arXiv subjects

Liang He

Publications and source records attributed to Liang He.

At least 91 records · Page 5Linked to original sources

arXiv2026

Primal-dual dynamics featuring Hessian-driven damping and variable mass for convex optimization problems

This paper deals with a new Tikhonov regularized primal-dual dynamical system with variable mass and Hessian-driven damping for solving a convex optimization problem with linear equality constraints. The system features several time-dependent parameters: variable mass, slow viscous damping, extrapolation, and temporal scaling. By employing the Lyapunov analysis approach, we obtain the strong convergence of the trajectory generated by the proposed system to the minimal norm solution of the optimization problem, as well as convergence rate results for the primal-dual gap, the objective residual, and the feasibility violation. We also show that the convergence rates of the primal-dual gap, the objective residual, and the feasibility violation can be improved by appropriately adjusting these parameters. Further, we conduct numerical experiments to demonstrate the effectiveness of the theoretical results.

math.OC↗

arXiv2026

FED-Bench: A Cross-Granular Benchmark for Disentangled Evaluation of Facial Expression Editing

Facial expression image editing requires fine-grained control to strictly preserve human identity and background while precisely manipulating expression. However, existing editing benchmarks primarily focus on general scenarios, lacking high-quality facial images and corresponding editing instructions. Furthermore, current evaluation metrics exhibit systemic biases in this task, often favoring lazy editing or overfit editing. To bridge these gaps, we propose FED-Bench, a comprehensive benchmark featuring rigorous testing and an accurate evaluation suite. First, we carefully construct a benchmark of 747 triplets through a cascaded and scalable pipeline, each comprising an original image, an editing instruction, and a ground-truth image for precise evaluation. Second, we introduce FED-Score, a cross-granularity evaluation protocol that disentangles assessment into three dimensions: Alignment for verifying instruction following, Fidelity for testing image quality and identity preservation, and Relative Expression Gain for quantifying the magnitude of expression changes, effectively mitigating the aforementioned evaluation biases. Third, we benchmark 18 image editing models, revealing that current approaches struggle to simultaneously achieve high fidelity and accurate expression manipulation, with fine-grained instruction following identified as the primary bottleneck. Finally, leveraging the scalable characteristic of introduced benchmark engine, we provide a 20k+ in-the-wild facial training set and demonstrate its effectiveness by fine-tuning a baseline model that achieves significant performance gains. Our benchmark and related code will be made publicly open soon.

cs.CV↗

arXiv2026

Multi-material Direct Ink Writing and Embroidery for Stretchable Wearable Sensors

The development of wearable sensing systems for sports performance tracking, rehabilitation, and injury prevention has driven growing demand for smart garments that combine comfort, durability, and accurate motion detection. This paper presents a textile-compatible fabrication workflow that integrates multi-material direct ink writing with automated embroidery to create stretchable strain sensors directly embedded into garments. The process combines sequential multi-material printing of a silicone-carbon grease-silicone stack with automated embroidery that provides both mechanical fixation and electrical interfacing in a single step. The resulting hybrid sensor demonstrates stretchability up to 120% strain while maintaining electrical continuity, with approximately linear behaviour up to 60% strain (R^2 = 0.99), a gauge factor of 31.4, and hysteresis of 22.9%. Repeated loading-unloading tests over 80 cycles show baseline and peak drift of 0.135% and 0.236% per cycle, respectively, indicating moderate cycle-to-cycle stability. Mechanical testing further confirms that the silicone-fabric interface remains intact under large deformation, with failure occurring in the textile rather than at the stitched boundary. As a preliminary proof of concept, the sensor was integrated into wearable elbow and knee sleeves for joint angle monitoring, showing a clear correlation between normalised resistance change and bending angle. By addressing both mechanical fixation and electrical interfacing through embroidery-based integration, this approach provides a reproducible and scalable pathway for incorporating printed stretchable electronics into textile systems for motion capture and soft robotic applications.

