearance and motion and hence ensure faithful temporality reasoning. In the experiments, we show that ATM outperforms previous approaches in terms of the accuracy on multiple VideoQAs and exhibits better true temporality reasoning ability.

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.

Jie Zhu, Jirong Zha, Ding Li et al.

Self-supervised learning shows promise in harnessing extensive unlabeled data, but it also confronts significant privacy concerns, especially in vision. In this paper, we perform membership inference on visual self-supervised models in a more realistic setting: self-supervised training method and details are unknown for an adversary when attacking as he usually faces a black-box system in practice. In this setting, considering that self-supervised model could be trained by completely different self-supervised paradigms, e.g., masked image modeling and contrastive learning, with complex training details, we propose a unified membership inference method called PartCrop. It is motivated by the shared part-aware capability among models and stronger part response on the training data. Specifically, PartCrop crops parts of objects in an image to query responses within the image in representation space. We conduct extensive attacks on self-supervised models with different training protocols and structures using three widely used image datasets. The results verify the effectiveness and generalization of PartCrop. Moreover, to defend against PartCrop, we evaluate two common approaches, i.e., early stop and differential privacy, and propose a tailored method called shrinking crop scale range. The defense experiments indicate that all of them are effective. Finally, besides prototype testing on toy visual encoders and small-scale image datasets, we quantitatively study the impacts of scaling from both data and model aspects in a realistic scenario and propose a scalable PartCrop-v2 by introducing two structural improvements to PartCrop. Our code is at https://github.com/JiePKU/PartCrop.

Huaixia Dou, Xinyu Tian, Xinglin Lyu et al.

Recent advancements in large language models (LLMs) have demonstrated their remarkable capabilities across various language tasks. Inspired by the success of text-to-text translation refinement, this paper investigates how LLMs can improve the performance of speech translation by introducing a joint refinement process. Through the joint refinement of speech translation (ST) and automatic speech recognition (ASR) transcription via LLMs, the performance of the ST model is significantly improved in both training-free in-context learning and parameter-efficient fine-tuning scenarios. Additionally, we explore the effect of document-level context on refinement under the context-aware fine-tuning scenario. Experimental results on the MuST-C and CoVoST 2 datasets, which include seven translation tasks, demonstrate the effectiveness of the proposed approach using several popular LLMs including GPT-3.5-turbo, LLaMA3-8B, and Mistral-12B. Further analysis further suggests that jointly refining both transcription and translation yields better performance compared to refining translation alone. Meanwhile, incorporating document-level context significantly enhances refinement performance. We release our code and datasets on GitHub.

Jie Zhu, Hanghang Ma, Jia Wang et al.

In this work, we introduce Wallaroo, a simple autoregressive baseline that leverages next-token prediction to unify multi-modal understanding, image generation, and editing at the same time. Moreover, Wallaroo supports multi-resolution image input and output, as well as bilingual support for both Chinese and English. We decouple the visual encoding into separate pathways and apply a four-stage training strategy to reshape the model's capabilities. Experiments are conducted on various benchmarks where Wallaroo produces competitive performance or exceeds other unified models, suggesting the great potential of autoregressive models in unifying multi-modality understanding and generation. Our code is available at https://github.com/JiePKU/Wallaroo.

Jie Zhu, Huaixia Dou, Junhui Li et al.

Effective customer support requires not only accurate problem solving but also structured and empathetic communication aligned with professional standards. However, existing dialogue datasets often lack strategic guidance, and real-world service data is difficult to access and annotate. To address this, we introduce the task of Customer Support Conversation (CSC), aimed at training customer service agents to respond using well-defined support strategies. We propose a structured CSC framework grounded in COPC guidelines, defining five conversational stages and twelve strategies to guide high-quality interactions. Based on this, we construct CSConv, an evaluation dataset of 1,855 real-world customer-agent conversations rewritten using LLMs to reflect deliberate strategy use, and annotated accordingly. Additionally, we develop a role-playing approach that simulates strategy-rich conversations using LLM-powered roles aligned with the CSC framework, resulting in the training dataset RoleCS. Experiments show that fine-tuning strong LLMs on RoleCS significantly improves their ability to generate high-quality, strategy-aligned responses on CSConv. Human evaluations further confirm gains in problem resolution. All code and data will be made publicly available at https://github.com/aliyun/qwen-dianjin.

