arXiv subjects

Wenhao Chai

Publications and source records attributed to Wenhao Chai.

At least 37 records · Page 2Linked to original sources

arXiv2026

OpenDeepThink: Parallel Reasoning via Bradley-Terry Aggregation

Test-time compute scaling is a primary axis for improving LLM reasoning. Existing methods primarily scale depth by extending a single reasoning trace. Scaling breadth by sampling multiple candidates in parallel is straightforward, but introduces a selection bottleneck: choosing the best candidate without a ground-truth verifier, since pointwise LLM judging is noisy and biased. To address this, we introduce OpenDeepThink, a population-based test-time compute framework that selects via pairwise Bradley-Terry comparison. Each generation, the LLM judges random pairs of candidates and aggregates votes via Bradley-Terry into a global ranking; top-ranked candidates are preserved and the top three quarters are mutated using the natural-language critiques produced during comparison; the bottom quarter is discarded. OpenDeepThink raises Gemini 3.1 Pro's effective Codeforces Elo by +405 points in eight sequential LLM-call rounds (~27 minutes wall-clock). The pipeline transfers across weaker and stronger models without retuning, and on the multi-domain HLE benchmark, gains appear concentrated in objectively verifiable domains and reverse in subjective ones. We release CF-73, a curated set of 73 expert-rated Codeforces problems with International Grandmaster annotation and 99% local-evaluation agreement against the official verdict.

cs.AI↗

arXiv2026

FrontierSmith: Synthesizing Open-Ended Coding Problems at Scale

Many real-world coding challenges are open-ended and admit no known optimal solution. Yet, recent progress in LLM coding has focused on well-defined tasks such as feature implementation, bug fixing, and competitive programming. Open-ended coding remains a weak spot for LLMs, largely because open-ended training problems are scarce and expensive to construct. Our goal is to synthesize open-ended coding problems at scale to train stronger LLM coders. We introduce FrontierSmith, an automated system for iteratively evolving open-ended problems from existing closed-ended coding tasks. Starting from competitive programming problems, FrontierSmith generates candidate open-ended variants by changing the problems'goals, restricting outputs, and generalizing inputs. It then uses a quantitative idea divergence metric to select problems that elicit genuinely diverse approaches from different solvers. Agents then generate test cases and verifiers for the surviving candidates. On two open-ended coding benchmarks, training on our synthesized data yields substantial gains over the base models: Qwen3.5-9B improves by +8.82 score on FrontierCS and +306.36 (Elo-rating-based performance) on ALE-bench; Qwen3.5-27B improves by +12.12 and +309.12, respectively. The synthesized problems also make agents take more turns and use more tokens, similar to human-curated ones, suggesting that closed-ended seeds can be a practical starting point for long-horizon coding data.

cs.LG↗

arXiv2026

Weak-to-Strong Knowledge Distillation Accelerates Visual Learning

Large-scale visual learning is increasingly limited by training cost. Existing knowledge distillation methods transfer from a stronger teacher to a weaker student for compression or final-accuracy improvement. We instead investigate distillation to accelerate the training of strong students. We propose a generalizable plug-and-play recipe that freezes a weaker teacher, applies distillation only in early training, and turns it off once the student reaches and surpasses teacher-level performance. For ImageNet and CIFAR classification, this strategy reaches target thresholds much earlier, with up to 4.8 times speedup measured by epochs. We confirm that the method generalizes to other tasks and report 1.7 times epoch speedup for object detection on the COCO dataset, and 2.5 times earlier target-FID crossing for diffusion generation on the CIFAR-10 dataset, measured in steps. These findings validate our method as a universal speedup mechanism for visual learning.

cs.CV↗

arXiv2026

Science-T2I: Addressing Scientific Illusions in Image Synthesis

Current image generation models produce visually compelling but scientifically implausible images, exposing a fundamental gap between visual fidelity and physical realism. In this work, we introduce ScienceT2I, an expert-annotated dataset comprising a training set of over 20k adversarial image pairs and 9k prompts across 16 scientific domains and an isolated test set of 454 challenging prompts. Using this benchmark, we evaluate 18 recent image generation models and find that none scores above 50 out of 100 under implicit scientific prompts, while explicit prompts that directly describe the intended outcome yield scores roughly 35 points higher, confirming that current models can render correct scenes when told what to depict but cannot reason from scientific cues to the correct visual outcome. To address this, we develop SciScore, a reward model fine-tuned from CLIP-H that captures fine-grained scientific phenomena without relying on language-guided inference, surpassing GPT-4o and experienced human evaluators by roughly 5 points. We further propose a two-stage alignment framework combining supervised fine-tuning with masked online fine-tuning to inject scientific knowledge into generative models. Applying this framework to FLUX.1[dev] yields a relative improvement exceeding 50% on SciScore, demonstrating that scientific reasoning in image generation can be substantially improved through targeted data and alignment.

