arXiv subjects

Haoyu Chen

Publications and source records attributed to Haoyu Chen.

At least 37 records · Page 2Linked to original sources

arXiv2026

LookME: Lookup-Based Multimodal Embeddings for Layer Injection in Vision-Language Models

Vision-Language Models (VLMs) have achieved strong progress in multimodal understanding. However, scaling dense or sparse Mixture-of-Experts (MoE) models to improve performance limits deployment in resource-constrained environments due to the trade-off between high memory usage from full loading and increased latency from on-demand loading. Recently, the Per-Layer Embedding (PLE) architecture addresses this by scaling models with large external embedding tables stored in read-only memory (ROM) and performing lightweight lookup to retrieve relevant embeddings to enhance token representations. Nevertheless, existing PLE-style methods are primarily designed for text embeddings due to the convenience of ID-based retrieval, limiting their effectiveness in VLMs where multimodal embeddings contain richer information for visual tasks. In this paper, we propose LookME, the first framework that enables lookup-based enhancement for multimodal embeddings in VLMs while supporting partitioned storage and on-demand loading. To efficiently lookup arbitrary continuous multimodal embeddings from large-scale embedding tables, we propose a hierarchical two-level lookup method employing a coarse-to-fine strategy that performs lookups from the scene-level to the intra-scene primitive-level. Furthermore, we integrate the lookup method with a sparse injection strategy, which adaptively prioritizes critical embeddings over voluminous multimodal embeddings within layers, and facilitates embedding table reuse across neighboring layers, improving the trade-off among efficiency, model size, and performance. Experiments on multiple visual benchmarks show that LookME outperforms text-only PLE-style methods, validating the effectiveness of lookup-based multimodal embedding enhancement.

cs.CV↗

arXiv2026

Compiling Large Multi-Modal Requirement Documents into Runnable Software Systems: From an Agentic Test-Driven Perspective

Large Language Models (LLMs) have significantly improved programming efficiency by translating natural language into code, yet their performance deteriorates when handling large-scale, multi-modal requirement documents containing hundreds of scenarios, often producing incorrect implementations or missing critical constraints. To address this challenge, we propose ARC (Agentic Requirement Compilation), a framework that compiles DSL-based requirement documents into runnable software systems while automatically generating modular software architecture, comprehensive test suites, and traceability across requirements, design, and code. ARC adopts a bidirectional test-driven agentic workflow, combining a top-down architecture design phase with a bottom-up implementation phase to ensure that generated code satisfies synthesized tests. We evaluate ARC on six runnable web system benchmarks and the AppForge benchmark of 101 mobile app generation tasks. Across three independent trials, ARC consistently outperforms state-of-the-art LLM-based baselines, achieving 50.6% more GUI tests passed on average for web systems, a 100% compilation success rate, and a 68.3% test pass rate on AppForge. A user study with 21 participants further shows that users with limited programming experience can write DSL-based requirement documents containing up to 174 scenarios within an average of 5.6 hours to generate maintainable runnable systems, including a real-world ticket-booking application of approximately 10K lines of code.

cs.SE↗

arXiv2026

MAC 2026: Advancing Micro-Action Analysis Towards Fine-Grained Understanding

Micro-Actions (MAs) are subtle and spontaneous human behaviors that provide important non-verbal cues in social interaction and affective communication. However, their short duration, weak motion patterns, and fine-grained semantic differences make them difficult to annotate, model, and evaluate in a standardized manner. To promote academic research on micro-action analysis, we proposed and have annually organized the Micro-Action Analysis Grand Challenge (MAC) as a public benchmark platform for this emerging field. The first two editions of MAC established standardized evaluation settings for micro-action recognition and detection, providing publicly accessible datasets and protocols. Building upon these editions, this paper presents the 3rd MAC, held in conjunction with ACM Multimedia 2026. Under the theme of moving from recognition to fine-grained micro-action understanding, this edition further expands the scope of the challenge beyond conventional recognition and detection. In particular, we introduce a new task named fine-grained micro-action understanding, evaluated with the assistance of multimodal large language models, aiming to assess models' ability to capture fine-grained semantic cues and interpret subtle human micro-actions at a deeper level. We summarize the datasets, task settings, evaluation protocols, competition results, and representative solutions from top-performing teams. Finally, we discuss future directions for micro-action analysis and its broader role in human-centric video understanding.

cs.CV↗

arXiv2026

PAUSE: A User-Centric Benchmark for Personal AI Assistants in Unified Service Environments

