arXiv subjects

He Hu

Publications and source records attributed to He Hu.

16 recordsLinked to original sources

arXiv2026

Attention Sensitivity Is Not Enough: Dissociating Attention-Level and Behavioural In-Context Learning under Fine-Tuning

In-context learning (ICL) lets large language models adapt to new tasks from demonstrations, and fine-tuning can erode this behaviour. Many preservation diagnostics inspect attention: if attention changes when demonstrations change, the model is treated as context-sensitive. This paper asks how far that proxy can be trusted once it is optimised. We formalise \emph{In-Context Sensitivity} (ICS), the average row distance between last-token attention on matched and mismatched demonstration prefixes, and pair it with \emph{ICL-GAP}, the behavioural accuracy gap between the same prefixes. In a controlled four-arm ablation on Llama-2-7B, an ICS-maximising regulariser ($\armKL$) drives ICS to $1.413$, within $0.5\%$ of its geometric ceiling. The behavioural readout tells a different story: ICL-GAP stays near zero and MMLU accuracy moves from $0.371$ to $0.279$, a Goodhart dissociation of the bounded attention proxy. Endpoint statistics locate the mechanism: attention grows sharp and near-disjoint across prefixes yet routes to formatting and demonstration-body tokens rather than labels. A random-label protocol confirms that the behavioural probe family retains dynamic range at the same checkpoints. In a constructive sweep, behaviour gating partially mitigates the effect, while objectives anchored to pretrained computation hold the high-MMLU, moderate-ICS region that divergence maximisers leave. The main lesson is diagnostic: attention-level ICL proxies earn their place as training targets only after validation against behavioural gaps.

cs.LG↗

arXiv2026

Neither Precision Nor Architecture Alone: Controlled Tests of Failure Remedies for Physics-Informed Neural Networks

Physics-Informed Neural Networks (PINNs) frequently fail on stiff or advection-dominated PDEs, and two recent accounts offer competing remedies: switching from FP32 to FP64 to repair an L-BFGS stopping artifact, or replacing the MLP with a state-space-model (SSM) backbone plus sub-sequence alignment to counter architectural simplicity bias. We test both under matched, seed-paired controls in a pre-registered 144-run study spanning convection, reaction, and wave, plus an independent 85-run convection/wave study; success is relative $\ell_2$ error below $0.05$. The two remedies act on disjoint regime-and-seed slices: neither substitutes for the other. On hard convection ($\beta{=}50$), alignment recovers 2/5 seeds in FP32 and 3/5 in FP64, where the unaligned SSM succeeds on 0/5 seeds at either precision and the vanilla MLP moves only from 0/5 to 1/5 across the precision switch---the recoveries trace to the alignment objective, not the backbone. On reaction the backbone alone already succeeds on 3/5--4/5 seeds, so each remedy covers a regime the other does not. Responses are also seed-specific: the same precision switch flips individual seeds in opposite directions and, on wave, lowers median error with no statistically significant success gain. Tightening the inner L-BFGS tolerance in an independent repeated-step runner likewise lowers median error at a large runtime cost, with success counts unchanged. Precision, stopping, backbone, and alignment must therefore be evaluated jointly and reported per seed.

cs.LG↗

arXiv2026

HOIMask: Towards Generative Masked Modeling for Human Object Interaction Generation

Diffusion-based methods have dominated the HOI generation, as they enable critical contact fusions or signals to guide the diffusion process. However, they often result in high artifacts and unstable interaction quality due to error accumulation during iterative denoising. In this work, we propose HOIMask, the first generative masked framework for modeling HOI motion in discrete space. HOIMask first encodes both motion sequences and contact-aware signals into discrete 2D human and object token maps via HOI Vector Quantization (VQ), preserving fine-grained spatial-temporal structure beyond conventional 1D representations. On this basis, a generative masked modeling framework is employed to jointly capture human-object interaction dynamics, leveraging a transformer architecture designed to model complex spatial-temporal and interaction dependencies. To generate more coherent and physically plausible motions, we further introduce a novel contact-aware reconstruction guidance in discrete space during inference, which fuses contact signals to optimize HOI tokens that forces the generated motion with higher spatio-temporal consistency. With craftily designed motion interaction tokens, dedicated architecture and guidance strategy, HOIMask outperforms state-of-the-art diffusion-based methods, generating more realistic and semantically aligned HOI motions. Please refer to https://jyhflash.github.io/HOIMask/ for more results.

