Wenyue Chong

Yaocheng Zhang, Yuanheng Zhu, Wenyue Chong et al.

Deep search agents have emerged as a promising paradigm for addressing complex information-seeking tasks, but their training remains challenging due to sparse rewards, weak credit assignment, and limited labeled data. Self-play offers a scalable route to reduce data dependence, but conventional self-play optimizes students only through sparse outcome rewards, leading to low learning efficiency. In this work, we observe that self-play naturally produces a question construction path (QCP) during task generation, an intermediate artifact that captures the reverse solution process. This reveals a new source of privileged information for self-distillation: self-play can itself provide high-quality privileged context for the teacher model in a low-cost and scalable manner, without relying on human feedback or curated privileged information. Leveraging this insight, we propose Privileged Information Self-Play ($π$-Play), a multi-agent self-evolution framework. In $π$-Play, an examiner generates tasks together with their QCPs, and a teacher model leverages QCP as privileged context to densely supervise a student via self-distillation. This design transforms conventional sparse-reward self-play into a dense-feedback self-evolution loop. Extensive experiments show that data-free $π$-Play surpasses fully supervised search agents and improves evolutionary efficiency by 2-3$\times$ over conventional self-play.

Jingbo Sun, Wenyue Chong, Songjun Tu et al.

Agentic retrieval-augmented generation (RAG) systems enable large language models (LLMs) to solve complex tasks through multi-step interaction with external retrieval tools. However, such multi-step interaction often involves redundant search steps, incurring substantial computational cost and latency. Prior work limits search depth (i.e., the number of search steps) to reduce cost, but this often leads to underexploration of complex questions. To address this, we first investigate how search depth affects accuracy and find a minimal sufficient search depth that defines an accuracy-efficiency trade-off, jointly determined by question complexity and the agent's capability. Furthermore, we propose AutoSearch, a reinforcement learning (RL) framework that evaluates each search step via self-generated intermediate answers. By a self-answering mechanism, AutoSearch identifies the minimal sufficient search depth and promotes efficient search by rewarding its attainment while penalizing over-searching. In addition, reward mechanisms are introduced to stabilize search behavior and improve answer quality on complex questions. Extensive experiments on multiple benchmarks show that AutoSearch achieves a superior accuracy-efficiency trade-off, alleviating over-searching while preserving search quality.

Wenyue Chong

Semantic correspondence aims to identify semantically meaningful relationships between different images and is a fundamental challenge in computer vision. It forms the foundation for numerous tasks such as 3D reconstruction, object tracking, and image editing. With the progress of large-scale vision models, semantic correspondence has achieved remarkable performance in controlled and high-quality conditions. However, the robustness of semantic correspondence in challenging scenarios is much less investigated. In this work, we establish a novel benchmark for evaluating semantic correspondence in adverse conditions. The benchmark dataset comprises 14 distinct challenging scenarios that reflect commonly encountered imaging issues, including geometric distortion, image blurring, digital artifacts, and environmental occlusion. Through extensive evaluations, we provide several key insights into the robustness of semantic correspondence approaches: (1) All existing methods suffer from noticeable performance drops under adverse conditions; (2) Using large-scale vision models can enhance overall robustness, but fine-tuning on these models leads to a decline in relative robustness; (3) The DINO model outperforms the Stable Diffusion in relative robustness, and their fusion achieves better absolute robustness; Moreover, We evaluate common robustness enhancement strategies for semantic correspondence and find that general data augmentations are ineffective, highlighting the need for task-specific designs. These results are consistent across both our dataset and real-world benchmarks.

Xiaoran Xu, Jiangang Yang, Wenyue Chong et al.

Single-Domain Generalized Object Detection~(S-DGOD) aims to train an object detector on a single source domain while generalizing well to diverse unseen target domains, making it suitable for multimedia applications that involve various domain shifts, such as intelligent video surveillance and VR/AR technologies. With the success of large-scale Vision-Language Models, recent S-DGOD approaches exploit pre-trained vision-language knowledge to guide invariant feature learning across visual domains. However, the utilized knowledge remains at a coarse-grained level~(e.g., the textual description of adverse weather paired with the image) and serves as an implicit regularization for guidance, struggling to learn accurate region- and object-level features in varying domains. In this work, we propose a new cross-modal feature learning method, which can capture generalized and discriminative regional features for S-DGOD tasks. The core of our method is the mechanism of Cross-modal and Region-aware Feature Interaction, which simultaneously learns both inter-modal and intra-modal regional invariance through dynamic interactions between fine-grained textual and visual features. Moreover, we design a simple but effective strategy called Cross-domain Proposal Refining and Mixing, which aligns the position of region proposals across multiple domains and diversifies them, enhancing the localization ability of detectors in unseen scenarios. Our method achieves new state-of-the-art results on S-DGOD benchmark datasets, with improvements of +8.8\%~mPC on Cityscapes-C and +7.9\%~mPC on DWD over baselines, demonstrating its efficacy.