cs.CV↗

arXiv2026

Agent Banana: High-Fidelity Image Editing with Agentic Thinking and Tooling

We study instruction-based image editing under professional workflows and identify three persistent challenges: (i) editors often over-edit, modifying content beyond the user's intent; (ii) existing models are largely single-turn, while multi-turn edits can alter object faithfulness; and (iii) evaluation at around 1K resolution is misaligned with real workflows that often operate on ultra high-definition images (e.g., 4K). We propose Agent Banana, a hierarchical agentic planner-executor framework for high-fidelity, object-aware, deliberative editing. Agent Banana introduces two key mechanisms: (1) Context Folding, which compresses long interaction histories into structured memory for stable long-horizon control; and (2) Image Layer Decomposition, which performs localized layer-based edits to preserve non-target regions while enabling native-resolution outputs. To support rigorous evaluation, we build HDD-Bench, a high-definition, dialogue-based benchmark featuring verifiable stepwise targets and native 4K images (11.8M pixels) for diagnosing long-horizon failures. On HDD-Bench, Agent Banana achieves the best multi-turn consistency and background fidelity (e.g., IC 0.871, SSIM-OM 0.84, LPIPS-OM 0.12) while remaining competitive on instruction following, and also attains strong performance on standard single-turn editing benchmarks. We hope this work advances reliable, professional-grade agentic image editing and its integration into real workflows.

cs.CV↗

arXiv2026

VideoNSA: Native Sparse Attention Scales Video Understanding

Video understanding in multimodal language models remains limited by context length: models often miss key transition frames and struggle to maintain coherence across long time scales. To address this, we adapt Native Sparse Attention (NSA) to video-language models. Our method, VideoNSA, adapts Qwen2.5-VL through end-to-end training on a 216K video instruction dataset. We employ a hardware-aware hybrid approach to attention, preserving dense attention for text, while employing NSA for video. Compared to token-compression and training-free sparse baselines, VideoNSA achieves improved performance on long-video understanding, temporal reasoning, and spatial benchmarks. Further ablation analysis reveals four key findings: (1) reliable scaling to 128K tokens; (2) an optimal global-local attention allocation at a fixed budget; (3) task-dependent branch usage patterns; and (4) the learnable combined sparse attention help induce dynamic attention sinks. Project Page: https://enxinsong.com/VideoNSA-web/, Code: https://github.com/Espere-1119-Song/VideoNSA

cs.CV↗

arXiv2026

UEval: A Benchmark for Unified Multimodal Generation

We introduce UEval, a benchmark to evaluate unified models, i.e., models capable of generating both images and text. UEval comprises 1,000 expert-curated questions that require both images and text in the model output, sourced from 8 real-world tasks. Our curated questions cover a wide range of reasoning types, from step-by-step guides to textbook explanations. Evaluating open-ended multimodal generation is non-trivial, as simple LLM-as-a-judge methods can miss the subtleties. Different from previous works that rely on multimodal Large Language Models (MLLMs) to rate image quality or text accuracy, we design a rubric-based scoring system in UEval. For each question, reference images and text answers are provided to a MLLM to generate an initial rubric, consisting of multiple evaluation criteria, and human experts then refine and validate these rubrics. In total, UEval contains 10,417 validated rubric criteria, enabling scalable and fine-grained automatic scoring. UEval is challenging for current unified models: GPT-5-Thinking scores only 66.4 out of 100, while the best open-source model reaches merely 49.1. We observe that reasoning models often outperform non-reasoning ones, and transferring reasoning traces from a reasoning model to a non-reasoning model significantly narrows the gap. This suggests that reasoning may be important for tasks requiring complex multimodal understanding and generation.

cs.CV↗

arXiv2026

Terminal-Bench: Benchmarking Agents on Hard, Realistic Tasks in Command Line Interfaces