Jie Zhu, Leye Wang

Text-to-image diffusion model since its propose has significantly influenced the content creation due to its impressive generation capability. However, this capability depends on large-scale text-image datasets gathered from web platforms like social media, posing substantial challenges in copyright compliance and personal privacy leakage. Though there are some efforts devoted to explore approaches for auditing data provenance in text-to-image diffusion models, existing work has unrealistic assumptions that can obtain model internal knowledge, e.g., intermediate results, or the evaluation is not reliable. To fill this gap, we propose a completely black-box auditing framework called Feature Semantic Consistency-based Auditing (FSCA). It utilizes two types of semantic connections within the text-to-image diffusion model for auditing, eliminating the need for access to internal knowledge. To demonstrate the effectiveness of our FSCA framework, we perform extensive experiments on LAION-mi dataset and COCO dataset, and compare with eight state-of-the-art baseline approaches. The results show that FSCA surpasses previous baseline approaches across various metrics and different data distributions, showcasing the superiority of our FSCA. Moreover, we introduce a recall balance strategy and a threshold adjustment strategy, which collectively allows FSCA to reach up a user-level accuracy of 90% in a real-world auditing scenario with only 10 samples/user, highlighting its strong auditing potential in real-world applications. Our code is made available at https://github.com/JiePKU/FSCA.

Jie Zhu, Shenggui Li, Yang You

Federated learning is proposed by Google to safeguard data privacy through training models locally on users' devices. However, with deep learning models growing in size to achieve better results, it becomes increasingly difficult to accommodate the whole model on one single device. Thus, model parallelism is then used to divide the model weights among several devices. With this logic, the approach currently used evenly allocates weights among devices. However, in reality, a computation bottleneck may occur resulting from variant computing power of different users' devices. To address this problem, load balancing is needed to allocate the model weights based on the computational capability of the device. In this paper, we proposed Sky Computing, a load-balanced model parallelism framework to adaptively allocate the weights to devices. Sky Computing outperforms the baseline method by 55% in training time when training 160-layer BERT with 64 nodes. The source code can be found at https://github.com/hpcaitech/SkyComputing.

Huaxi Jiang, Jie Zhu, Li Yang et al.

The release note is an essential software artifact of open-source software that documents crucial information about changes, such as new features and bug fixes. With the help of release notes, both developers and users could have a general understanding of the latest version without browsing the source code. However, it is a daunting and time-consuming job for developers to produce release notes. Although prior studies have provided some automatic approaches, they generate release notes mainly by extracting information from code changes. This will result in language-specific and not being general enough to be applicable. Therefore, helping developers produce release notes effectively remains an unsolved challenge. To address the problem, we first conduct a manual study on the release notes of 900 GitHub projects, which reveals that more than 54% of projects produce their release notes with pull requests. Based on the empirical finding, we propose a deep learning based approach named DeepRelease (Deep learning based Release notes generator) to generate release notes according to pull requests. The process of release notes generation in DeepRelease includes the change entries generation and the change category (i.e., new features or bug fixes) generation, which are formulated as a text summarization task and a multi-class classification problem, respectively. Since DeepRelease fully employs text information from pull requests to summarize changes and identify the change category, it is language-agnostic and can be used for projects in any language. We build a dataset with over 46K release notes and evaluate DeepRelease on the dataset. The experimental results indicate that DeepRelease outperforms four baselines and can generate release notes similar to those manually written ones in a fraction of the time.

Laila Rasmy, Jie Zhu, Zhiheng Li et al.

