28 papers indexed
Hao Wang, Kun Yuan, Wenlin Zhong, Minglei Zhang, Han Xiao, Ming Sun, et al.
Open-weight language models from different families exhibit complementary capabilities, motivating their consolidation into a compact student through on-policy distillation (OPD). However, full-vocabulary OPD typically assumes a shared tokenizer, while existing cross-tokenizer me…
Kaixiong Gong, Xin Cai, Bin Lin, Hao Wang, Yunlong Lin, Mingzhe Zheng, et al.
Unified multimodal models seek a shared visual token space that supports both multimodal understanding and image generation. Discrete methods unify the interface via a shared codebook, whereas continuous pipelines often rely on two disparate representations -- semantic features (…
Wenyuan Wang, Yusong Zhao, Zihao Xu, Hengyi Wang, Qi Xu, Zhigang Hua, et al.
Modern recommender systems are typically based on deep learning (DL) models, where a dense encoder learns representations of users and items. As a result, these systems often suffer from the black-box nature and computational complexity of the underlying models, making it difficu…
Hao Wang, Haoran Geng, Xiaotong Yang, Jing Tang, Songlin Wei, Linlong Lang, et al.
Stereo matching is a fundamental task in 3D reconstruction. Despite remarkable advances, the prevailing paradigms formulate stereo matching as a deterministic regression problem, collapsing the multimodal distribution modeling into a single-point estimation. This formulation suff…
Guangyan Chen, Meiling Wang, Te Cui, Zichen Zhou, Qi Shao, Shalfun Li, et al.
The ability to acquire skills rapidly and effortlessly while retaining those already mastered is essential for robots. However, current methods still rely on a cumbersome training-time loop that is costly and slow, while eroding skills already mastered. In this paper, we introduc…
Hao Wu, Jinjing Zhu, Nanyu Wu, Qianyi Cai, Heyi Lin, Hao Wang, et al.
Edit-conditioned 3D scene retrieval pairs a reference 3D room with a natural-language modification and retrieves rooms from a corpus that satisfy the edit. Three lines of prior work each fall short on this task. 2D composed image retrieval reasons over pixel-level edits and has n…
Hao Wu, Heyi Lin, Zilin Wang, Huizai Yao, Hao Wang, Hui Xiong
Large pretrained vision models have substantially improved appearance-based 3D shape retrieval, but they still confuse shapes that look similar while differing in geometry. Although geometry-aware features can reduce these errors, naive fusion of geometry and appearance may hurt…
Xianghong Fang, Litao Guo, Hengchao Chen, Yuxuan Zhang, XiaofanXia, Dingjie Song, et al.
The effectiveness of modern visual representation learning and autoregressive models critically depends on vector quantization (VQ), which discretizes continuous feature representations using a learnable codebook. Despite its widespread use, existing VQ methods often suffer from…
Ximeng Zhai, Zheng Wang, Yaohong Chen, Hao Wang, Ming-Ming Cheng, Yimian Dai
Due to the optical diffraction limit and long imaging distances, Closely-Spaced Infrared Small Targets (CSIST) typically exhibit energy overlap, manifesting as indistinguishable blobs in infrared images. This ambiguity invalidates the one-to-one mapping assumption of traditional…
Yi Zhao, Jiajun Gao, Chenyang Xu, Yuxi Zhou, Hao Wang
Deploying deep learning models for automated electrocardiogram classification on resource-constrained wearable devices remains challenging due to high computational costs. To address this, we propose LSTrans, a lightweight hybrid model designed for efficient and sensitive ECG ana…
Wenyuan Wang, Lianyu Hu, Hao Wang, Yang Liu
Video-language models (VLMs) have achieved remarkable performance on video understanding and visual question answering, yet they remain unreliable in reasoning about physical plausibility, where understanding object interactions, causal dynamics, and fundamental physical principl…
Jie Liu, Yue Wang, Danping Zeng, Xiaowen Mao, Yunyun Yu, Huijun Feng, et al.
Lulin Liu, Nuo Chen, Yan Wang, Bangya Liu, Wenyan Cong, Hezhen Hu, et al.
