arXiv subjects

Bin Li

Publications and source records attributed to Bin Li.

At least 109 records · Page 6Linked to original sources

arXiv2026

ScribbleDose: Scribble-Guided Dose Prediction in Radiotherapy

Anatomical structure masks are widely adopted in radiotherapy dose prediction, as they provide explicit geometric constraints that facilitate structure-dose coupling. However, conventional manual delineation of these masks requires precise annotation of structure boundaries relevant to radiotherapy, which is time-consuming and labor-intensive. To address these limitations, we propose a scribble-guided dose prediction framework that relies solely on anatomical structures annotated with sparse scribbles. Specifically, we design a Scribble Completion Module (SCM) to generate dense anatomical masks by propagating sparse scribble labels to semantically similar voxels. During the propagation process, a supervoxel-based regularization is introduced to preserve geometric boundary consistency to ensure anatomical plausibility. Furthermore, we propose a Structure-Guided Dose Generation Module (SGDGM) to strengthen the correspondence between sparse structural cues and dose distribution. Herein, the completed dense masks derived from scribbles serve as structural guidance to condition the dose prediction network. This scribble-mask-dose consistency encourages high-dose concentration within target volumes while effectively sparing surrounding organs-at-risk. Extensive experiments on the open-source GDP-HMM dataset demonstrate that the proposed method maintains superior dose prediction performance while substantially reducing annotation cost, providing a practical paradigm for dose prediction under sparse structural annotation. The code and reannotated scribbles are made publicly available at https://github.com/iCherishxixixi/ScribbleDose.

cs.CV↗

arXiv2026

VIMCAN: Visual-Inertial 3D Human Pose Estimation with Hybrid Mamba-Cross-Attention Network

The rapid advances in deep learning have significantly enhanced the accuracy of multimodal 3D human pose estimation (HPE). However, the state-of-the-art (SOTA) HPE pipelines still rely on Transformers, whose quadratic complexity makes real-time processing for long sequences impractical. Mamba addresses this issue through selective state-space modeling, enabling efficient sequence processing without sacrificing representational power. Nevertheless, it struggles to capture complex spatial dependencies in multimodal settings. To bridge this gap, we propose VIMCAN, a hybrid architecture that combines the efficient sequence modeling of Mamba with the spatial reasoning of Cross-Attention, and performs robust visual-inertial fusion and human pose estimation between RGB keypoints and wearable IMU data. By leveraging Mamba's dynamic parameterization for temporal modeling and Attention for spatial dependency extraction, VIMCAN achieves superior accuracy, with mean per-joint position errors (MPJPE) of 17.2 mm on TotalCapture and 45.3 mm on 3DPW. VIMCAN outperforms prior Transformer-based and other SOTA approaches while supporting real-time inference at over 60 frames per second on consumer-grade hardware. The source code is available at \href{https://github.com/Eddieyzp/VIMCAN}{this GitHub repository}.

cs.CV↗

arXiv2026

Defect screening and load transfer in minimal hard-soft double networks

Double network (DN) materials exhibit anomalous strength and toughness that far exceed the sum of their constituents. While widely exploited, the fundamental physical mechanisms underlying this synergy remain elusive. Here, we show that a minimal three-dimensional model of two coupled, disordered linear-elastic networks is sufficient to capture the essential physics of DN nonlinear mechanics. The model reproduces the full suite of unique mechanical behaviors, including yielding, necking, strain hardening, and the brittle-to-ductile transition. Mechanical contrast between the hard and soft networks drives inter-network load transfer, which screens defects and suppresses stress concentrations in the hard network. By defining a stress-concentration factor, K_sc, we find that the hard-network failure strain scales universally as 1/K_sc, directly bridging microscopic defect screening to macroscopic yielding. We further show that complete defect screening triggers the shift from localized necking to delocalized damage. Furthermore, the stable necking plateau is identified as an energetic selection governed by the balance between potential energy release and irreversible dissipation. These findings reveal that a simple linear-elastic framework can account for the rich nonlinear landscape of DN materials, providing a general principle for designing next-generation tough solids.

cond-mat.soft↗

arXiv2026

Evolving-RL: End-to-End Optimization of Experience-Driven Self-Evolving Capability within Agents