Recently, there is great interest to investigate the application of deep learning models for the prediction of clinical events using electronic health records (EHR) data. In EHR data, a patient's history is often represented as a sequence of visits, and each visit contains multiple events. As a result, deep learning models developed for sequence modeling, like recurrent neural networks (RNNs) are common architecture for EHR-based clinical events predictive models. While a large variety of RNN models were proposed in the literature, it is unclear if complex architecture innovations will offer superior predictive performance. In order to move this field forward, a rigorous evaluation of various methods is needed. In this study, we conducted a thorough benchmark of RNN architectures in modeling EHR data. We used two prediction tasks: the risk for developing heart failure and the risk of early readmission for inpatient hospitalization. We found that simple gated RNN models, including GRUs and LSTMs, often offer competitive results when properly tuned with Bayesian Optimization, which is in line with similar to findings in the natural language processing (NLP) domain. For reproducibility, Our codebase is shared at https://github.com/ZhiGroup/pytorch_ehr.

Zhixian Ma, Jie Zhu, Weitian Li et al.

The morphology of a radio galaxy is highly affected by its central active galactic nuclei (AGN), which is studied to reveal the evolution of the super massive black hole (SMBH). In this work, we propose a morphology generation framework for two typical radio galaxies namely Fanaroff-Riley type-I (FRI) and type-II (FRII) with deep neural network based autoencoder (DNNAE) and Gaussian mixture models (GMMs). The encoder and decoder subnets in the DNNAE are symmetric aside a fully-connected layer namely code layer hosting the extracted feature vectors. By randomly generating the feature vectors later with a three-component Gaussian Mixture models, new FRI or FRII radio galaxy morphologies are simulated. Experiments were demonstrated on real radio galaxy images, where we discussed the length of feature vectors, selection of lost functions, and made comparisons on batch normalization and dropout techniques for training the network. The results suggest a high efficiency and performance of our morphology generation framework. Code is available at: https://github.com/myinxd/dnnae-gmm.

Jie Zhu, Yinliang Xu, Hongbin Sun

The growing penetration of renewable energy necessitates high-frequency real-time scheduling. While neural network-based surrogates enable computationally efficient scheduling, strictly enforcing nonconvex power flow constraints without external solvers remains a fundamental challenge. To bridge this gap, this letter proposes a solver-free neural dispatch framework with rigorous feasibility guarantees. A convex inner approximation of the DistFlow model is first derived via the convex envelope theorem. Building upon this approximation, a robust optimization-based affine policy is formulated to yield a theoretically certified interior-point mapping rule, which is then embedded within a bisection-based projection scheme to efficiently recover feasibility for infeasible NN outputs without any external solver. Experimental results demonstrate that the proposed method restores feasibility on the order of $10^{-3}$ s while maintaining near-optimal performance.

Jie Zhu, Yixiong Chen, Mingyu Ding et al.

Text-to-image diffusion has attracted vast attention due to its impressive image-generation capabilities. However, when it comes to human-centric text-to-image generation, particularly in the context of faces and hands, the results often fall short of naturalness due to insufficient training priors. We alleviate the issue in this work from two perspectives. 1) From the data aspect, we carefully collect a human-centric dataset comprising over one million high-quality human-in-the-scene images and two specific sets of close-up images of faces and hands. These datasets collectively provide a rich prior knowledge base to enhance the human-centric image generation capabilities of the diffusion model. 2) On the methodological front, we propose a simple yet effective method called Mixture of Low-rank Experts (MoLE) by considering low-rank modules trained on close-up hand and face images respectively as experts. This concept draws inspiration from our observation of low-rank refinement, where a low-rank module trained by a customized close-up dataset has the potential to enhance the corresponding image part when applied at an appropriate scale. To validate the superiority of MoLE in the context of human-centric image generation compared to state-of-the-art, we construct two benchmarks and perform evaluations with diverse metrics and human studies. Datasets, model, and code are released at https://sites.google.com/view/mole4diffuser/.