Scaling robust driving policies is fundamentally bottlenecked by the scarcity of edge cases in curated datasets. While the real world continuously captures these critical events, such long-tail events remain underutilized when collected from heterogeneous sources. Specifically, d…
Ligong Han, Kai Xu, Hao Wang, Ruijiang Gao, Han Gao, Akash Srivastava
Fully homomorphic encryption (FHE) enables computation on encrypted data, but practical encrypted Transformer inference is bottlenecked by the sequential composition of many nonlinear blocks. We study whether Structured Newton Layer Parallelism (SNLP) can make this inter-layer co…
Tianwen Zhu, Hao Wang, Yonggang Wen
Public battery aging datasets are a critical asset for advanced health management, but their practical use is often limited by inconsistent formats, unclear schemas, and metadata scattered across repositories and publications. Current curation remains largely manual and hard to r…
Hao Wang, Di Yang, Jiangtao Wang
Skeleton-based action recognition models have recently shown strong performance on large-scale benchmarks with general actions. However, directly transferring them to domain-specific tasks e.g., healthcare monitoring, is often suboptimal, as such tasks are narrow in scope and may…
Yifan Wang, Li Li, Youlong Wang, Chengyuan Yang, Sherwin Stephen Chan, Jiaye Chen, et al.
The decline of human balance control due to aging and pathological conditions increases fall risk, a major concern in geriatric care and rehabilitation. Gait training is essential for balance recovery, enhancing walking ability and postural control. However, existing overground r…
Yufan Lu, Xinhui Liu, Chenyang Xu, Yuxi Zhou, Hao Wang
In multi-source ECG deployment, models may need to incorporate new data sources when earlier raw ECGs cannot be retained or replayed. Freezing a pretrained backbone and assigning each source an isolated classifier prevents parameter interference, but deployment still requires sel…
Xin Li, Wenhui Zhu, Xuanzhao Dong, Xiwen Chen, Yanxi Chen, Yujian Xiong, et al.
Medical image segmentation is dominated by U-Net-style encoder-decoder architectures. Vision Transformers (ViTs) overcome the limited receptive field of convolutional networks through self-attention, enabling modeling of long-range dependencies. Early ViT-based segmentation metho…
Ziyang Lian, Qingya Zhang, Hao Wang, Huiwen Xiong, Qi Yang, Lingyi Meng, et al.
Professional translation is often a team-based process: translators, reviewers, and project managers must coordinate terminology, legal force, and accountability across documents. Yet many LLM-based translation tools treat human corrections as isolated edits. Expert decisions mad…
Sicheng Yu, Dongxu Shen, Beizhen Zhao, Guanzhi Ding, Hao Wang
Scaling monocular 3D Gaussian Splatting (3DGS) SLAM to kilometer-level outdoor environments poses two tightly coupled challenges: fragile long-term pose tracking and excessive memory overhead during large-scale mapping. In this paper, we propose KiloGS-SLAM, a highly efficient an…
Maolin Liu, Fanyu Xu, Ruoqing Xu, Jiahang Zhang, Hao Wang, Rui Wang
Navigating the deluge of heterogeneous medical data, from academic literature (PubMed) to clinical guidelines (Web) and private knowledge bases, remains a critical bottleneck for evidence-based medicine. While commercial black-box tools lack transparency, standard open-source RAG…
Hao Wang, Jiuzhou Lei, Dayou Li, Bangya Liu, Minghui Zheng, Manling Li, et al.
Behavior-cloned policies often learn multiple behavior modes from demonstration datasets, including modes that are unsafe or otherwise undesired at deployment. For example, a policy trained on diverse handover demonstrations may learn to pass a knife blade-first. Standard remedie…
Liton Kumar Biswas, M Shafkat M Khan, Himanandhan Reddy Kottur, Hao Wang, Hamed Dalir, Navid Asadizanjani
Silicon photonics enables integration of optical components using standard semiconductor processes, greatly improving data communication bandwidth and energy efficiency. However, photonics integrated circuits (PICs) face unique security challenges, such as counterfeit or tamperin…
Vincent Chen, Starrick Liu, Regis Cheng, Dance Yang, Shalfun Li, Ryan Yu, et al.
Matrix-orthogonalization-based optimizers, exemplified by Muon, have demonstrated strong convergence behavior across a wide range of modern deep learning workloads. The matrix-aware updates offer a compelling alternative to conventional element-wise optimization, particularly as…
Hao Wang, Omkar Salunkhe, Annalena Hartmann, Sven Ekered, Patrick Bründl, Jörg Franke, et al.