Personal AI assistants are increasingly deployed as task-oriented, tool-augmented agents that operate within unified service environments to support everyday user activities. In realistic settings, such assistants must reason over persistent user state, respect user-specific configurations and permissions, and sustain long-horizon, constraint-aware interactions across multiple services. Existing benchmarks, however, often fragment service contexts or abstract away user state, limiting their ability to evaluate user-centric personal assistant behavior in realistic service settings. We introduce PAUSE, a user-centric benchmark for evaluating personal AI assistants in stateful, service-integrated environments. PAUSE captures core challenges of real-world assistant deployment by requiring agents to coordinate actions across heterogeneous user-owned resources while maintaining consistency with environment state, authorization constraints over multi-turn interactions. The benchmark incorporates explicit user-agent interaction via realistic user simulation, enabling evaluation beyond static tool execution. To support principled and reproducible evaluation, PAUSE adopts a multi-regime evaluation framework aligned with task characteristics. Open-ended service management tasks are assessed using semantic and trajectory-level behavioral metrics, while constraint-intensive tasks admit deterministic, state-based verification. Benchmark results show that even state-of-the-art proprietary models fail to reach 70% task completion on scenarios requiring stateful reasoning and configuration awareness, revealing consistent and interpretable failure patterns. Finally, we present a user-centric synthesis pipeline that enables scalable generation of coherent service environments, user configurations, and reliably annotated tasks, supporting benchmark extensibility and future research.

cs.AI↗

arXiv2026

MemoBench: Benchmarking World Modeling in Dynamically Changing Environments

Video generation models aspire to simulate dynamic environments, and several benchmarks now evaluate memory consistency across frames. However, most assess consistency only while the target remains in view, and the few that force objects out of view evaluate static scenes where nothing changes during occlusion. To bridge this gap, we introduce MemoBench, a diagnostic benchmark built around the disappear-and-reappear paradigm in dynamically changing environments: a target object undergoes a physical process, disappears from view, and must be correctly recovered in its updated state upon reappearance. We curate 360 ground-truth clips spanning synthetic and real-world scenes, and design an evaluation suite combining automated metrics with VQA-based assessment across four diagnostic pillars. Evaluation of eight state-of-the-art models reveals key insights and open challenges regarding memory consistency under the disappear-and-reappear paradigm.

cs.CV↗

arXiv2026

Infra-Swarm: Robust Vision-Based Multi-Robot Swarming via Near-Infrared Spectral Vision

Distributed swarms typically rely on either active wireless communication or passive vision, and they are frequently hindered by bandwidth constraints or environmental sensitivity. This paper proposes Infra-Swarm, a robust vision-based swarm. Each robot is equipped with a near-infrared light source and four ordinary gray-scale cameras. The Infra-Swarm system directly measures the centimeter-level 3D position of neighbors based on the position (bearing) and intensity (strength) of optical flares in the captured images. By utilizing 940 nm narrow-band filters to physically reject 99.2% of ambient light interference, the perception front-end achieves hardware-level robustness against illumination variations. Furthermore, its minimal computational overhead provides a resilient foundation for the massive scalability of robotic collectives on resource-constrained hardware.

cs.RO↗

arXiv2026

A Bearing-Strength Method for Motion Estimation of Unknown Energy Emitters

This paper studies motion estimation of moving energy emitters using passive sensors. The emitters may be light, acoustic, or radio sources. While the bearing vector pointing from the sensor to the emitter can be easily obtained, existing approaches mainly rely on the bearing-only motion estimation method. However, this method suffers from a fundamental limitation that the sensor must have lateral motion to ensure observability. Unfortunately, this lateral motion requirement often conflicts with the sensor's desired motion in many tasks. In this paper, we point out that the received signal strength, which can also be obtained easily in many ways, can greatly enhance motion estimation. Surprisingly, this strength information has not been well explored so far. Here, we propose a new bearing-strength method to fully exploit both the bearing and strength measurements. Our theoretical analysis shows that the system observability is significantly enhanced in the sense that the lateral motion condition is not required anymore. Real-world experimental results verify the proposed method and the theoretical analysis. It is notable that the benefit of the proposed method comes with no additional cost since it simply utilizes the received strength information that has not been fully exploited in the past.