Jie Zhu, Qian Chen, Huaixia Dou et al.

Effective reasoning remains a core challenge for large language models (LLMs) in the financial domain, where tasks often require domain-specific knowledge, precise numerical calculations, and strict adherence to compliance rules. We propose DianJin-R1, a reasoning-enhanced framework designed to address these challenges through reasoning-augmented supervision and reinforcement learning. Central to our approach is DianJin-R1-Data, a high-quality dataset constructed from CFLUE, FinQA, and a proprietary compliance corpus (Chinese Compliance Check, CCC), combining diverse financial reasoning scenarios with verified annotations. Our models, DianJin-R1-7B and DianJin-R1-32B, are fine-tuned from Qwen2.5-7B-Instruct and Qwen2.5-32B-Instruct using a structured format that generates both reasoning steps and final answers. To further refine reasoning quality, we apply Group Relative Policy Optimization (GRPO), a reinforcement learning method that incorporates dual reward signals: one encouraging structured outputs and another rewarding answer correctness. We evaluate our models on five benchmarks: three financial datasets (CFLUE, FinQA, and CCC) and two general reasoning benchmarks (MATH-500 and GPQA-Diamond). Experimental results show that DianJin-R1 models consistently outperform their non-reasoning counterparts, especially on complex financial tasks. Moreover, on the real-world CCC dataset, our single-call reasoning models match or even surpass the performance of multi-agent systems that require significantly more computational cost. These findings demonstrate the effectiveness of DianJin-R1 in enhancing financial reasoning through structured supervision and reward-aligned learning, offering a scalable and practical solution for real-world applications.

Jie Zhu, Leye Wang, Xiao Han et al.

The size of deep learning models in artificial intelligence (AI) software is increasing rapidly, hindering the large-scale deployment on resource-restricted devices (e.g., smartphones). To mitigate this issue, AI software compression plays a crucial role, which aims to compress model size while keeping high performance. However, the intrinsic defects in a big model may be inherited by the compressed one. Such defects may be easily leveraged by adversaries, since a compressed model is usually deployed in a large number of devices without adequate protection. In this article, we aim to address the safe model compression problem from the perspective of safety-performance co-optimization. Specifically, inspired by the test-driven development (TDD) paradigm in software engineering, we propose a test-driven sparse training framework called SafeCompress. By simulating the attack mechanism as safety testing, SafeCompress can automatically compress a big model to a small one following the dynamic sparse training paradigm. Then, considering two kinds of representative and heterogeneous attack mechanisms, i.e., black-box membership inference attack and white-box membership inference attack, we develop two concrete instances called BMIA-SafeCompress and WMIA-SafeCompress. Further, we implement another instance called MMIA-SafeCompress by extending SafeCompress to defend against the occasion when adversaries conduct black-box and white-box membership inference attacks simultaneously. We conduct extensive experiments on five datasets for both computer vision and natural language processing tasks. The results show the effectiveness and generalizability of our framework. We also discuss how to adapt SafeCompress to other attacks besides membership inference attack, demonstrating the flexibility of SafeCompress.

Weisong Sun, Yiran Zhang, Jie Zhu et al.

Commenting code is a crucial activity in software development, as it aids in facilitating future maintenance and updates. To enhance the efficiency of writing comments and reduce developers' workload, researchers has proposed various automated code summarization (ACS) techniques to automatically generate comments/summaries for given code units. However, these ACS techniques primarily focus on generating summaries for code units at the method level. There is a significant lack of research on summarizing higher-level code units, such as file-level and module-level code units, despite the fact that summaries of these higher-level code units are highly useful for quickly gaining a macro-level understanding of software components and architecture. To fill this gap, in this paper, we conduct a systematic study on how to use LLMs for commenting higher-level code units, including file level and module level. These higher-level units are significantly larger than method-level ones, which poses challenges in handling long code inputs within LLM constraints and maintaining efficiency. To address these issues, we explore various summarization strategies for ACS of higher-level code units, which can be divided into three types: full code summarization, reduced code summarization, and hierarchical code summarization. The experimental results suggest that for summarizing file-level code units, using the full code is the most effective approach, with reduced code serving as a cost-efficient alternative. However, for summarizing module-level code units, hierarchical code summarization becomes the most promising strategy. In addition, inspired by the research on method-level ACS, we also investigate using the LLM as an evaluator to evaluate the quality of summaries of higher-level code units. The experimental results demonstrate that the LLM's evaluation results strongly correlate with human evaluations.