Experience-driven self-evolving agents aim to overcome the static nature of large language models by distilling reusable experience from past interactions, thus enabling adaptation to novel tasks at deployment time. This process places substantial demands on the foundation model's capacities for abstraction, generalization, and in-context learning. However, most existing studies focus primarily on system-level design choices, such as how experience is represented and managed, neglecting the inherent capabilities of the underlying model. While some recent works have started to optimize the experience utilization stage via reinforcement learning, they still fail to treat self-evolution as a unified process to be jointly optimized. To this end, we propose Evolving-RL, an efficient algorithmic framework that jointly improves the experience extraction and utilization capabilities required for self-evolution. Specifically, we center the learning process on experience extraction and evaluation, using the two supervisory signals derived from evaluation to optimize the extractor and solver separately and thus enable their coordinated co-evolution. Experiments on ALFWorld and Mind2Web show that Evolving-RL effectively enhances LLMs' ability to extract and reuse experience, leading to strong performance gains on out-of-distribution tasks (up to 98.7% relative improvement over the GRPO baseline on ALFWorld unseen tasks and 35.8% on Mind2Web), and these gains are fully unlocked only through the coordinated co-evolution of experience extraction and utilization. Furthermore, Evolving-RL inherently functions as an experience-augmented RL algorithm. By internalizing reusable experience patterns directly into model parameters, it achieves remarkable performance gains over standard baselines on both seen and unseen tasks, even in the absence of test-time experience accumulation.

cs.AI↗

arXiv2026

When Large Vision-Language Models Meet Person Re-Identification

Large Vision-Language Models (LVLMs) that incorporate visual models and large language models have achieved impressive results across cross-modal understanding and reasoning tasks. In recent years, person re-identification (ReID) has also started to explore cross-modal semantics to improve the accuracy of identity recognition. However, effectively utilizing LVLMs for ReID remains an open challenge. While LVLMs operate under a generative paradigm by predicting the next output word, ReID requires the extraction of discriminative identity features to match pedestrians across cameras. In this paper, we propose LVLM-ReID, a novel framework that harnesses the strengths of LVLMs to promote ReID. Specifically, we employ instructions to guide the LVLM in generating one semantic token that encapsulates key appearance semantics from the person image. This token is further refined through our Semantic-Guided Interaction (SGI) module, establishing a reciprocal interaction between the semantic token and visual tokens. Ultimately, the reinforced semantic token serves as the representation of pedestrian identity. Our framework integrates the semantic understanding and generation capabilities of LVLM into end-to-end ReID training, allowing LVLM to capture rich semantic cues during both training and inference. LVLM-ReID achieves competitive results on multiple benchmarks without additional image-text annotations, demonstrating the potential of LVLM-generated semantics to advance person ReID.

cs.CV↗

arXiv2026

KORE: Enhancing Knowledge Injection for Large Multimodal Models via Knowledge-Oriented Controls

Large Multimodal Models encode extensive factual knowledge in their pre-trained weights. However, its knowledge remains static and limited, unable to keep pace with real-world developments, which hinders continuous knowledge acquisition. Effective knowledge injection thus becomes critical, involving two goals: knowledge adaptation (injecting new knowledge) and knowledge retention (preserving old knowledge). Existing methods often struggle to learn new knowledge and suffer from catastrophic forgetting. To address this, we propose KORE, a synergistic method of KnOwledge-oRientEd augmentations and constraints for injecting new knowledge into large multimodal models while preserving old knowledge. Unlike general text or image data augmentation, KORE automatically converts individual knowledge items into structured and comprehensive knowledge to ensure that the model accurately learns new knowledge, enabling accurate adaptation. Meanwhile, KORE stores previous knowledge in the covariance matrix of LMM's linear layer activations and initializes the adapter by projecting the original weights into the matrix's null space, defining a fine-tuning direction that minimizes interference with previous knowledge, enabling powerful retention. Extensive experiments on various LMMs, including LLaVA-v1.5-7B, LLaVA-v1.5-13B, and Qwen2.5-VL-7B, show that KORE achieves superior new knowledge injection performance and effectively mitigates catastrophic forgetting.

