|
我们实验室专注于研究机器学习在开放世界环境中的泛化能力，涵盖理论研究、算法创新和系统开发。此外，我们尤其关注面向具身智能的多模态大模型研究，如自动驾驶和机器人技术。我们致力于最大限度地发挥人工智能的潜力，并将其应用于基础科学，从而推动AI for Science的实际进展。我们还特别专注于探索大脑的认知机制，开发能够在标签有限的情况下适应新领域和新模态的通用且高效的机器学习系统。 Our laboratory focuses on the generalization of machine learning in open-world settings, encompassing theoretical research, algorithm innovation, and system development. Additionally, we pay special attentions to multimodal large language model domains, such as autonomous driving and robotics. We are committed to fully leveraging the potential of artificial intelligence and applying it in fundamental sciences to truly advance AI for Science. We also place particular emphasis on exploring cognitive mechanisms of the brain to develop universal and efficient machine learning systems capable of adapting to new domains and modalities with limited labeling. |
 Dr. Shanghang Zhang is a Tenure Track Assistant Professor at the Computer Science Department of Peking University. She has been the postdoc research fellow at Berkeley AI Research Lab (BAIR), EECS, UC Berkeley. Her research focuses on OOD Generalization that enables the machine learning systems to generalize to new domains, categories, and modalities using limited labels, with applications to autonomous driving and robotics, as reflected in her over 40 papers on top-tier journals and conference proceedings (Google Scholar Citations: 13366, H-index: 42, I10-index: 98). Her recent work “Informer: Beyond Efficient Transformer for Long Sequence Time-Series Forecasting” has received the AAAI 2021 Best Paper Award. It ranks the 1st place of Trending Research on PaperWithCode and its Github receives 2,600+ Stars. Shanghang has been selected to “2018 Rising Stars in EECS, USA”. She has also been awarded the Adobe Academic Collaboration Fund, Qualcomm Innovation Fellowship (QInF) Finalist Award, and Chiang Chen Overseas Graduate Fellowship. Dr. Zhang has been the chief organizer of several workshops on ICML/NeurIPS, and the special issue on ICMR. Dr. Zhang received her Ph.D. from Carnegie Mellon University in 2018, and her Master from Peking University.. |
|
|
|
| Our team proposed RoboMamba, which is an advanced multimodal system excels in reasoning and manipulation, tripling inference speed compared to existing models. It achieves precise task completion in simulations and real-world experiments. | Our team has jointly released a significant dataset RoboMind, in collaboration with The National Joint Engineering Laboratory of Intelligent Robotics and Systems in the field of embodied intelligence. | Our team has created an AI multilingual commentary system for the Hangzhou Asian Games, enriching the viewer experience with integrated visual, language, and speech inputs. |
|
© Copyright 2025 HMI Lab - All rights reserved
|