Feng Liu, Nicholas Chimitt, Lanqing Guo et al. · gatech

We address the problem of whole-body person recognition in unconstrained environments. This problem arises in surveillance scenarios such as those in the IARPA Biometric Recognition and Identification at Altitude and Range (BRIAR) program, where biometric data is captured at long standoff distances, elevated viewing angles, and under adverse atmospheric conditions (e.g., turbulence and high wind velocity). To this end, we propose FarSight, a unified end-to-end system for person recognition that integrates complementary biometric cues across face, gait, and body shape modalities. FarSight incorporates novel algorithms across four core modules: multi-subject detection and tracking, recognition-aware video restoration, modality-specific biometric feature encoding, and quality-guided multi-modal fusion. These components are designed to work cohesively under degraded image conditions, large pose and scale variations, and cross-domain gaps. Extensive experiments on the BRIAR dataset, one of the most comprehensive benchmarks for long-range, multi-modal biometric recognition, demonstrate the effectiveness of FarSight. Compared to our preliminary system, this system achieves a 34.1% absolute gain in 1:1 verification accuracy (TAR@0.1% FAR), a 17.8% increase in closed-set identification (Rank-20), and a 34.3% reduction in open-set identification errors (FNIR@1% FPIR). Furthermore, FarSight was evaluated in the 2025 NIST RTE Face in Video Evaluation (FIVE), which conducts standardized face recognition testing on the BRIAR dataset. These results establish FarSight as a state-of-the-art solution for operational biometric recognition in challenging real-world conditions.

Chang Yu, Fang Liu, Jie Zhu et al. · amazon-science

This paper proposes a hybrid framework combining LSTM (Long Short-Term Memory) networks with LightGBM and CatBoost for stock price prediction. The framework processes time-series financial data and evaluates performance using seven models: Artificial Neural Networks (ANNs), Convolutional Neural Networks (CNNs), Bidirectional LSTM (BiLSTM), vanilla LSTM, XGBoost, LightGBM, and standard Neural Networks (NNs). Key metrics, including MAE, R-squared, MSE, and RMSE, are used to establish benchmarks across different time scales. Building on these benchmarks, we develop an ensemble model that combines the strengths of sequential and tree-based approaches. Experimental results show that the proposed framework improves accuracy by 10 to 15 percent compared to individual models and reduces error during market changes. This study highlights the potential of ensemble methods for financial forecasting and provides a flexible design for integrating new machine learning techniques.

Jie Zhu, Chihao Shen, Ziyang Li et al.

Directed fuzzing aims to find program inputs that lead to specified target program states. It has broad applications, such as debugging system crashes, confirming reported bugs, and generating exploits for potential vulnerabilities. This task is inherently challenging because target states are often deeply nested in the program, while the search space manifested by numerous possible program inputs is prohibitively large. Existing approaches rely on branch distances or manually-specified constraints to guide the search; however, the branches alone are often insufficient to precisely characterize progress toward reaching the target states, while the manually specified constraints are often tailored for specific bug types and thus difficult to generalize to diverse target states and programs. We present Locus, a novel framework to improve the efficiency of directed fuzzing. Our key insight is to synthesize predicates to capture fuzzing progress as semantically meaningful intermediate states, serving as milestones towards reaching the target states. When used to instrument the program under fuzzing, they can reject executions unlikely to reach the target states, while providing additional coverage guidance. To automate this task and generalize to diverse programs, Locus features an agentic framework with program analysis tools to synthesize and iteratively refine the candidate predicates, while ensuring the predicates strictly relax the target states to prevent false rejections via symbolic execution. Our evaluation shows that Locus substantially improves the efficiency of eight state-of-the-art fuzzers in discovering real-world vulnerabilities, achieving an average speedup of 41.6x. So far, Locus has found eight previously unpatched bugs, with one already acknowledged with a draft patch.