cs.CL↗

arXiv2026

Wireless Communication for Low-Altitude Economy with UAV Swarm Enabled Two-Level Movable Antenna System

Unmanned aerial vehicle (UAV) is regarded as a key enabling platform for low-altitude economy, due to its advantages such as 3D maneuverability, flexible deployment, and LoS air-to-air/ground communication links. In particular, the intrinsic high mobility renders UAV especially suitable for operating as a movable antenna (MA) from the sky. In this paper, by exploiting the flexible mobility of UAV swarm and antenna position adjustment of MA, we propose a novel UAV swarm enabled two-level MA system, where UAVs not only individually deploy a local MA array, but also form a larger-scale MA system with their individual MA arrays via swarm coordination. We formulate a general optimization problem to maximize the minimum achievable rate over all ground user equipments (UEs), by jointly optimizing the 3D UAV swarm placement positions, their individual MAs' positions, and receive beamforming for different UEs. To gain useful insights, we first consider the special case where each UAV has only one antenna, under different scenarios of one single UE, two UEs, and arbitrary number of UEs. In particular, for the two-UE case, we derive the optimal UAV swarm placement positions in closed-form that achieves IUI-free communication when the uniform plane wave (UPW) model holds, where the UAV swarm forms a uniform sparse array (USA) satisfying minimum safe distance constraint. While for the general case with arbitrary number of UEs, we propose an efficient alternating optimization algorithm to solve the formulated non-convex optimization problem. Then, we extend the results to the case where each UAV is equipped with multiple antennas. Numerical results verify that the proposed low-altitude UAV swarm enabled MA system significantly outperforms various benchmark schemes, thanks to the exploitation of two-level mobility to create more favorable channel conditions for multi-UE communications.

cs.IT↗

arXiv2026

A Pose-only Geometric Constraint for Multi-Camera Pose Adjustment

Multi-camera systems offer rich observation capabilities for visual navigation and 3D scene reconstruction; however, the resulting feature redundancy often compromises computational efficiency. This challenge is particularly pronounced during bundle adjustment, where the non-linear optimization of both system poses and scene points incurs substantial computational overhead. To address this challenge, this paper introduces a pose-only geometric constraint for multi-camera systems and proposes a corresponding pose adjustment algorithm. Specifically, we use generalized camera model to establish a unified representation of the multi-camera system. Building upon this model, we formulate the multi-camera pose-only constraint, which implicitly represents a 3D scene point using two base observations and their associated poses, thereby achieving a pose-only representation of the projection geometry. Subsequently, we introduce a multi-camera pose adjustment algorithm that eliminates 3D points from the parameter space, thereby achieving efficient and focused pose optimization. Experimental results on both synthetic and real-world datasets demonstrate that the proposed algorithm outperforms baseline bundle adjustment methods in computational efficiency, while maintaining or even improving pose estimation accuracy.

cs.CV↗

arXiv2026

Photometry and physical characterization of near-Earth asteroid 2025 FA22 from one apparition

We present comprehensive photometric characterisation of 2025 FA$_{22}$, a Potentially Hazardous Asteroid (PHA) discovered on 29 March 2025 and observed during the seventh International Asteroid Warning Network (IAWN) coordinated campaign. The asteroid's close approach at 2 lunar distances on 18 September 2025 provided an opportunity for rapid physical characterisation in a simulated virtual impactor scenario. Photometric observations were conducted from 17 September to 1 October 2025, during which 2025 FA$_{22}$ traversed a $150^{\circ}$ arc and spanned solar phase angles from $\sim20^{\circ}$ to $\sim70^{\circ}$. This geometry enabled detailed physical characterisation, including determination of the spin axis and shape. Convex inversion yields a sidereal rotation period of $P_{\mathrm{sid}} = 13.07366 \pm 0.00076$ h and a spin axis at ecliptic coordinates $(λ, β) = (246^{\circ} \pm 9^{\circ}, 60^{\circ} \pm 9^{\circ})$. The absolute magnitude was derived as $H_{\mathrm{V}} = 21.39^{+0.07}_{-0.08}$ mag ($G_1 = 0.8228$, $G_2 = 0.0194$), with colour indices $B-V = 0.71 \pm 0.05$ mag, $V-R = 0.39 \pm 0.03$ mag, and $R-I = 0.39 \pm 0.04$ mag, consistent with X-complex classification in the Bus--DeMeo taxonomy. Assuming a geometric albedo of $p_v = 0.15^{+0.05}_{-0.04}$, representative of the moderate-albedo X-complex asteroids, we estimate $D_{\mathrm{eff}} = 181^{+31}_{-25}$ m, consistent with "China Compound Eye" radar dimensions of ${\sim}100 \times 320$ m ($D_{\mathrm{eff}} \sim 186$ m), which also reveal a contact-binary morphology consistent with the axis ratios $a:b:c \sim 2.68:1.96:1.00$ derived from our photometric shape model. The bilobate morphology is consistent with YORP-driven spin-up and deformation during a previous YORP cycle, with the current slow rotation possibly explained by internal reconfiguration or tidal braking during close Earth encounters.