cs.RO↗

arXiv2026

PRMB: Benchmarking Reward Models in Long-Horizon CBT-based Counseling Dialogue

Large language models (LLMs) hold potential for mental healthcare applications, particularly in cognitive behavioral therapy (CBT)-based counseling, where reward models play a critical role in aligning LLMs with preferred therapeutic behaviors. However, existing reward model evaluations often fail to capture alignment effectiveness in long-horizon interventions due to limited coverage of process-oriented datasets and misalignment between evaluation targets and psychological alignment objectives. To address these limitations, we present PRMB, a comprehensive benchmark tailored for evaluating reward models in multi-session CBT counseling. PRMB spans 6 sessions and 21 diverse negative scenarios, incorporating both pairwise and Best-of-N preference evaluations. We demonstrate a positive correlation between our benchmark and downstream counseling dialogue performance. Based on our benchmark, we conduct extensive analysis on the state-of-the-art reward models, revealing their generalization defects that were not discovered by previous benchmarks and highlighting the potential of generative reward models. Furthermore, we delve into examining the effectiveness of inference-time strategy for the evaluation of reward models and analyzing the impact factors of generative reward models. This work advances intelligent informatics for personalized healthcare by establishing a framework for reward model assessment in mental health dialogues. Evaluation code and datasets are publicly available at https://github.com/YouKenChaw/PRMB

cs.DB↗

arXiv2026

FEALPy: A Cross-platform Intelligent Numerical Simulation Engine

In resent years, the software ecosystem for numerical simulation still remains fragmented, with different algorithms and discretization methods often implemented in isolation, each with distinct data structures and programming conventions. This fragmentation is compounded by the growing divide between packages from different research fields and the lack of a unified, universal data structure, hindering the development of integrated, cross-platform solutions. In this work, we introduce FEALPy, a numerical simulation engine built around a unified tensor abstraction layer in a modular design. It enables seamless integration between diverse numerical methods along with deep learning workflows. By supporting multiple computational backends such as NumPy, PyTorch, and JAX, FEALPy ensures consistent adaptability across CPU and GPU hardware systems. Its modular architecture facilitates the entire simulation pipeline, from mesh handling and assembly to solver execution, with built-in support for automatic differentiation. In this paper, the versatility and efficacy of the framework are demonstrated through applications spanning linear elasticity, high-order PDEs, moving mesh methods, inverse problems and path planning.

math.NA↗

arXiv2026

Black-Hole Signatures in the Finite-Temperature Critical Ising Chain

We demonstrate that the finite-temperature critical transverse-field Ising chain exhibits quantitative signatures of black-hole physics in its dual gravitational description within the AdS/CFT correspondence. Its finite-temperature dynamics and thermodynamics are consistently captured by a mixed thermal-AdS/BTZ black hole saddle, leading to three mutually compatible observations. First, antipodal excitation transport collapses onto a universal temperature-dependent curve determined by the relative AdS and BTZ contributions to the gravitational partition function, reflecting horizon absorption. Second, in the high-temperature regime, the retarded response exhibits exponential relaxation governed by the lowest quasi-normal mode of the dual black hole. Third, the temperature derivative of the von Neumann entropy develops a pronounced minimum at a temperature consistent with the Hawking-Page transition. These results identify critical quantum spin chains as minimal and experimentally accessible platforms for probing dynamical and thermodynamic aspects of quantum black holes in controllable many-body systems.

quant-ph↗

arXiv2026

AutoSkill: Experience-Driven Lifelong Learning via Skill Self-Evolution

In practical LLM applications, users repeatedly express stable preferences and requirements, such as reducing hallucinations, following institutional writing conventions, or avoiding overly technical wording, yet such interaction experience is seldom consolidated into reusable knowledge. Consequently, LLM agents often fail to accumulate personalized capabilities across sessions. We present AutoSkill, an experience-driven lifelong learning framework that enables LLM agents to automatically derive, maintain, and reuse skills from dialogue and interaction traces. AutoSkill abstracts skills from user experience, supports their continual self-evolution, and dynamically injects relevant skills into future requests without retraining the underlying model. Designed as a model-agnostic plugin layer, it is compatible with existing LLMs and introduces a standardized skill representation for sharing and transfer across agents, users, and tasks. In this way, AutoSkill turns ephemeral interaction experience into explicit, reusable, and composable capabilities. This paper describes the motivation, architecture, skill lifecycle, and implementation of AutoSkill, and positions it with respect to prior work on memory, retrieval, personalization, and agentic systems. AutoSkill highlights a practical and scalable path toward lifelong personalized agents and personal digital surrogates.