Jie Zhu, Yiyang Su, Minchul Kim et al.

Whole-body biometric recognition is a challenging multimodal task that integrates various biometric modalities, including face, gait, and body. This integration is essential for overcoming the limitations of unimodal systems. Traditionally, whole-body recognition involves deploying different models to process multiple modalities, achieving the final outcome by score-fusion (e.g., weighted averaging of similarity matrices from each model). However, these conventional methods may overlook the variations in score distributions of individual modalities, making it challenging to improve final performance. In this work, we present \textbf{Q}uality-guided \textbf{M}ixture of score-fusion \textbf{E}xperts (QME), a novel framework designed for improving whole-body biometric recognition performance through a learnable score-fusion strategy using a Mixture of Experts (MoE). We introduce a novel pseudo-quality loss for quality estimation with a modality-specific Quality Estimator (QE), and a score triplet loss to improve the metric performance. Extensive experiments on multiple whole-body biometric datasets demonstrate the effectiveness of our proposed approach, achieving state-of-the-art results across various metrics compared to baseline methods. Our method is effective for multimodal and multi-model, addressing key challenges such as model misalignment in the similarity score domain and variability in data quality.

Donghua Yu, Zhengyuan Lin, Chen Yang et al.

Speaker-Attributed, Time-Stamped Transcription (SATS) aims to transcribe what is said and to precisely determine the timing of each speaker, which is particularly valuable for meeting transcription. Existing SATS systems rarely adopt an end-to-end formulation and are further constrained by limited context windows, weak long-range speaker memory, and the inability to output timestamps. To address these limitations, we present MOSS Transcribe Diarize, a unified multimodal large language model that jointly performs Speaker-Attributed, Time-Stamped Transcription in an end-to-end paradigm. Trained on extensive real wild data and equipped with a 128k context window for up to 90-minute inputs, MOSS Transcribe Diarize scales well and generalizes robustly. Across comprehensive evaluations, it outperforms state-of-the-art commercial systems on multiple public and in-house benchmarks.

Yiyang Su, Minchul Kim, Jie Zhu et al.

Open-set biometrics faces challenges with probe subjects who may not be enrolled in the gallery, as traditional biometric systems struggle to detect these non-mated probes. Despite the growing prevalence of multi-sample galleries in real-world deployments, most existing methods collapse intra-subject variability into a single global representation, leading to suboptimal decision boundaries and poor open-set robustness. To address this issue, we propose LocalScore, a simple yet effective scoring algorithm that explicitly incorporates the local density of the gallery feature distribution using the k-th nearest neighbors. LocalScore is architecture-agnostic, loss-independent, and incurs negligible computational overhead, making it a plug-and-play solution for existing biometric systems. Extensive experiments across multiple modalities demonstrate that LocalScore consistently achieves substantial gains in open-set retrieval (FNIR@FPIR reduced from 53% to 40%) and verification (TAR@FAR improved from 51% to 74%). We further provide theoretical analysis and empirical validation explaining when and why the method achieves the most significant gains based on dataset characteristics.

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.

Jie Zhu, Junhui Li, Yalong Wen et al.