AI agents may soon become capable of autonomously completing valuable, long-horizon tasks in diverse domains. Current benchmarks either do not measure real-world tasks, or are not sufficiently difficult to meaningfully measure frontier models. To this end, we present Terminal-Bench 2.0: a carefully curated hard benchmark composed of 89 tasks in computer terminal environments inspired by problems from real workflows. Each task features a unique environment, human-written solution, and comprehensive tests for verification. We show that frontier models and agents score less than 65\% on the benchmark and conduct an error analysis to identify areas for model and agent improvement. We publish the dataset and evaluation harness to assist developers and researchers in future work at https://www.tbench.ai/ .

cs.SE↗

arXiv2026

Reasoning Matters for 3D Visual Grounding

The recent development of Large Language Models (LLMs) with strong reasoning ability has driven research in various domains such as mathematics, coding, and scientific discovery. Meanwhile, 3D visual grounding, as a fundamental task in 3D understanding, still remains challenging due to the limited reasoning ability of recent 3D visual grounding models. Most of the current methods incorporate a text encoder and visual feature encoder to generate cross-modal fuse features and predict the referring object. These models often require supervised training on extensive 3D annotation data. On the other hand, recent research also focus on scaling synthetic data to train stronger 3D visual grounding LLM, however, the performance gain remains limited and non-proportional to the data collection cost. In this work, we propose a 3D visual grounding data pipeline, which is capable of automatically synthesizing 3D visual grounding data along with corresponding reasoning process. Additionally, we leverage the generated data for LLM fine-tuning and introduce Reason3DVG-8B, a strong 3D visual grounding LLM that outperforms previous LLM-based method 3D-GRAND using only 1.6% of their training data, demonstrating the effectiveness of our data and the importance of reasoning in 3D visual grounding.

cs.CV↗

arXiv2025

Next-Embedding Prediction Makes Strong Vision Learners

Inspired by the success of generative pretraining in natural language, we ask whether the same principles can yield strong self-supervised visual learners. Instead of training models to output features for downstream use, we train them to generate embeddings to perform predictive tasks directly. This work explores such a shift from learning representations to learning models. Specifically, models learn to predict future patch embeddings conditioned on past ones, using causal masking and stop gradient, which we refer to as Next-Embedding Predictive Autoregression (NEPA). We demonstrate that a simple Transformer pretrained on ImageNet-1k with next embedding prediction as its sole learning objective is effective - no pixel reconstruction, discrete tokens, contrastive loss, or task-specific heads. This formulation retains architectural simplicity and scalability, without requiring additional design complexity. NEPA achieves strong results across tasks, attaining 83.8% and 85.3% top-1 accuracy on ImageNet-1K with ViT-B and ViT-L backbones after fine-tuning, and transferring effectively to semantic segmentation on ADE20K. We believe generative pretraining from embeddings provides a simple, scalable, and potentially modality-agnostic alternative to visual self-supervised learning.

cs.CV↗

arXiv2025

FrontierCS: Evolving Challenges for Evolving Intelligence

We introduce FrontierCS, a benchmark of 156 open-ended problems across diverse areas of computer science, designed and reviewed by experts, including CS PhDs and top-tier competitive programming participants and problem setters. Unlike existing benchmarks that focus on tasks with known optimal solutions, FrontierCS targets problems where the optimal solution is unknown, but the quality of a solution can be objectively evaluated. Models solve these tasks by implementing executable programs rather than outputting a direct answer. FrontierCS includes algorithmic problems, which are often NP-hard variants of competitive programming problems with objective partial scoring, and research problems with the same property. For each problem we provide an expert reference solution and an automatic evaluator. Combining open-ended design, measurable progress, and expert curation, FrontierCS provides a benchmark at the frontier of computer-science difficulty. Empirically, we find that frontier reasoning models still lag far behind human experts on both the algorithmic and research tracks, that increasing reasoning budgets alone does not close this gap, and that models often over-optimize for generating merely workable code instead of discovering high-quality algorithms and system designs.

cs.LG↗

arXiv2025

UniHPR: Unified Human Pose Representation via Singular Value Contrastive Learning

