arXiv subjects

Han Zhou

Publications and source records attributed to Han Zhou.

At least 73 records · Page 4Linked to original sources

arXiv2026

STEP3-VL-10B Technical Report

We present STEP3-VL-10B, a lightweight open-source foundation model designed to redefine the trade-off between compact efficiency and frontier-level multimodal intelligence. STEP3-VL-10B is realized through two strategic shifts: first, a unified, fully unfrozen pre-training strategy on 1.2T multimodal tokens that integrates a language-aligned Perception Encoder with a Qwen3-8B decoder to establish intrinsic vision-language synergy; and second, a scaled post-training pipeline featuring over 1k iterations of reinforcement learning. Crucially, we implement Parallel Coordinated Reasoning (PaCoRe) to scale test-time compute, allocating resources to scalable perceptual reasoning that explores and synthesizes diverse visual hypotheses. Consequently, despite its compact 10B footprint, STEP3-VL-10B rivals or surpasses models 10$\times$-20$\times$ larger (e.g., GLM-4.6V-106B, Qwen3-VL-235B) and top-tier proprietary flagships like Gemini 2.5 Pro and Seed-1.5-VL. Delivering best-in-class performance, it records 92.2% on MMBench and 80.11% on MMMU, while excelling in complex reasoning with 94.43% on AIME2025 and 75.95% on MathVision. We release the full model suite to provide the community with a powerful, efficient, and reproducible baseline.

cs.CV↗

arXiv2025

A Reinforcement Learning Method to Factual and Counterfactual Explanations for Session-based Recommendation

Session-based Recommendation (SR) systems have recently achieved considerable success, yet their complex, "black box" nature often obscures why certain recommendations are made. Existing explanation methods struggle to pinpoint truly influential factors, as they frequently depend on static user profiles or fail to grasp the intricate dynamics within user sessions. In response, we introduce FCESR (Factual and Counterfactual Explanations for Session-based Recommendation), a novel framework designed to illuminate SR model predictions by emphasizing both the sufficiency (factual) and necessity (counterfactual) of recommended items. By recasting explanation generation as a combinatorial optimization challenge and leveraging reinforcement learning, our method uncovers the minimal yet critical sequence of items influencing recommendations. Moreover, recognizing the intrinsic value of robust explanations, we innovatively utilize these factual and counterfactual insights within a contrastive learning paradigm, employing them as high-quality positive and negative samples to fine-tune and significantly enhance SR accuracy. Extensive qualitative and quantitative evaluations across diverse datasets and multiple SR architectures confirm that our framework not only boosts recommendation accuracy but also markedly elevates the quality and interpretability of explanations, thereby paving the way for more transparent and trustworthy recommendation systems.

cs.SI↗

arXiv2025

Skypilot: Fine-Tuning LLM with Physical Grounding for AAV Coverage Search

Autonomous aerial vehicles (AAVs) have played a pivotal role in coverage operations and search missions. Recent advances in large language models (LLMs) offer promising opportunities to augment AAV intelligence. These advances help address complex challenges like area coverage optimization, dynamic path planning, and adaptive decision-making. However, the absence of physical grounding in LLMs leads to hallucination and reproducibility problems in spatial reasoning and decision-making. To tackle these issues, we present Skypilot, an LLM-enhanced two-stage framework that grounds language models in physical reality by integrating monte carlo tree search (MCTS). In the first stage, we introduce a diversified action space that encompasses generate, regenerate, fine-tune, and evaluate operations, coupled with physics-informed reward functions to ensure trajectory feasibility. In the second stage, we fine-tune Qwen3-4B on 23,000 MCTS-generated samples, achieving substantial inference acceleration while maintaining solution quality. Extensive numerical simulations and real-world flight experiments validate the efficiency and superiority of our proposed approach. Detailed information and experimental results are accessible at https://sky-pilot.top.

cs.RO↗

arXiv2025

Zero-Reference Joint Low-Light Enhancement and Deblurring via Visual Autoregressive Modeling with VLM-Derived Modulation