cs.CV↗

arXiv2026

EmoTrans: A Benchmark for Understanding, Reasoning, and Predicting Emotion Transitions in Multimodal LLMs

Recent multimodal large language models (MLLMs) have shown strong capabilities in perception, reasoning, and generation, and are increasingly used in applications such as social robots and human-computer interaction, where understanding human emotions is essential. However, existing benchmarks mainly formulate emotion understanding as a static recognition problem, leaving it largely unclear whether current MLLMs can understand emotion as a dynamic process that evolves, shifts between states, and unfolds across diverse social contexts. To bridge this gap, we present EmoTrans, a benchmark for evaluating emotion dynamics understanding in multimodal videos. EmoTrans contains 1,000 carefully collected and manually annotated video clips, covering 12 real-world scenarios, and further provides over 3,000 task-specific question-answer (QA) pairs for fine-grained evaluation. The benchmark introduces four tasks, namely Emotion Change Detection (ECD), Emotion State Identification (ESI), Emotion Transition Reasoning (ETR), and Next Emotion Prediction (NEP), forming a progressive evaluation framework from coarse-grained detection to deeper reasoning and prediction. We conduct a comprehensive evaluation of 18 state-of-the-art MLLMs on EmoTrans and obtain two main findings. First, although current MLLMs show relatively stronger performance on coarse-grained emotion change detection, they still struggle with fine-grained emotion dynamics modeling. Second, socially complex settings, especially multi-person scenarios, remain substantially challenging, while reasoning-oriented variants do not consistently yield clear improvements. To facilitate future research, we publicly release the benchmark, evaluation protocol, and code at https://github.com/Emo-gml/EmoTrans.

cs.CV↗

arXiv2026

A Principle-Driven Adaptive Policy for Group Cognitive Stimulation Dialogue for Elderly with Cognitive Impairment

Cognitive impairment is becoming a major public health challenge. Cognitive Stimulation Therapy (CST) is an effective intervention for cognitive impairment, but traditional methods are difficult to scale, and existing digital systems struggle with group dialogues and cognitive stimulation principles. While Large Language Models (LLMs) are powerful, their application in this context faces key challenges: cognitive stimulation dialogue paradigms, a lack of therapeutic reasoning, and static-only user modeling. To address these issues, we propose a principle-driven adaptive policy actualized through a Group Cognitive Stimulation Dialogue (GCSD) system. We first construct a dataset with over 500 hours of real-world CST conversations and 10,000+ simulated dialogues generated via our Principle-Guided Scenario Simulation strategy. Our GCSD system then integrates four core modules to overcome LLM limitations: (i) a multi-speaker context controller to resolve role confusion; (ii) dynamic participant cognitive state modeling for personalized interaction; (iii) a cognitive stimulation-focused attention loss to instill cognitive stimulation reasoning; and (iv) a multi-dimensional reward strategy to enhance response value. Experimental results demonstrate that GCSD significantly outperforms baseline models across various evaluation metrics. Future work will focus on long-term clinical validation to bridge the gap between computational performance and clinical efficacy.

cs.CL↗

arXiv2025

Investigating Data Pruning for Pretraining Biological Foundation Models at Scale

Biological foundation models (BioFMs), pretrained on large-scale biological sequences, have recently shown strong potential in providing meaningful representations for diverse downstream bioinformatics tasks. However, such models often rely on millions to billions of training sequences and billions of parameters, resulting in prohibitive computational costs and significant barriers to reproducibility and accessibility, particularly for academic labs. To address these challenges, we investigate the feasibility of data pruning for BioFM pretraining and propose a post-hoc influence-guided data pruning framework tailored to biological domains. Our approach introduces a subset-based self-influence formulation that enables efficient estimation of sample importance at low computational cost, and builds upon it two simple yet effective selection strategies, namely Top-k Influence (Top I) and Coverage-Centric Influence (CCI). We empirically validate our method on two representative BioFMs, RNA-FM and ESM-C. For RNA, our framework consistently outperforms random selection baselines under an extreme pruning rate of over 99 percent, demonstrating its effectiveness. Furthermore, we show the generalizability of our framework on protein-related tasks using ESM-C. In particular, our coreset even outperforms random subsets that are ten times larger in both RNA and protein settings, revealing substantial redundancy in biological sequence datasets. These findings underscore the potential of influence-guided data pruning to substantially reduce the computational cost of BioFM pretraining, paving the way for more efficient, accessible, and sustainable biological AI research.