In recent years, there has been a growing interest in developing effective alignment pipelines to generate unified representations from different modalities for multi-modal fusion and generation. As an important component of Human-Centric applications, Human Pose representations are critical in many downstream tasks, such as Human Pose Estimation, Action Recognition, Human-Computer Interaction, Object tracking, etc. Human Pose representations or embeddings can be extracted from images, 2D keypoints, 3D skeletons, mesh models, and lots of other modalities. Yet, there are limited instances where the correlation among all of those representations has been clearly researched using a contrastive paradigm. In this paper, we propose UniHPR, a unified Human Pose Representation learning pipeline, which aligns Human Pose embeddings from images, 2D and 3D human poses. To align more than two data representations at the same time, we propose a novel singular value-based contrastive learning loss, which better aligns different modalities and further boosts performance. To evaluate the effectiveness of the aligned representation, we choose 2D and 3D Human Pose Estimation (HPE) as our evaluation tasks. In our evaluation, with a simple 3D human pose decoder, UniHPR achieves remarkable performance metrics: MPJPE 49.9mm on the Human3.6M dataset and PA-MPJPE 51.6mm on the 3DPW dataset with cross-domain evaluation. Meanwhile, we are able to achieve 2D and 3D pose retrieval with our unified human pose representations in Human3.6M dataset, where the retrieval error is 9.24mm in MPJPE.

cs.CV↗

arXiv2025

MM-HELIX: Boosting Multimodal Long-Chain Reflective Reasoning with Holistic Platform and Adaptive Hybrid Policy Optimization

While current Multimodal Large Language Models (MLLMs) have demonstrated proficiency in reasoning tasks such as mathematics and logic, their capacity for long-chain reflective reasoning, a prerequisite for solving complex real-world problems, remains largely underexplored. In this work, we first conduct an extensive empirical investigation to evaluate this capability. Leveraging a carefully designed data synthesis engine, we construct MM-HELIX, a multimodal benchmark consisting 1,260 samples of 42 challenging synthetic tasks that require iterative thinking and backtracking. Empirical results on this benchmark reveal that existing MLLMs exhibit significant performance deficits in long-chain reflective reasoning. To address this limitation, we generate post-training data and further explore learning paradigms for exploiting such data. We first develop the Step-Elicited Response Generation pipeline to create MM-HELIX-100K, a large-scale dataset of 100k high-quality, reflective reasoning traces for instruction-tuning stage. Given that standard Reinforcement Learning fails on complex tasks due to sparse reward signals and catastrophic forgetting after Supervised Fine-Tuning, we propose Adaptive Hybrid Policy Optimization (AHPO), a novel training strategy that dynamically unifies offline supervision and online optimization into a single stage. This strategy enables the model to learn from expert data when rewards are sparse and conduct independent exploration once proficient. When applied to the Qwen2.5-VL-7B baseline, our method achieves a +18.6\% accuracy improvement on MM-HELIX benchmark and demonstrates strong generalization with a +5.7\% average performance gain on general mathematic and logic tasks. Our work demonstrate that reflective reasoning in MLLMs can be effectively learned and generalized, paving the way for developing more capable MLLMs.

cs.CV↗

arXiv2025

AutoCode: LLMs as Problem Setters for Competitive Programming

Writing competitive programming problems is exacting. Authors must: set constraints, input distributions, and edge cases that rule out shortcuts; target specific algorithms (e.g., max-flow, dynamic programming, data structures); and calibrate complexity beyond the reach of most competitors. We argue that this makes for an ideal test of general large language model capabilities and study whether they can do this reliably. We introduce AutoCode, which uses multiple rounds of validation to yield competition-grade problem statements and test cases. On held-out problems, AutoCode test suites approach 99% consistency with official judgments, a significant improvement over current state-of-the-art methods like HardTests, which achieve less than 81%. Furthermore, starting with a random seed problem, AutoCode can create novel variants with reference and brute-force solutions. By cross-verifying these generated solutions against test cases, we can further filter out malformed problems. Our system ensures high correctness, as verified by human experts. AutoCode successfully produces novel problems judged by Grandmaster-level (top 0.3%) competitive programmers to be of contest quality.