cs.RO↗

arXiv2026

SimWorlds: A Multi-Agent System for Dynamic 3D Scene Creation

LLM agents are increasingly used to translate natural language into 3D scenes in a procedural way, but existing systems focus on static output. Dynamic 4D scenes from text alone, in which liquids flow, particles emit, rigid bodies cascade, and articulated mechanisms move, remain largely unexplored despite their value as editable content and as physics-grounded training data for video generation and embodied AI. Two challenges set the dynamic case apart from static text-to-scene work: an agent must jointly coordinate spatial layout, multiple physics solvers, temporal sequencing, camera, and lighting in a single coherent scene, and verifying motion correctness from rendered video is fundamentally harder than judging a single image. We present SimWorlds: a multi-agent framework that produces dynamic, editable 4D scenes from text, with Blender-specific procedural knowledge, a planner-coder-reviewer workflow driving a fixed ordered sequence of construction stages, a layered scene protocol enforced by a deterministic verifier, and a runtime-state inspection tool suite that catches mechanism failures the rendered image cannot reveal. We also introduce 4DBuildBench, a benchmark for assessing both visual fidelity and physical consistency of the procedural dynamic 3D scenes generated from text prompts. Experiments show that SimWorlds outperforms prior dynamic Blender generation baselines.

cs.AI↗

arXiv2026

iMiGUE-3K: A Large-Scale Benchmark for Micro-Gesture Analysis with Self-Supervised Learning

Emotion understanding is a fundamental challenge in affective computing and artificial intelligence. While existing approaches predominantly focus on facial expressions and speech, they often overlook the rich emotional cues conveyed through body language. Recently, micro-gestures (MGs), unintentional, subconscious movements driven by inner feelings, have attracted increasing attention as an alternative to other cues. However, there are no existing large-scale datasets supporting the pre-training of the MG foundation model. To advance MG research, we present a new benchmark for micro-gesture-based emotion understanding, featuring key contributions with a novel dataset (iMiGUE-3K) and a series of foundation models for different tasks. Using a model-based crowd-sourcing data collection strategy, we construct iMiGUE-3K, the largest MG dataset to date. It comprises video recordings from 332 distinct professional tennis players' public press interviews over the past seven years, totaling more than 3.4K long video clips and 37 million frames. The dataset includes 32 micro-gesture classes with rich descriptive annotations, making it the first large-scale, in-the-wild, video dataset for fine-grained gesture-based emotion analysis. Built on iMiGUE-3K, we propose MG-FMs, a discriminative foundation model for transferable gesture presentation learning. Based on the foundation model, we establish five comprehensive evaluation tasks: MG recognition (unsupervised, semi-supervised, supervised), MG retrieval, and MG emotion recognition. Our systematic evaluation of representative methods demonstrates that micro-gesture-based analysis significantly improves emotion understanding. We hope this work can provide comprehensive tools for MG analysis and set a solid foundation for future research in psychological diagnostics, affective computing, and advanced human-computer interaction.

cs.CV↗

arXiv2026

MOTOR-Bench: A Real-world Dataset and Multi-agent Framework for Zero-shot Human Mental State Understanding

Understanding human mental states from natural behavior is crucial for intelligent systems in the real world. However, most current research focuses on predicting isolated mental state labels, lacking structured annotations of complex interpersonal interactions. To support structured analysis, we introduce MOTOR-Bench, a carefully-designed benchmark with a real-world dataset MOTOR-dataset, containing 1,440 multimodal video clips in collaborative learning scenarios, reflecting key real-world data challenges including natural class imbalance, visual noise, and domain-specific language. Each sample is labeled by educational experts based on self-regulated learning theory. We further evaluate several state-of-the-art multimodal large language models and multi-agent systems in a zero-shot setting on our MOTOR-Bench. However, their performance on this task remains limited, suggesting that existing methods still struggle with structured reasoning from observable behavior to deeper mental states. To address this challenge, we propose a reasoning multi-agent framework, named MOTOR-MAS. It coordinates multiple agents through a structured agent coordination mechanism to infer explicit behaviors, internal cognitions, and psychological emotions. Experimental results show that our MOTOR-MAS outperforms the best single-model benchmark by 15.93 points in Macro-F1 scores for the three labels of behavior, cognition, and emotion, and outperforms the general multi-agent benchmark by 10.2 points in internal cognition prediction.

cs.CV↗

arXiv2026

SayNext-Bench: Why Do LLMs Struggle with Next-Utterance Anticipation?