astro-ph.EP↗

arXiv2026

Chain Of Interaction Benchmark (COIN): When Reasoning meets Embodied Interaction

Generalist embodied agents must perform interactive, causally-dependent reasoning, continually interacting with the environment, acquiring information, and updating plans to solve long-horizon tasks before they could be adopted in real-life scenarios. For instance, retrieving an apple from a cabinet may require opening multiple doors and drawers before the apple becomes visible and reachable, demanding sequential interaction under partial observability. However, existing benchmarks fail to systematically evaluate this essential capability. We introduce COIN, a benchmark designed to assess interactive reasoning in realistic robotic manipulation through three key contributions. First, we construct COIN-50: 50 interactive tasks in daily scenarios, and create COIN-Primitive required by causally-dependent tasks, and COIN-Composition with mid-term complexity for skill learning and generalization evaluation. Second, we develop a low-cost mobile AR teleoperation system and collect the COIN-Primitive Dataset with 50 demonstrations per primitive task (1,000 in total). Third, we develop systematic evaluation metrics about execution stability and generalization robustness to evaluate CodeAsPolicy, VLA, and language-conditioned H-VLA approaches. Our comprehensive evaluation reveals critical limitations in current methods: models struggle with interactive reasoning tasks due to significant gaps between visual understanding and motor execution. We provide fine-grained analysis of these limitations.

cs.RO↗

arXiv2026

Optimal Insurance Menu Design under the Expected-Value Premium Principle

This paper studies optimal insurance design under asymmetric information in a Stackelberg framework, where a monopolistic insurer faces uncertainty about both the insured's risk attitude, captured by a risk-aversion parameter, and the insured's risk type, characterized by the loss distribution. In particular, when the risk type is unobservable, we allow the risk-aversion parameter to depend on the risk type. We construct a menu of contracts that maximizes the mean-variance utilities of both parties under the expected-value premium principle, subject to a truth-telling constraint that ensures the truthful revelation of private information. We show that when risk attitude is private information, the optimal coverage takes the form of excess-of-loss insurance with linear pricing in terms of the risk loading (defined as the premium minus the expected loss), designed to screen risk preferences. In contrast, when risk type is unobserved, we restrict the coverage function to an excess-of-loss form and derive an ordinary differential equation that characterizes the optimal risk loading. Under mild conditions, we establish the existence and uniqueness of the solution. The results show that equilibrium contracts exhibit nonlinear pricing with decreasing risk loadings, implying that higher-risk individuals face lower risk loadings in order to induce self-selection. Finally, numerical illustrations demonstrate how parameter values and the distributions of unobserved heterogeneity affect the structure of optimal contracts and the resulting pricing schedule.

q-fin.RM↗

arXiv2026

Simplicity Prevails: The Emergence of Generalizable AIGI Detection in Visual Foundation Models