Real-world dark images commonly exhibit not only low visibility and contrast but also complex noise and blur, posing significant restoration challenges. Existing methods often rely on paired data or fail to model dynamic illumination and blur characteristics, leading to poor generalization. To tackle this, we propose a generative framework based on visual autoregressive (VAR) modeling, guided by perceptual priors from the vision-language model (VLM). Specifically, to supply informative conditioning cues for VAR models, we deploy an adaptive curve estimation scheme to modulate the diverse illumination based on VLM-derived visibility scores. In addition, we integrate dynamic and spatial-frequency-aware Rotary Positional Encodings (SF-RoPE) into VAR to enhance its ability to model structures degraded by blur. Furthermore, we propose a recursive phase-domain modulation strategy that mitigates blur-induced artifacts in the phase domain via bounded iterative refinement guided by VLM-assessed blur scores. Our framework is fully unsupervised and achieves state-of-the-art performance on benchmark datasets.

cs.CV↗

arXiv2025

NTIRE 2025 Challenge on Low Light Image Enhancement: Methods and Results

This paper presents a comprehensive review of the NTIRE 2025 Low-Light Image Enhancement (LLIE) Challenge, highlighting the proposed solutions and final outcomes. The objective of the challenge is to identify effective networks capable of producing brighter, clearer, and visually compelling images under diverse and challenging conditions. A remarkable total of 762 participants registered for the competition, with 28 teams ultimately submitting valid entries. This paper thoroughly evaluates the state-of-the-art advancements in LLIE, showcasing the significant progress.

cs.CV↗

arXiv2025

Adaptive Modality Balanced Online Knowledge Distillation for Brain-Eye-Computer based Dim Object Detection

Advanced cognition can be extracted from the human brain using brain-computer interfaces. Integrating these interfaces with computer vision techniques, which possess efficient feature extraction capabilities, can achieve more robust and accurate detection of dim targets in aerial images. However, existing target detection methods primarily concentrate on homogeneous data, lacking efficient and versatile processing capabilities for heterogeneous multimodal data. In this paper, we first build a brain-eye-computer based object detection system for aerial images under few-shot conditions. This system detects suspicious targets using region proposal networks, evokes the event-related potential (ERP) signal in electroencephalogram (EEG) through the eye-tracking-based slow serial visual presentation (ESSVP) paradigm, and constructs the EEG-image data pairs with eye movement data. Then, an adaptive modality balanced online knowledge distillation (AMBOKD) method is proposed to recognize dim objects with the EEG-image data. AMBOKD fuses EEG and image features using a multi-head attention module, establishing a new modality with comprehensive features. To enhance the performance and robust capability of the fusion modality, simultaneous training and mutual learning between modalities are enabled by end-to-end online knowledge distillation. During the learning process, an adaptive modality balancing module is proposed to ensure multimodal equilibrium by dynamically adjusting the weights of the importance and the training gradients across various modalities. The effectiveness and superiority of our method are demonstrated by comparing it with existing state-of-the-art methods. Additionally, experiments conducted on public datasets and system validations in real-world scenarios demonstrate the reliability and practicality of the proposed system and the designed method.

cs.CV↗

arXiv2025

Maestro: Self-Improving Text-to-Image Generation via Agent Orchestration

Text-to-image (T2I) models, while offering immense creative potential, are highly reliant on human intervention, posing significant usability challenges that often necessitate manual, iterative prompt engineering over often underspecified prompts. This paper introduces Maestro, a novel self-evolving image generation system that enables T2I models to autonomously self-improve generated images through iterative evolution of prompts, using only an initial prompt. Maestro incorporates two key innovations: 1) self-critique, where specialized multimodal LLM (MLLM) agents act as 'critics' to identify weaknesses in generated images, correct for under-specification, and provide interpretable edit signals, which are then integrated by a 'verifier' agent while preserving user intent; and 2) self-evolution, utilizing MLLM-as-a-judge for head-to-head comparisons between iteratively generated images, eschewing problematic images, and evolving creative prompt candidates that align with user intents. Extensive experiments on complex T2I tasks using black-box models demonstrate that Maestro significantly improves image quality over initial prompts and state-of-the-art automated methods, with effectiveness scaling with more advanced MLLM components. This work presents a robust, interpretable, and effective pathway towards self-improving T2I generation.