Recent breakthroughs in large language models (LLMs) have led to the development of new benchmarks for evaluating their performance in the financial domain. However, current financial benchmarks often rely on news articles, earnings reports, or announcements, making it challenging to capture the real-world dynamics of financial meetings. To address this gap, we propose a novel benchmark called $\texttt{M$^3$FinMeeting}$, which is a multilingual, multi-sector, and multi-task dataset designed for financial meeting understanding. First, $\texttt{M$^3$FinMeeting}$ supports English, Chinese, and Japanese, enhancing comprehension of financial discussions in diverse linguistic contexts. Second, it encompasses various industry sectors defined by the Global Industry Classification Standard (GICS), ensuring that the benchmark spans a broad range of financial activities. Finally, $\texttt{M$^3$FinMeeting}$ includes three tasks: summarization, question-answer (QA) pair extraction, and question answering, facilitating a more realistic and comprehensive evaluation of understanding. Experimental results with seven popular LLMs reveal that even the most advanced long-context models have significant room for improvement, demonstrating the effectiveness of $\texttt{M$^3$FinMeeting}$ as a benchmark for assessing LLMs' financial meeting comprehension skills.

Jie Zhu, Bo Peng, Zhe Zhang et al.

Learning-based Multi-View Stereo (MVS) methods have made remarkable progress in recent years. However, how to effectively train the network without using real-world labels remains a challenging problem. In this paper, driven by the recent advancements of vision foundation models, a novel method termed DFM-MVS, is proposed to leverage the depth foundation model to generate the effective depth prior, so as to boost MVS in the absence of real-world labels. Specifically, a depth prior-based pseudo-supervised training mechanism is developed to simulate realistic stereo correspondences using the generated depth prior, thereby constructing effective supervision for the MVS network. Besides, a depth prior-guided error correction strategy is presented to leverage the depth prior as guidance to mitigate the error propagation problem inherent in the widely-used coarse-to-fine network structure. Experimental results on DTU and Tanks & Temples datasets demonstrate that the proposed DFM-MVS significantly outperforms existing MVS methods without using real-world labels.

Xiaoqing Lian, Jie Zhu, Tianxu Lv et al.

Significant differences in protein structures hinder the generalization of existing drug-target interaction (DTI) models, which often rely heavily on pre-learned binding principles or detailed annotations. In contrast, BioBridge designs an Inductive-Associative pipeline inspired by the workflow of scientists who base their accumulated expertise on drawing insights into novel drug-target pairs from weakly related references. BioBridge predicts novel drug-target interactions using limited sequence data, incorporating multi-level encoders with adversarial training to accumulate transferable binding principles. On these principles basis, BioBridge employs a dynamic prototype meta-learning framework to associate insights from weakly related annotations, enabling robust predictions for previously unseen drug-target pairs. Extensive experiments demonstrate that BioBridge surpasses existing models, especially for unseen proteins. Notably, when only homologous protein binding data is available, BioBridge proves effective for virtual screening of the epidermal growth factor receptor and adenosine receptor, underscoring its potential in drug discovery.

Jie Zhu, Huabin Huang, Banghuai Li et al.

Modern semantic segmentation methods devote much effect to adjusting image feature representations to improve the segmentation performance in various ways, such as architecture design, attention mechnism, etc. However, almost all those methods neglect the particularity of class weights (in the classification layer) in segmentation models. In this paper, we notice that the class weights of categories that tend to share many adjacent boundary pixels lack discrimination, thereby limiting the performance. We call this issue Boundary-caused Class Weights Confusion (BCWC). We try to focus on this problem and propose a novel method named Embedded Conditional Random Field (E-CRF) to alleviate it. E-CRF innovatively fuses the CRF into the CNN network as an organic whole for more effective end-to-end optimization. The reasons are two folds. It utilizes CRF to guide the message passing between pixels in high-level features to purify the feature representation of boundary pixels, with the help of inner pixels belonging to the same object. More importantly, it enables optimizing class weights from both scale and direction during backpropagation. We make detailed theoretical analysis to prove it. Besides, superpixel is integrated into E-CRF and served as an auxiliary to exploit the local object prior for more reliable message passing. Finally, our proposed method yields impressive results on ADE20K, Cityscapes, and Pascal Context datasets.