cs.SE↗

arXiv2025

Dense Video Understanding with Gated Residual Tokenization

High temporal resolution is essential for capturing fine-grained details in video understanding. However, current video large language models (VLLMs) and benchmarks mostly rely on low-frame-rate sampling, such as uniform sampling or keyframe selection, discarding dense temporal information. This compromise avoids the high cost of tokenizing every frame, which otherwise leads to redundant computation and linear token growth as video length increases. While this trade-off works for slowly changing content, it fails for tasks like lecture comprehension, where information appears in nearly every frame and requires precise temporal alignment. To address this gap, we introduce Dense Video Understanding (DVU), which enables high-FPS video comprehension by reducing both tokenization time and token overhead. Existing benchmarks are also limited, as their QA pairs focus on coarse content changes. We therefore propose DIVE (Dense Information Video Evaluation), the first benchmark designed for dense temporal reasoning. To make DVU practical, we present Gated Residual Tokenization (GRT), a two-stage framework: (1) Motion-Compensated Inter-Gated Tokenization uses pixel-level motion estimation to skip static regions during tokenization, achieving sub-linear growth in token count and compute. (2) Semantic-Scene Intra-Tokenization Merging fuses tokens across static regions within a scene, further reducing redundancy while preserving dynamic semantics. Experiments on DIVE show that GRT outperforms larger VLLM baselines and scales positively with FPS. These results highlight the importance of dense temporal information and demonstrate that GRT enables efficient, scalable high-FPS video understanding.

cs.CV↗

arXiv2025

AuroraLong: Bringing RNNs Back to Efficient Open-Ended Video Understanding

The challenge of long video understanding lies in its high computational complexity and prohibitive memory cost, since the memory and computation required by transformer-based LLMs scale quadratically with input sequence length. We propose AuroraLong to address this challenge by replacing the LLM component in MLLMs with a linear RNN language model that handles input sequence of arbitrary length with constant-size hidden states. To further increase throughput and efficiency, we combine visual token merge with linear RNN models by reordering the visual tokens by their sizes in ascending order. Despite having only 2B parameters and being trained exclusively on public data, AuroraLong achieves performance comparable to Transformer-based models of similar size trained on private datasets across multiple video benchmarks. This demonstrates the potential of efficient, linear RNNs to democratize long video understanding by lowering its computational entry barrier. To our best knowledge, we are the first to use a linear RNN based LLM backbone in a LLaVA-like model for open-ended video understanding.

cs.CV↗

arXiv2025

ToSA: Token Merging with Spatial Awareness

Token merging has emerged as an effective strategy to accelerate Vision Transformers (ViT) by reducing computational costs. However, existing methods primarily rely on the visual token's feature similarity for token merging, overlooking the potential of integrating spatial information, which can serve as a reliable criterion for token merging in the early layers of ViT, where the visual tokens only possess weak visual information. In this paper, we propose ToSA, a novel token merging method that combines both semantic and spatial awareness to guide the token merging process. ToSA leverages the depth image as input to generate pseudo spatial tokens, which serve as auxiliary spatial information for the visual token merging process. With the introduced spatial awareness, ToSA achieves a more informed merging strategy that better preserves critical scene structure. Experimental results demonstrate that ToSA outperforms previous token merging methods across multiple benchmarks on visual and embodied question answering while largely reducing the runtime of the ViT, making it an efficient solution for ViT acceleration. The code will be available at: https://github.com/hsiangwei0903/ToSA

cs.CV↗

arXiv2025

LiveCodeBench Pro: How Do Olympiad Medalists Judge LLMs in Competitive Programming?

Recent reports claim that large language models (LLMs) now outperform elite humans in competitive programming. Drawing on knowledge from a group of medalists in international algorithmic contests, we revisit this claim, examining how LLMs differ from human experts and where limitations still remain. We introduce LiveCodeBench Pro, a benchmark composed of problems from Codeforces, ICPC, and IOI that are continuously updated to reduce the likelihood of data contamination. A team of Olympiad medalists annotates every problem for algorithmic categories and conducts a line-by-line analysis of failed model-generated submissions. Using this new data and benchmark, we find that frontier models still have significant limitations: without external tools, the best model achieves only 53% pass@1 on medium-difficulty problems and 0% on hard problems, domains where expert humans still excel. We also find that LLMs succeed at implementation-heavy problems but struggle with nuanced algorithmic reasoning and complex case analysis, often generating confidently incorrect justifications. High performance appears largely driven by implementation precision and tool augmentation, not superior reasoning. LiveCodeBench Pro thus highlights the significant gap to human grandmaster levels, while offering fine-grained diagnostics to steer future improvements in code-centric LLM reasoning.

cs.SE↗