cs.AI↗

arXiv2025

A Novel Characterization of the Population Area Under the Risk Coverage Curve (AURC) and Rates of Finite Sample Estimators

The selective classifier (SC) has been proposed for rank based uncertainty thresholding, which could have applications in safety critical areas such as medical diagnostics, autonomous driving, and the justice system. The Area Under the Risk-Coverage Curve (AURC) has emerged as the foremost evaluation metric for assessing the performance of SC systems. In this work, we present a formal statistical formulation of population AURC, presenting an equivalent expression that can be interpreted as a reweighted risk function. Through Monte Carlo methods, we derive empirical AURC plug-in estimators for finite sample scenarios. The weight estimators associated with these plug-in estimators are shown to be consistent, with low bias and tightly bounded mean squared error (MSE). The plug-in estimators are proven to converge at a rate of $\mathcal{O}(\sqrt{\ln(n)/n})$ demonstrating statistical consistency. We empirically validate the effectiveness of our estimators through experiments across multiple datasets, model architectures, and confidence score functions (CSFs), demonstrating consistency and effectiveness in fine-tuning AURC performance.

stat.ML↗

arXiv2025

A Comprehensive Survey of Self-Evolving AI Agents: A New Paradigm Bridging Foundation Models and Lifelong Agentic Systems

Recent advances in large language models have sparked growing interest in AI agents capable of solving complex, real-world tasks. However, most existing agent systems rely on manually crafted configurations that remain static after deployment, limiting their ability to adapt to dynamic and evolving environments. To this end, recent research has explored agent evolution techniques that aim to automatically enhance agent systems based on interaction data and environmental feedback. This emerging direction lays the foundation for self-evolving AI agents, which bridge the static capabilities of foundation models with the continuous adaptability required by lifelong agentic systems. In this survey, we provide a comprehensive review of existing techniques for self-evolving agentic systems. Specifically, we first introduce a unified conceptual framework that abstracts the feedback loop underlying the design of self-evolving agentic systems. The framework highlights four key components: System Inputs, Agent System, Environment, and Optimisers, serving as a foundation for understanding and comparing different strategies. Based on this framework, we systematically review a wide range of self-evolving techniques that target different components of the agent system. We also investigate domain-specific evolution strategies developed for specialised fields such as biomedicine, programming, and finance, where optimisation objectives are tightly coupled with domain constraints. In addition, we provide a dedicated discussion on the evaluation, safety, and ethical considerations for self-evolving agentic systems, which are critical to ensuring their effectiveness and reliability. This survey aims to provide researchers and practitioners with a systematic understanding of self-evolving AI agents, laying the foundation for the development of more adaptive, autonomous, and lifelong agentic systems.

cs.AI↗

arXiv2025

ArgusCogito: Chain-of-Thought for Cross-Modal Synergy and Omnidirectional Reasoning in Camouflaged Object Segmentation

Camouflaged Object Segmentation (COS) poses a significant challenge due to the intrinsic high similarity between targets and backgrounds, demanding models capable of profound holistic understanding beyond superficial cues. Prevailing methods, often limited by shallow feature representation, inadequate reasoning mechanisms, and weak cross-modal integration, struggle to achieve this depth of cognition, resulting in prevalent issues like incomplete target separation and imprecise segmentation. Inspired by the perceptual strategy of the Hundred-eyed Giant-emphasizing holistic observation, omnidirectional focus, and intensive scrutiny-we introduce ArgusCogito, a novel zero-shot, chain-of-thought framework underpinned by cross-modal synergy and omnidirectional reasoning within Vision-Language Models (VLMs). ArgusCogito orchestrates three cognitively-inspired stages: (1) Conjecture: Constructs a strong cognitive prior through global reasoning with cross-modal fusion (RGB, depth, semantic maps), enabling holistic scene understanding and enhanced target-background disambiguation. (2) Focus: Performs omnidirectional, attention-driven scanning and focused reasoning, guided by semantic priors from Conjecture, enabling precise target localization and region-of-interest refinement. (3) Sculpting: Progressively sculpts high-fidelity segmentation masks by integrating cross-modal information and iteratively generating dense positive/negative point prompts within focused regions, emulating Argus' intensive scrutiny. Extensive evaluations on four challenging COS benchmarks and three Medical Image Segmentation (MIS) benchmarks demonstrate that ArgusCogito achieves state-of-the-art (SOTA) performance, validating the framework's exceptional efficacy, superior generalization capability, and robustness.