We explore the use of large language models (LLMs) for next-utterance anticipation in human dialogue. Despite recent advances in LLMs demonstrating their ability to engage in natural conversations with users, we show that even leading models surprisingly struggle to anticipate a human speaker's next utterance. Instead, humans can readily anticipate forthcoming utterances based on multi-modal cues -- such as gestures, gaze, and emotional tone -- from the context. To systematically examine this gap, we propose SayNext-Bench, a benchmark evaluating MLLMs on anticipating context-conditioned responses across diverse real-world scenarios. To support it, we build SayNext-PC, a large-scale multimodal dialogue dataset, and carefully design a multi-level evaluation framework spanning lexical similarity, emotion-intention consistency, and LLM-based overall alignment. Building on this, we develop SayNext-Chat, a cognitively inspired dual-route MLLM that incorporates learnable priming tokens to fuse perceptual cues with anticipatory priors. Extensive experiments demonstrate that SayNext-Chat consistently outperforms state-of-the-art MLLMs across all evaluation levels, corroborated by user studies and LLM-as-Judge evaluations. Our results emphasize the (i) indispensable role of multimodal cues and (ii) active anticipatory processing as foundations of natural human interaction currently missing in MLLMs.

cs.AI↗

arXiv2026

DeepSeek-V4: Towards Highly Efficient Million-Token Context Intelligence

We present a preview version of DeepSeek-V4 series, including two strong Mixture-of-Experts (MoE) language models -- DeepSeek-V4-Pro with 1.6T parameters (49B activated) and DeepSeek-V4-Flash with 284B parameters (13B activated) -- both supporting a context length of one million tokens. DeepSeek-V4 series incorporate several key upgrades in architecture and optimization: (1) a hybrid attention architecture that combines Compressed Sparse Attention (CSA) and Heavily Compressed Attention (HCA) to improve long-context efficiency; (2) Manifold-Constrained Hyper-Connections (mHC) that enhance conventional residual connections; (3) and the Muon optimizer for faster convergence and greater training stability. We pre-train both models on more than 32T diverse and high-quality tokens, followed by a comprehensive post-training pipeline that unlocks and further enhances their capabilities. DeepSeek-V4-Pro-Max, the maximum reasoning effort mode of DeepSeek-V4-Pro, redefines the state-of-the-art for open models, outperforming its predecessors in core tasks. Meanwhile, DeepSeek-V4 series are highly efficient in long-context scenarios. In the one-million-token context setting, DeepSeek-V4-Pro requires only 27% of single-token inference FLOPs and 10% of KV cache compared with DeepSeek-V3.2. This enables us to routinely support one-million-token contexts, thereby making long-horizon tasks and further test-time scaling more feasible. The model checkpoints are available at https://huggingface.co/collections/deepseek-ai/deepseek-v4.

cs.CL↗

arXiv2026

MTT-Bench: Predicting Social Dominance in Mice via Multimodal Large Language Models

Understanding social dominance in animal behavior is critical for neuroscience and behavioral studies. In this work, we explore the capability of Multimodal Large Language Models(MLLMs) to analyze raw behavioral video of mice and predict their dominance hierarchy. We introduce MTT-Bench, a novel benchmark comprising annotated videos of pairwise mouse interactions for Mouse Tube Test analysis. Building on existing MLLM architectures, we fine-tune these models to perform zero-shot inference on unseen behavioral sequences, predicting social dominance without explicit labels during testing. Our framework demonstrates promising results, showing high agreement with tube test rankings. This work opens a new direction for applying foundation models to ethology and social behavior analysis, without the need to design domain-specific models.

eess.IV↗

arXiv2026

Amoeba: Runtime Tensor Parallel Transformation for LLM Inference Services

In Large Language Model (LLM) inference services, it is challenging to make a parallelism strategy configuration, to efficiently process the requests of variance context lengths. Requests of long context require high degree of parallelism to provide more memory for Key-Value (KV) Cache, while requests of short context prefer low degree of parallelism to increase concurrency, thus improving throughput. To maintain high throughput while supporting large context lengths on demand, we propose Amoeba, a runtime Tensor Parallel (TP) transformation for online LLM inference services, which adaptively adjusts the TP of running instances to align with the dynamics of incoming requests. Evaluations using real-world traces show that Amoeba improves throughput by 1.75x-6.57x compared to state-of-the-art solutions.

cs.DC↗

arXiv2026

Adaptive Action Chunking at Inference-time for Vision-Language-Action Models

In Vision-Language-Action (VLA) models, action chunking (i.e., executing a sequence of actions without intermediate replanning) is a key technique to improve robotic manipulation abilities. However, a large chunk size reduces the model's responsiveness to new information, while a small one increases the likelihood of mode-jumping, jerky behavior resulting from discontinuities between chunks. Therefore, selecting the optimal chunk size is an urgent demand to balance the model's reactivity and consistency. Unfortunately, a dominant trend in current VLA models is an empirical fixed chunk length at inference-time, hindering their superiority and scalability across diverse manipulation tasks. To address this issue, we propose a novel Adaptive Action Chunking (AAC) strategy, which exploits action entropy as the cue to adaptively determine the chunk size based on current predictions. Extensive experiments on a wide range of simulated and real-world robotic manipulation tasks have demonstrated that our approach substantially improves performance over the state-of-the-art alternatives. The videos and source code are publicly available at https://lance-lot.github.io/adaptive-chunking.github.io/.