While specialized detectors for AI-Generated Images (AIGI) achieve near-perfect accuracy on curated benchmarks, they suffer from a dramatic performance collapse in realistic, in-the-wild scenarios. In this work, we demonstrate that simplicity prevails over complex architectural designs. A simple linear classifier trained on the frozen features of modern Vision Foundation Models , including Perception Encoder, MetaCLIP 2, and DINOv3, establishes a new state-of-the-art. Through a comprehensive evaluation spanning traditional benchmarks, unseen generators, and challenging in-the-wild distributions, we show that this baseline not only matches specialized detectors on standard benchmarks but also decisively outperforms them on in-the-wild datasets, boosting accuracy by striking margins of over 30\%. We posit that this superior capability is an emergent property driven by the massive scale of pre-training data containing synthetic content. We trace the source of this capability to two distinct manifestations of data exposure: Vision-Language Models internalize an explicit semantic concept of forgery, while Self-Supervised Learning models implicitly acquire discriminative forensic features from the pretraining data. However, we also reveal persistent limitations: these models suffer from performance degradation under recapture and transmission, remain blind to VAE reconstruction and localized editing. We conclude by advocating for a paradigm shift in AI forensics, moving from overfitting on static benchmarks to harnessing the evolving world knowledge of foundation models for real-world reliability.

cs.CV↗

arXiv2026

CoD-Lite: Real-Time Diffusion-Based Generative Image Compression

Recent advanced diffusion methods typically derive strong generative priors by scaling diffusion transformers. However, scaling fails to generalize when adapted for real-time compression scenarios that demand lightweight models. In this paper, we explore the design of real-time and lightweight diffusion codecs by addressing two pivotal questions. First, does diffusion pre-training benefit lightweight diffusion codecs? Through systematic analysis, we find that generation-oriented pre-training is less effective at small model scales whereas compression-oriented pre-training yields consistently better performance. Second, are transformers essential? We find that while global attention is crucial for standard generation, lightweight convolutions suffice for compression-oriented diffusion when paired with distillation. Guided by these findings, we establish a one-step lightweight convolution diffusion codec that achieves real-time $60$~FPS encoding and $42$~FPS decoding at 1080p. Further enhanced by distillation and adversarial learning, the proposed codec reduces bitrate by 85\% at a comparable FID to MS-ILLM, bridging the gap between generative compression and practical real-time deployment. Codes are released at https://github.com/microsoft/GenCodec/tree/main/CoD_Lite

cs.CV↗

arXiv2026

DyBBT: Dynamic Balance via Bandit-inspired Targeting for Dialog Policy with Cognitive Dual-Systems

Task oriented dialog systems often rely on static exploration strategies that do not adapt to dynamic dialog contexts, leading to inefficient exploration and suboptimal performance. We propose DyBBT, a novel dialog policy learning framework that formalizes the exploration challenge through a structured cognitive state space capturing dialog progression, user uncertainty, and slot dependency. DyBBT proposes a bandit inspired meta-controller that dynamically switches between a fast intuitive inference (System 1) and a slow deliberative reasoner (System 2) based on real-time cognitive states and visitation counts. Extensive experiments on single- and multi-domain benchmarks show that DyBBT achieves state-of-the-art performance in success rate, efficiency, and generalization, with human evaluations confirming its decisions are well aligned with expert judgment.

cs.CL↗

arXiv2026

HiCoLoRA: Addressing Context-Prompt Misalignment via Hierarchical Collaborative LoRA for Zero-Shot DST

Zero-shot Dialog State Tracking (zs-DST) is essential for enabling Task-Oriented Dialog Systems (TODs) to generalize to new domains without costly data annotation. A central challenge lies in the semantic misalignment between dynamic dialog contexts and static prompts, leading to inflexible cross-layer coordination, domain interference, and catastrophic forgetting. To tackle this, we propose Hierarchical Collaborative Low-Rank Adaptation (HiCoLoRA), a framework that enhances zero-shot slot inference through robust prompt alignment. It features a hierarchical LoRA architecture for dynamic layer-specific processing (combining lower-layer heuristic grouping and higher-layer full interaction), integrates Spectral Joint Domain-Slot Clustering to identify transferable associations (feeding an Adaptive Linear Fusion Mechanism), and employs Semantic-Enhanced SVD Initialization (SemSVD-Init) to preserve pre-trained knowledge. Experiments on multi-domain datasets MultiWOZ and SGD show that HiCoLoRA outperforms baselines, achieving SOTA in zs-DST. Code is available at https://github.com/carsonz/HiCoLoRA.