cs.CV↗

arXiv2025

A kernel-free boundary integral method for elliptic interface problems on surfaces

This work presents a generalized boundary integral method for elliptic equations on surfaces, encompassing both boundary value and interface problems. The method is kernel-free, implying that the explicit analytical expression of the kernel function is not required when solving the boundary integral equations. The numerical integration of single- and double-layer potentials or volume integrals at the boundary is replaced by interpolation of the solution to an equivalent interface problem, which is then solved using a fast multigrid solver on Cartesian grids. This paper provides detailed implementation of the second-order version of the kernel-free boundary integral method for elliptic PDEs defined on an embedding surface in $\mathbb{R}^3$ and presents numerical experiments to demonstrate the efficiency and accuracy of the method for both boundary value and interface problems.

math.NA↗

arXiv2025

Harmonized Gradient Descent for Class Imbalanced Data Stream Online Learning

Many real-world data are sequentially collected over time and often exhibit skewed class distributions, resulting in imbalanced data streams. While existing approaches have explored several strategies, such as resampling and reweighting, for imbalanced data stream learning, our work distinguishes itself by addressing the imbalance problem through training modification, particularly focusing on gradient descent techniques. We introduce the harmonized gradient descent (HGD) algorithm, which aims to equalize the norms of gradients across different classes. By ensuring the gradient norm balance, HGD mitigates under-fitting for minor classes and achieves balanced online learning. Notably, HGD operates in a streamlined implementation process, requiring no data-buffer, extra parameters, or prior knowledge, making it applicable to any learning models utilizing gradient descent for optimization. Theoretical analysis, based on a few common and mild assumptions, shows that HGD achieves a satisfied sub-linear regret bound. The proposed algorithm are compared with the commonly used online imbalance learning methods under several imbalanced data stream scenarios. Extensive experimental evaluations demonstrate the efficiency and effectiveness of HGD in learning imbalanced data streams.

cs.LG↗

arXiv2025

AU-IQA: A Benchmark Dataset for Perceptual Quality Assessment of AI-Enhanced User-Generated Content

AI-based image enhancement techniques have been widely adopted in various visual applications, significantly improving the perceptual quality of user-generated content (UGC). However, the lack of specialized quality assessment models has become a significant limiting factor in this field, limiting user experience and hindering the advancement of enhancement methods. While perceptual quality assessment methods have shown strong performance on UGC and AIGC individually, their effectiveness on AI-enhanced UGC (AI-UGC) which blends features from both, remains largely unexplored. To address this gap, we construct AU-IQA, a benchmark dataset comprising 4,800 AI-UGC images produced by three representative enhancement types which include super-resolution, low-light enhancement, and denoising. On this dataset, we further evaluate a range of existing quality assessment models, including traditional IQA methods and large multimodal models. Finally, we provide a comprehensive analysis of how well current approaches perform in assessing the perceptual quality of AI-UGC. The access link to the AU-IQA is https://github.com/WNNGGU/AU-IQA-Dataset.

cs.CV↗

arXiv2025

Step-3 is Large yet Affordable: Model-system Co-design for Cost-effective Decoding

Large language models (LLMs) face low hardware efficiency during decoding, especially for long-context reasoning tasks. This paper introduces Step-3, a 321B-parameter VLM with hardware-aware model-system co-design optimized for minimizing decoding costs. Step-3 innovates in two key dimensions: (1) A novel Multi-Matrix Factorization Attention (MFA) mechanism that significantly reduces both KV cache size and computation while maintaining high attention expressiveness, and (2) Attention-FFN Disaggregation (AFD), a distributed inference system that decouples attention and Feed-Forward Network (FFN) layers into specialized subsystems. This co-design achieves unprecedented cost efficiency: Step-3 significantly reduces theoretical decoding costs compared with models like DeepSeek-V3 and Qwen3 MoE 235B, with the gains widening at longer context. Step-3 achieves low cost while activating 38B parameters per token (more than DeepSeek-V3 and Qwen3 MoE 235B), demonstrating that hardware-aligned attention arithmetic intensity, MoE sparsity, and AFD are critical to cost-effectiveness. We perform a head-to-head comparison with DeepSeek-V3 in its favorable scenarios. Our implementation on Hopper GPUs achieves a decoding throughput of up to 4,039 tokens per second per GPU under 50ms TPOT SLA (4K context, FP8, no MTP). It is higher than DeepSeek-V3's 2,324 in the same setup and sets a new Pareto frontier for LLM decoding.