cs.LG↗

arXiv2025

TheraMind: A Strategic and Adaptive Agent for Longitudinal Psychological Counseling

The shortage of mental health professionals has driven the web to become a primary avenue for accessible psychological support. While Large Language Models (LLMs) offer promise for scalable web-based counseling, existing approaches often lack emotional understanding, adaptive strategies, and long-term memory. These limitations pose risks to digital well-being, as disjointed interactions can fail to support vulnerable users effectively. To address these gaps, we introduce TheraMind, a strategic and adaptive agent designed for trustworthy online longitudinal counseling. The cornerstone of TheraMind is a novel dual-loop architecture that decouples the complex counseling process into an Intra-Session Loop for tactical dialogue management and a Cross-Session Loop for strategic therapeutic planning. The Intra-Session Loop perceives the patient's emotional state to dynamically select response strategies while leveraging cross-session memory to ensure continuity. Crucially, the Cross-Session Loop empowers the agent with long-term adaptability by evaluating the efficacy of the applied therapy after each session and adjusting the method for subsequent interactions. We validate our approach in a high-fidelity simulation environment grounded in real clinical cases. Extensive evaluations show that TheraMind outperforms other methods, especially on multi-session metrics like Coherence, Flexibility, and Therapeutic Attunement, validating the effectiveness of its dual-loop design in emulating strategic, adaptive, and longitudinal therapeutic behavior. The code is publicly available at https://github.com/Emo-gml/TheraMind.

cs.AI↗

arXiv2025

Beyond Empathy: Integrating Diagnostic and Therapeutic Reasoning with Large Language Models for Mental Health Counseling

Large language models (LLMs) hold significant potential for mental health support, capable of generating empathetic responses and simulating therapeutic conversations. However, existing LLM-based approaches often lack the clinical grounding necessary for real-world psychological counseling, particularly in explicit diagnostic reasoning aligned with standards like the DSM/ICD and incorporating diverse therapeutic modalities beyond basic empathy or single strategies. To address these critical limitations, we propose PsyLLM, the first large language model designed to systematically integrate both diagnostic and therapeutic reasoning for mental health counseling. To develop PsyLLM, we design a novel automated data synthesis pipeline that processes real-world mental health posts collected from Reddit, where users frequently share psychological distress and seek community support. This pipeline processes real-world mental health posts, generates multi-turn dialogue structures, and leverages LLMs guided by international diagnostic standards (e.g., DSM/ICD) and multiple therapeutic frameworks (e.g., CBT, ACT, psychodynamic) to simulate detailed clinical reasoning processes. Rigorous multi-dimensional filtering ensures the generation of high-quality, clinically aligned dialogue data. In addition, we introduce a new benchmark and evaluation protocol, assessing counseling quality across four key dimensions. Our experiments demonstrate that PsyLLM significantly outperforms state-of-the-art baseline models on this benchmark. The model weights and dataset have been publicly released at https://github.com/Emo-gml/PsyLLM.

cs.CL↗

arXiv2025

EmoBench-M: Benchmarking Emotional Intelligence for Multimodal Large Language Models

With the integration of multimodal large language models (MLLMs) into robotic systems and AI applications, embedding emotional intelligence (EI) capabilities is essential for enabling these models to perceive, interpret, and respond to human emotions effectively in real-world scenarios. Existing static, text-based, or text-image benchmarks overlook the multimodal complexities of real interactions and fail to capture the dynamic, context-dependent nature of emotional expressions, rendering them inadequate for evaluating MLLMs' EI capabilities. To address these limitations, we introduce EmoBench-M, a systematic benchmark grounded in established psychological theories, designed to evaluate MLLMs across 13 evaluation scenarios spanning three hierarchical dimensions: foundational emotion recognition (FER), conversational emotion understanding (CEU), and socially complex emotion analysis (SCEA). Evaluation was conducted on 27 state-of-the-art MLLMs, using both objective task-specific metrics and LLM-based evaluation, revealing a substantial performance gap relative to human-level competence. Even the best performing models, Gemini-3.0-Pro and GPT-5.2, achieve the highest scores on EmoBench-M, 70.5 and 66.5 points respectively. Specialized models such as AffectGPT exhibit uneven performance across EmoBench-M, demonstrating strengths in certain scenarios but generally lacking comprehensive emotional intelligence. By providing a comprehensive, multimodal evaluation framework, EmoBench-M captures both the strengths and weaknesses of current MLLMs across diverse emotional contexts. All benchmark resources, including datasets and code, are publicly available at https://emo-gml.github.io/, facilitating further research and advancement in MLLM emotional intelligence.