cs.CL↗

arXiv2026

DarwinTOD: LLM-driven Lifelong Self-evolution for Task-oriented Dialog Systems

Traditional task-oriented dialog systems are unable to evolve from ongoing interactions or adapt to new domains after deployment, that is a critical limitation in real-world dynamic environments. Continual learning approaches depend on episodic retraining with human curated data, failing to achieve autonomy lifelong improvement. While evolutionary computation and LLM driven self improvement offer promising mechanisms for dialog optimization, they lack a unified framework for holistic, iterative strategy refinement. To bridge this gap, we propose DarwinTOD, a lifelong self evolving dialog framework that systematically integrates these two paradigms, enabling continuous strategy optimization from a zero-shot base without task specific fine-tuning. DarwinTOD maintains an Evolvable Strategy Bank and operates through a dual-loop process: online multi-agent dialog execution with peer critique, and offline structured evolutionary operations that refine the strategy bank using accumulated feedback. This closed-loop design enables autonomous continuous improvement without human intervention. Extensive experiments show that DarwinTOD surpasses previous state-of-the-art methods and exhibits continuous performance gains throughout evolution. Our work provides a novel framework for building dialog systems with lifelong self evolution capabilities.

cs.MA↗

arXiv2026

MINED: Probing and Updating with Multimodal Time-Sensitive Knowledge for Large Multimodal Models

Large Multimodal Models (LMMs) encode rich factual knowledge via cross-modal pre-training, yet their static representations struggle to maintain an accurate understanding of time-sensitive factual knowledge. Existing benchmarks remain constrained by static designs, inadequately evaluating LMMs' ability to understand time-sensitive knowledge. To address this gap, we propose MINED, a comprehensive benchmark that evaluates temporal awareness along 6 key dimensions and 11 challenging tasks: cognition, awareness, trustworthiness, understanding, reasoning, and robustness. MINED is constructed from Wikipedia by two professional annotators, containing 2,104 time-sensitive knowledge samples spanning six knowledge types. Evaluating 15 widely used LMMs on MINED shows that Gemini-2.5-Pro achieves the highest average CEM score of 63.07, while most open-source LMMs still lack time understanding ability. Meanwhile, LMMs perform best on organization knowledge, whereas their performance is weakest on sport. To address these challenges, we investigate the feasibility of updating time-sensitive knowledge in LMMs through knowledge editing methods and observe that LMMs can effectively update knowledge via knowledge editing methods in single editing scenarios.

cs.CL↗

arXiv2026

Event-Based Method for High-Speed 3D Deformation Measurement under Extreme Illumination Conditions

Background: Large engineering structures, such as space launch towers and suspension bridges, are subjected to extreme forces that cause high-speed 3D deformation and compromise safety. These structures typically operate under extreme illumination conditions. Traditional cameras often struggle to handle strong light intensity, leading to overexposure due to their limited dynamic range. Objective: Event cameras have emerged as a compelling alternative to traditional cameras in high dynamic range and low-latency applications. This paper presents an integrated method, from calibration to measurement, using a multi-event camera array for high-speed 3D deformation monitoring of structures in extreme illumination conditions. Methods: Firstly, the proposed method combines the characteristics of the asynchronous event stream and temporal correlation analysis to extract the corresponding marker center point. Subsequently, the method achieves rapid calibration by solving the Kruppa equations in conjunction with a parameter optimization framework. Finally, by employing a unified coordinate transformation and linear intersection, the method enables the measurement of 3D deformation of the target structure. Results: Experiments confirmed that the relative measurement error is below 0.08%. Field experiments under extreme illumination conditions, including self-calibration of a multi-event camera array and 3D deformation measurement, verified the performance of the proposed method. Conclusions: This paper addressed the critical limitation of traditional cameras in measuring high-speed 3D deformations under extreme illumination conditions. The experimental results demonstrate that, compared to other methods, the proposed method can accurately measure 3D deformations of structures under harsh lighting conditions, and the relative error of the measured deformation is less than 0.1%.

cs.CV↗