cs.AI↗

arXiv2026

Inertial dynamical systems and accelerated algorithms with implicit Hessian-driven damping for nonconvex optimization

This paper is devoted to the investigation of inertial dynamical systems with implicit Hessian-driven damping for strongly quasiconvex optimization which is a specific class of nonconvex optimization problems. We first establish exponential convergence rate properties for this system without requiring Lipschitz continuity of the gradient on the function. Then, we obtain an inertial accelerated algorithm for minimizing strongly quasiconvex functions through natural explicit time discretization to the dynamical system. Meanwhile, we consider an exogenous additive perturbation term to this dynamical system and obtain the corresponding algorithm. By utilizing the Lyapunov method, we establish convergence rates of iterative sequences and their function values. Furthermore, we conduct numerical experiments to illustrate the theoretical results.

math.OC↗

arXiv2026

Dynamic Multimodal Activation Steering for Hallucination Mitigation in Large Vision-Language Models

Large Vision-Language Models (LVLMs) exhibit outstanding performance on vision-language tasks but struggle with hallucination problems. Through in-depth analysis of LVLM activation patterns, we reveal two key findings: 1) truthfulness and visual perception capabilities predominantly engage different subsets of attention heads within the model architecture; and 2) truthfulness steering vectors vary significantly across different semantic contexts. Based on these observations, we propose Dynamic Multimodal Activation Steering, a training-free approach for hallucination mitigation. Our method constructs a semantic-based truthfulness steering vector database and computes visual perception steering vectors, enabling context-aware interventions during inference by dynamically selecting the most relevant steering vectors based on input semantic similarity and applying them to the most influential attention heads. We conduct comprehensive experiments across multiple models and datasets, demonstrating that our approach significantly enhances model performance, outperforming existing state-of-the-art methods.

cs.CV↗

arXiv2026

A Survey on the Optimization of Large Language Model-based Agents

With the rapid development of Large Language Models (LLMs), LLM-based agents have been widely adopted in various fields, becoming essential for autonomous decision-making and interactive tasks. However, current work typically relies on prompt design or fine-tuning strategies applied to vanilla LLMs, which often leads to limited effectiveness or suboptimal performance in complex agent-related environments. Although LLM optimization techniques can improve model performance across many general tasks, they lack specialized optimization towards critical agent functionalities such as long-term planning, dynamic environmental interaction, and complex decision-making. Although numerous recent studies have explored various strategies to optimize LLM-based agents for complex agent tasks, a systematic review summarizing and comparing these methods from a holistic perspective is still lacking. In this survey, we provide a comprehensive review of LLM-based agent optimization approaches, categorizing them into parameter-driven and parameter-free methods. We first focus on parameter-driven optimization, covering fine-tuning-based optimization, reinforcement learning-based optimization, and hybrid strategies, analyzing key aspects such as trajectory data construction, fine-tuning techniques, reward function design, and optimization algorithms. Additionally, we briefly discuss parameter-free strategies that optimize agent behavior through prompt engineering and external knowledge retrieval. Finally, we summarize the datasets and benchmarks used for evaluation and tuning, review key applications of LLM-based agents, and discuss major challenges and promising future directions. Our repository for related references is available at https://github.com/YoungDubbyDu/LLM-Agent-Optimization.

cs.AI↗

arXiv2026

Suppression of Decoherence at Exceptional Transitions

Decoherence is strongly influenced by environmental criticality, with conventional Hermitian critical points typically enhancing the loss of quantum coherence. Here, we show that this paradigm is fundamentally altered in non-Hermitian environments. Focusing on qubits coupled to non-Hermitian spin chains and interacting ultracold Fermi gases, we find that approaching exceptional points can either enhance or strongly suppress decoherence, depending on the balance between Hermitian and non-Hermitian system-environment couplings. In particular, when these couplings are comparable, decoherence is dramatically suppressed at exceptional transitions. We trace this behavior to the distinct response of the environmental ground state near non-Hermitian degeneracies and demonstrate the robustness of this effect across multiple models. Finally, we show that the predicted suppression of decoherence is directly observable on current digital quantum simulation platforms. Our results establish exceptional points as a concrete mechanism for suppressing decoherence and identify non-Hermitian criticality as a new avenue for coherence control in open quantum systems and quantum technologies.