cs.CL↗

arXiv2024

MemoryMamba: Memory-Augmented State Space Model for Defect Recognition

As automation advances in manufacturing, the demand for precise and sophisticated defect detection technologies grows. Existing vision models for defect recognition methods are insufficient for handling the complexities and variations of defects in contemporary manufacturing settings. These models especially struggle in scenarios involving limited or imbalanced defect data. In this work, we introduce MemoryMamba, a novel memory-augmented state space model (SSM), designed to overcome the limitations of existing defect recognition models. MemoryMamba integrates the state space model with the memory augmentation mechanism, enabling the system to maintain and retrieve essential defect-specific information in training. Its architecture is designed to capture dependencies and intricate defect characteristics, which are crucial for effective defect detection. In the experiments, MemoryMamba was evaluated across four industrial datasets with diverse defect types and complexities. The model consistently outperformed other methods, demonstrating its capability to adapt to various defect recognition scenarios.

cs.CV↗

arXiv2024

Data-CUBE: Data Curriculum for Instruction-based Sentence Representation Learning

Recently, multi-task instruction tuning has been applied into sentence representation learning, which endows the capability of generating specific representations with the guidance of task instruction, exhibiting strong generalization ability on new tasks. However, these methods mostly neglect the potential interference problems across different tasks and instances, which may affect the training and convergence of the model. To address it, we propose a data curriculum method, namely Data-CUBE, that arranges the orders of all the multi-task data for training, to minimize the interference risks from the two views. In the task level, we aim to find the optimal task order to minimize the total cross-task interference risk, which is exactly the traveling salesman problem, hence we utilize a simulated annealing algorithm to find its solution. In the instance level, we measure the difficulty of all instances per task, then divide them into the easy-to-difficult mini-batches for training. Experiments on MTEB sentence representation evaluation tasks show that our approach can boost the performance of state-of-the-art methods. Our code and data are publicly available at the link: \url{https://github.com/RUCAIBox/Data-CUBE}.

cs.CL↗

arXiv2022

C2-CRS: Coarse-to-Fine Contrastive Learning for Conversational Recommender System

Conversational recommender systems (CRS) aim to recommend suitable items to users through natural language conversations. For developing effective CRSs, a major technical issue is how to accurately infer user preference from very limited conversation context. To address issue, a promising solution is to incorporate external data for enriching the context information. However, prior studies mainly focus on designing fusion models tailored for some specific type of external data, which is not general to model and utilize multi-type external data. To effectively leverage multi-type external data, we propose a novel coarse-to-fine contrastive learning framework to improve data semantic fusion for CRS. In our approach, we first extract and represent multi-grained semantic units from different data signals, and then align the associated multi-type semantic units in a coarse-to-fine way. To implement this framework, we design both coarse-grained and fine-grained procedures for modeling user preference, where the former focuses on more general, coarse-grained semantic fusion and the latter focuses on more specific, fine-grained semantic fusion. Such an approach can be extended to incorporate more kinds of external data. Extensive experiments on two public CRS datasets have demonstrated the effectiveness of our approach in both recommendation and conversation tasks.

cs.CL↗

arXiv2021

Exotic dielectric behaviors induced by pseudo-spin texture in magnetic twisted bilayer

Twisted van der Waals bilayers provide an ideal platform to study the electron correlation in solids. Of particular interest is the 30 degree twisted bilayer honeycomb lattice system, which possesses an incommensurate moire pattern and uncommon electronic behaviors may appear due to the absence of phase coherence. Such system is extremely sensitive to further twist and many intriguing phenomena will occur. In this work, based on first-principles calculations we show that, for further twist near 30 degree, there could induce dramatically different dielectric behaviors of electron between left and right twisted cases. Specifically, it is found that the left and right twists show suppressed and amplified dielectric response under vertical electric field, respectively. Further analysis demonstrate that such exotic dielectric property can be attributed to the stacking dependent charge redistribution due to twist, which forms twist-dependent pseudospin textures. We will show that such pseudospin textures are robust under small electric field. As a result, for the right twisted case, there is almost no electric dipole formation exceeding the monolayer thickness when the electric field is applied. Whereas for the left case, the system could even demonstrate negative susceptibility, i.e. the induced polarization is opposite to the applied field, which is very rare in the nature. Such findings not only enrich our understanding on moire systems but also open an appealing route toward functional 2D materials design for electronic, optical and even energy storage devices.