cs.LG↗

arXiv2025

Bayesian Optimization over Bounded Domains with the Beta Product Kernel

Bayesian optimization with Gaussian processes (GP) is commonly used to optimize black-box functions. The Matérn and the Radial Basis Function (RBF) covariance functions are used frequently, but they do not make any assumptions about the domain of the function, which may limit their applicability in bounded domains. To address the limitation, we introduce the Beta kernel, a non-stationary kernel induced by a product of Beta distribution density functions. Such a formulation allows our kernel to naturally model functions on bounded domains. We present statistical evidence supporting the hypothesis that the kernel exhibits an exponential eigendecay rate, based on empirical analyses of its spectral properties across different settings. Our experimental results demonstrate the robustness of the Beta kernel in modeling functions with optima located near the faces or vertices of the unit hypercube. The experiments show that our kernel consistently outperforms a wide range of kernels, including the well-known Matérn and RBF, in different problems, including synthetic function optimization and the compression of vision and language models.

cs.LG↗

arXiv2025

NTIRE 2025 Image Shadow Removal Challenge Report

This work examines the findings of the NTIRE 2025 Shadow Removal Challenge. A total of 306 participants have registered, with 17 teams successfully submitting their solutions during the final evaluation phase. Following the last two editions, this challenge had two evaluation tracks: one focusing on reconstruction fidelity and the other on visual perception through a user study. Both tracks were evaluated with images from the WSRD+ dataset, simulating interactions between self- and cast-shadows with a large number of diverse objects, textures, and materials.

cs.CV↗

arXiv2025

NeSyPack: A Neuro-Symbolic Framework for Bimanual Logistics Packing

This paper presents NeSyPack, a neuro-symbolic framework for bimanual logistics packing. NeSyPack combines data-driven models and symbolic reasoning to build an explainable hierarchical system that is generalizable, data-efficient, and reliable. It decomposes a task into subtasks via hierarchical reasoning, and further into atomic skills managed by a symbolic skill graph. The graph selects skill parameters, robot configurations, and task-specific control strategies for execution. This modular design enables robustness, adaptability, and efficient reuse - outperforming end-to-end models that require large-scale retraining. Using NeSyPack, our team won the First Prize in the What Bimanuals Can Do (WBCD) competition at the 2025 IEEE International Conference on Robotics and Automation.

cs.RO↗

arXiv2025

MANBench: Is Your Multimodal Model Smarter than Human?

The rapid advancement of Multimodal Large Language Models (MLLMs) has ignited discussions regarding their potential to surpass human performance in multimodal tasks. In response, we introduce MANBench (Multimodal Ability Norms Benchmark), a bilingual benchmark (English and Chinese) comprising 1,314 questions across nine tasks, spanning knowledge-based and non-knowledge-based domains. MANBench emphasizes intuitive reasoning, seamless cross-modal integration, and real-world complexity, providing a rigorous evaluation framework. Through extensive human experiments involving diverse participants, we compared human performance against state-of-the-art MLLMs. The results indicate that while MLLMs excel in tasks like Knowledge and Text-Image Understanding, they struggle with deeper cross-modal reasoning tasks such as Transmorphic Understanding, Image Consistency, and Multi-image Understanding. Moreover, both humans and MLLMs face challenges in highly complex tasks like Puzzles and Spatial Imagination. MANBench highlights the strengths and limitations of MLLMs, revealing that even advanced models fall short of achieving human-level performance across many domains. We hope MANBench will inspire efforts to bridge the gap between MLLMs and human multimodal capabilities. The code and dataset are available at https://github.com/micdz/MANBench.

cs.CL↗