quant-ph↗

arXiv2026

Not Seeing the Whole Picture: Challenges and Opportunities in Using AI for Co-Making Physical DIY-AT for People with Visual Impairments

Existing assistive technologies (AT) often adopt a one-size-fits-all approach, overlooking the diverse needs of people with visual impairments (PVI). Do-it-yourself AT (DIY-AT) toolkits offer one path toward customization, but most remain limited--targeting co-design with engineers or requiring programming expertise. Non-professionals with disabilities, including PVI, also face barriers such as inaccessible tools, lack of confidence, and insufficient technical knowledge. These gaps highlight the need for prototyping technologies that enable PVI to directly make their own AT. Building on emerging evidence that large language models (LLMs) can serve not only as visual aids but also as co-design partners, we present an exploratory study of how LLM-based AI can support PVI in the tangible DIY-AT co-making process. Our findings surface key challenges and design opportunities: the need for greater spatial and visual support, strategies for mitigating novel AI errors, and implications for designing more accessible AI-assisted prototypes.

cs.HC↗

arXiv2026

Easy-Poly: An Easy Polyhedral Framework For 3D Multi-Object Tracking

Recent 3D multi-object tracking (3D MOT) methods mainly follow tracking-by-detection pipelines, but often suffer from high false positives, missed detections, and identity switches, especially in crowded and small-object scenarios. To address these challenges, we propose Easy-Poly, a filter-based 3D MOT framework with four key innovations: (1) CNMSMM, a novel Camera-LiDAR fusion detection method combining multi-modal augmentation and an efficient NMS with a new loss function to improve small target detection; (2) Dynamic Track-Oriented (DTO) data association that robustly handles uncertainties and occlusions via class-aware optimal assignment and parallel processing strategies; (3) Dynamic Motion Modeling (DMM) using a confidence-weighted Kalman filter with adaptive noise covariance to enhance tracking accuracy; and (4) an extended life-cycle management system reducing identity switches and false terminations. Experimental results show that Easy-Poly outperforms state-of-the-art methods such as Poly-MOT and Fast-Poly, achieving notable gains in mAP (e.g., from 63.30% to 65.65% with LargeKernel3D) and AMOTA (e.g., from 73.1% to 75.6%), while also running in real-time. Our framework advances robustness and adaptability in complex driving environments, paving the way for safer autonomous driving perception.

cs.CV↗

arXiv2026

InternAgent-1.5: A Unified Agentic Framework for Long-Horizon Autonomous Scientific Discovery

We introduce InternAgent-1.5, a unified system designed for end-to-end scientific discovery across computational and empirical domains. The system is built on a structured architecture composed of three coordinated subsystems for generation, verification, and evolution. These subsystems are supported by foundational capabilities for deep research, solution optimization, and long horizon memory. The architecture allows InternAgent-1.5 to operate continuously across extended discovery cycles while maintaining coherent and improving behavior. It also enables the system to coordinate computational modeling and laboratory experimentation within a single unified system. We evaluate InternAgent-1.5 on scientific reasoning benchmarks such as GAIA, HLE, GPQA, and FrontierScience, and the system achieves leading performance that demonstrates strong foundational capabilities. Beyond these benchmarks, we further assess two categories of discovery tasks. In algorithm discovery tasks, InternAgent-1.5 autonomously designs competitive methods for core machine learning problems. In empirical discovery tasks, it executes complete computational or wet lab experiments and produces scientific findings in earth, life, biological, and physical domains. Overall, these results show that InternAgent-1.5 provides a general and scalable framework for autonomous scientific discovery.

cs.AI↗

arXiv2026

FluxLab: Creating 3D Printable Shape-Changing Devices with Integrated Deformation Sensing

We present FluxLab, a system comprising interactive tools for creating custom 3D-printable shape-changing devices with integrated deformation sensing. To achieve this, we propose a 3D printable nesting structure, consisting of a central SMA channel for sensing and actuation, lattice-based padding in the middle for structural support and controllable elasticity, and parallel helix-based surface wires that preserve the overall form and provide anchoring struts for guided deformation. We developed a design editor to embed these structures into custom 3D models for printing with elastic silicone resin on a consumer-grade SLA 3D printer and minimal post-printing assembly. A deformation authoring tool was also developed for users to build a machine learning-based classifier that distinguishes desired deformation behaviors using inductive sensing. Finally, we demonstrate the potential of our system through example applications, including a self-deformable steamer bowl clip, a remotely controllable gripper, and an interactive desk lamp.