cond-mat.mtrl-sci↗

arXiv2020

Concept of the half-valley-metal and quantum anomalous valley Hall effect

Valley, the energy extrema in the electronic band structure at momentum space, is regarded as a new degree of freedom of electrons, in addition to charge and spin. The studies focused on valley degree of freedom now form an emerging field of condensed matter physics, i.e. valleytronics, whose development is exactly following that of spintronics which focuses on the spin degree of freedom. Here, in analogy to half-metals in spintronics with one spin channel is conducting whereas the other is insulating, we propose the concept of half-valley-metal, in which conduction electrons are intrinsically 100% valley polarized, as well as 100% spin-polarized even when spin-orbit interactions are considered. Combining first-principles calculations with two-band kp model, the physical mechanism to form the half-valley-metal is illuminated. Taking the ferrovalley H-FeCl2 monolayer with strong exchange interaction as an example, we find that the strong electron correlation effect can induce the ferrovalley to half-valley-metal transition. Due to the valley-dependent optical selection rules, such system could be transparent to, e.g., left-circularly polarized light, yet the right-circularly polarized light will be reflected, which can in turn be used as a crucial method to detect half-valley-metal state. In addition, we find that in the so obtained half-valley-metal state, the conduction valley demonstrates Dirac cone-like linear energy dispersion. Interestingly, with the increase of the correlation effect, the system becomes insulating again with all valleys follow same optical selection rule. We confirm that in this specific case, the valence bands, which consist of single spin, possess non-zero Chern number and consequently intrinsic quantum anomalous valley Hall effect emerges. Our findings open an appealing route toward functional 2D materials design of valleytronics.

cond-mat.mtrl-sci↗

arXiv2019

Emergent exotic chirality dependent dielectricity in magnetic twisted bilayer system

Twisted van der Waals bilayers provide an ideal platform to study the electron correlation in solids. Of particular interest is the 30 degree twisted bilayer honeycomb lattice system, which possesses an incommensurate Moire pattern and uncommon electronic behaviors may appear due to the absence of phase coherence. Such system is extremely sensitive to further twist and many intriguing phenomena will occur. In this work, we show that due to the twist induced spatial inhomogeneity of interlayer coupling, there emerges an U(1) gauge field in magnetic transition-metal dichalcogenides (TMD) bilayers. Interestingly, for further twist near 30 degree, the induced gauge field could form a chirality dependent real-space skyrmion pattern, or magnetic charge. Moreover, such twist also induces the topology dependent electronic polarization of the bilayer system through the nonzero flux of the real-space Berry curvature. Further analysis proves that the antiferromagnetically coupled twisted bilayer system is indeed also antiferroelectric! When an external electric field is applied to break the potential balance between layers, there will emerge novel magnetoelectric coupling and exotic chirality dependent dielectricity. Such findings not only enrich our understanding on Moire systems, but also open an appealing route toward functional 2D materials design for electronic, optical and even energy storage devices.

cond-mat.mtrl-sci↗

arXiv2018

Giant dielectric difference in chiral asymmetric bilayers

Twistronics rooted in the twist operation towards bilayer van der Waals crystals is of both theoretical and technological importance. The realize of the correlated electronic behaviors under this operation encourages enormous effort to the research on magic-angle systems which possess sensitive response to the external field. Here, a giant dielectric difference between 30 plus or minus degree twist case is observed in a typical magnetic system 2H-VSe2 bilayer. It is shown that due to the structural inversion asymmetry in its monolayer, the different stacking of the two cases corresponds to the two kind of valley polarized states: interlayer ferrovalley and interlayer antiferrovalley. Further investigations reveal that such different dielectric response between the two states stems from the different Fermi wave vectors coupled to the electric field. More interestingly, we even obtain the selective circularly polarized optical absorption by tuning the interlayer twist. These findings open an appealing route toward functional 2D materials design for electric and optical devices.

cond-mat.mes-hall↗