cs.RO↗

arXiv2026

Bias Is a Subspace, Not a Coordinate: A Geometric Rethinking of Post-hoc Debiasing in Vision-Language Models

Vision-Language Models (VLMs) have become indispensable for multimodal reasoning, yet their representations often encode and amplify demographic biases, resulting in biased associations and misaligned predictions in downstream tasks. Such behavior undermines fairness and distorts the intended alignment between vision and language. Recent post-hoc approaches attempt to mitigate bias by replacing the most attribute-correlated embedding coordinates with neutral values. However, our systematic analysis reveals three critical limitations of this coordinate-wise approach: feature entanglement, poor cross-dataset generalization, and incomplete bias removal. We find that bias is not localized to a few coordinates but is instead distributed across a few linear subspaces. To address these limitations, we propose $\textbf{S}$ubspace $\textbf{P}$rojection $\textbf{D}$ebiasing ($\textbf{SPD}$), a geometrically principled framework that identifies and removes the entire subspace of linearly decodable bias while reinserting a neutral mean component to preserve semantic fidelity. Extensive experiments across zero-shot classification, text-to-image retrieval, and image generation validate the effectiveness of SPD: our method achieves more robust debiasing with an average improvement of $18.5\%$ across four fairness metrics, while maintaining minimal loss in task performance compared to the best debiasing baseline.

cs.CV↗

arXiv2026

Composer 2 Technical Report

Composer 2 is a specialized model designed for agentic software engineering. The model demonstrates strong long-term planning and coding intelligence while maintaining the ability to efficiently solve problems for interactive use. The model is trained in two phases: first, continued pretraining to improve the model's knowledge and latent coding ability, followed by large-scale reinforcement learning to improve end-to-end coding performance through stronger reasoning, accurate multi-step execution, and coherence on long-horizon realistic coding problems. We develop infrastructure to support training in the same Cursor harness that is used by the deployed model, with equivalent tools and structure, and use environments that match real problems closely. To measure the ability of the model on increasingly difficult tasks, we introduce a benchmark derived from real software engineering problems in large codebases including our own. Composer 2 is a frontier-level coding model and demonstrates a process for training strong domain-specialized models. On our CursorBench evaluations the model achieves a major improvement in accuracy compared to previous Composer models (61.3). On public benchmarks the model scores 61.7 on Terminal-Bench and 73.7 on SWE-bench Multilingual in our harness, comparable to state-of-the-art systems.

cs.SE↗

arXiv2026

MVRD-Bench: Multi-View Learning and Benchmarking for Dynamic Remote Photoplethysmography under Occlusion

Remote photoplethysmography (rPPG) is a non-contact technique that estimates physiological signals by analyzing subtle skin color changes in facial videos. Existing rPPG methods often encounter performance degradation under facial motion and occlusion scenarios due to their reliance on static and single-view facial videos. Thus, this work focuses on tackling the motion-induced occlusion problem for rPPG measurement in unconstrained multi-view facial videos. Specifically, we introduce a Multi-View rPPG Dataset (MVRD), a high-quality benchmark dataset featuring synchronized facial videos from three viewpoints under stationary, speaking, and head movement scenarios to better match real-world conditions. We also propose MVRD-rPPG, a unified multi-view rPPG learning framework that fuses complementary visual cues to maintain robust facial skin coverage, especially under motion conditions. Our method integrates an Adaptive Temporal Optical Compensation (ATOC) module for motion artifact suppression, a Rhythm-Visual Dual-Stream Network to disentangle rhythmic and appearance-related features, and a Multi-View Correlation-Aware Attention (MVCA) for adaptive view-wise signal aggregation. Furthermore, we introduce a Correlation Frequency Adversarial (CFA) learning strategy, which jointly enforces temporal accuracy, spectral consistency, and perceptual realism in the predicted signals. Extensive experiments and ablation studies on the MVRD dataset demonstrate the superiority of our approach. In the MVRD movement scenario, MVRD-rPPG achieves an MAE of 0.90 and a Pearson correlation coefficient (R) of 0.99. The source code and dataset will be made available.

cs.CV↗