cs.HC↗

arXiv2026

Hyperbolic Additive Margin Softmax with Hierarchical Information for Speaker Verification

Speaker embedding learning based on Euclidean space has achieved significant progress, but it is still insufficient in modeling hierarchical information within speaker features. Hyperbolic space, with its negative curvature geometric properties, can efficiently represent hierarchical information within a finite volume, making it more suitable for the feature distribution of speaker embeddings. In this paper, we propose Hyperbolic Softmax (H-Softmax) and Hyperbolic Additive Margin Softmax (HAM-Softmax) based on hyperbolic space. H-Softmax incorporates hierarchical information into speaker embeddings by projecting embeddings and speaker centers into hyperbolic space and computing hyperbolic distances. HAM-Softmax further enhances inter-class separability by introducing margin constraint on this basis. Experimental results show that H-Softmax and HAM-Softmax achieve average relative EER reductions of 27.84% and 14.23% compared with standard Softmax and AM-Softmax, respectively, demonstrating that the proposed methods effectively improve speaker verification performance and at the same time preserve the capability of hierarchical structure modeling. The code will be released at https://github.com/PunkMale/HAM-Softmax.

cs.SD↗

arXiv2026

Building Self-Evolving Agents via Experience-Driven Lifelong Learning: A Framework and Benchmark

As AI advances toward general intelligence, the focus is shifting from systems optimized for static tasks to creating open-ended agents that learn continuously. In this paper, we introduce Experience-driven Lifelong Learning (ELL), a framework for building self-evolving agents capable of continuous growth through real-world interaction. The framework is built on four core principles: (1) Experience Exploration: Agents learn through continuous, self-motivated interaction with dynamic environments, navigating interdependent tasks and generating rich experiential trajectories. (2) Long-term Memory: Agents preserve and structure historical knowledge, including personal experiences, domain expertise, and commonsense reasoning, into a persistent memory system. (3) Skill Learning: Agents autonomously improve by abstracting recurring patterns from experience into reusable skills, which are actively refined and validated for application in new tasks. (4) Knowledge Internalization: Agents internalize explicit and discrete experiences into implicit and intuitive capabilities as "second nature". We also introduce StuLife, a benchmark dataset for ELL that simulates a student's holistic college journey, from enrollment to academic and personal development, across three core phases and ten detailed sub-scenarios. StuLife is designed around three key paradigm

cs.AI↗

arXiv2026

PsychEval: A Multi-Session and Multi-Therapy Benchmark for High-Realism AI Psychological Counselor

To develop a reliable AI for psychological assessment, we introduce \texttt{PsychEval}, a multi-session, multi-therapy, and highly realistic benchmark designed to address three key challenges: \textbf{1) Can we train a highly realistic AI counselor?} Realistic counseling is a longitudinal task requiring sustained memory and dynamic goal tracking. We propose a multi-session benchmark (spanning 6-10 sessions across three distinct stages) that demands critical capabilities such as memory continuity, adaptive reasoning, and longitudinal planning. The dataset is annotated with extensive professional skills, comprising over 677 meta-skills and 4577 atomic skills. \textbf{2) How to train a multi-therapy AI counselor?} While existing models often focus on a single therapy, complex cases frequently require flexible strategies among various therapies. We construct a diverse dataset covering five therapeutic modalities (Psychodynamic, Behaviorism, CBT, Humanistic Existentialist, and Postmodernist) alongside an integrative therapy with a unified three-stage clinical framework across six core psychological topics. \textbf{3) How to systematically evaluate an AI counselor?} We establish a holistic evaluation framework with 18 therapy-specific and therapy-shared metrics across Client-Level and Counselor-Level dimensions. To support this, we also construct over 2,000 diverse client profiles. Extensive experimental analysis fully validates the superior quality and clinical fidelity of our dataset. Crucially, \texttt{PsychEval} transcends static benchmarking to serve as a high-fidelity reinforcement learning environment that enables the self-evolutionary training of clinically responsible and adaptive AI counselors.

cs.AI↗