天桥脑科学研究院

Our News

研究院动态

重度抑郁症(MDD)是全球致残的主要原因之一,三分之一的病例对治疗产生抵抗性。症状的异质性表明受影响的大脑网络存在差异,这推动了基于脑回路的个性化神经调节干预的研究。例如,个性化的深部脑刺激(DBS)已通过根据表型或映射刺激反应选择不同的治疗靶点实现。然而,DBS是侵入性的,且在动态适应性大脑中,长期最佳治疗效果的稳定性尚不清楚。非侵入性方法,如经颅磁刺激(TMS),在调节假定的情绪网络方面显示出前景,但无法针对较深的皮质下区域。

经颅超声刺激(TUS)或低强度聚焦超声是一种新兴的非侵入性方法,具有毫米级的空间精度,能够独特地实现深层皮质下神经调节。TUS可以可逆地调节大脑网络,并产生持久的行为效应。初步研究表明,将TUS应用于经典的TMS和DBS靶点可以改善焦虑、担忧、回避行为和情绪。为了探究动态引导的TUS是否能够在MDD中识别个性化的治疗亚区域,陈氏学者Joline M. Fan博士在加州大学旧金山分校与其他研究人员合作,使用双相阵列交叉光束聚焦刺激皮质下与情绪相关的神经回路。

22

2024-11

在一项新研究中,研究人员报告称,神奇蘑菇中的活性化合物裸盖菇素(psilocybin)会暂时扰乱大脑中负责内省思维的关键网络,比如白日梦和回忆。

食用含有裸盖菇素的蘑菇(即通常所说的神奇蘑菇)的人通常会经历一种超现实的体验,他们的空间、时间和自我感知会被扭曲。

长期以来,支持者们一直认为,在适当的条件下,迷幻体验可以减轻精神痛苦,而一些科学研究也表明事实可能确实如此。准确理解这种药物如何影响大脑将帮助科学家和医生更好地利用其治疗潜力。

新的研究成果从神经生物学的角度解释了这种药物扭曲大脑感官的原因,并为开发基于裸盖菇素的疗法奠定了基础,此类疗法主要针对治疗抑郁症和创伤后应激障碍等精神疾病。

“刚开始的影响会比较大,但会逐渐减弱,最后只剩下微弱且范围有限的影响,”该研究的资深合著者、圣路易斯华盛顿大学医学院神经学教授Nico U.F. Dosenbach说道。

“这正是我们希望潜在药物能做到的。我们并不希望药物导致病患的大脑好几天完全宕机,但也不希望它立即恢复原状。你希望药效的持续时间足够长,能够真的起效。”

这项研究最近发表在《自然》期刊上,为其他科学家提供了一张可以遵循的路线图,以评估精神活性药物对大脑功能的影响,从而加速精神疾病相关药物的开发工作。

这项工作的支持来自泰勒家族创新精神病研究基金、健康心灵实验室、麦克唐纳系统神经科学中心、华盛顿大学临床和转化科学研究所;美国国立卫生研究院(NIH)、全国痉挛性发声障碍协会、拉尔夫·梅茨纳教授基金、天桥脑科学研究院、华盛顿大学智力与发育障碍研究中心、吉瓦尼斯基金会、华盛顿大学希望中心神经疾病研究所以及马林克罗特放射学研究所。

此外,这项研究还利用了由NIH资助的青少年大脑认知发展研究的数据。

22

2024-11

近日,天桥脑科学研究院(Tianqiao and Chrissy Chen Institute, TCCI)当选《中国慈善家》杂志技术公益慈善影响力年度案例。《中国慈善家》是中国新闻社主管、《中国新闻周刊》主办的央媒体系中唯一的公益慈善领域专业媒体。

据《影响力慈善研究报告2023》介绍,研究院聚焦于利用人工智能和前沿技术提升大脑功能,将慈善资金与前沿技术相结合,致力于在大脑科学领域取得突破性进展。报告表示,TCCI已在大脑探知、大脑相关疾病治疗以及大脑功能开发等技术应用领域取得突破。比如支持、资助了中科院微系统所和华山医院神经外科团队的脑机接口项目,研发出一种超薄、超柔、高通量的神经信号采集电极。

22

2024-11

 

今年12月6日和7日,天桥脑科学研究院(Tianqiao and Chrissy Chen Institute, TCCl)与BCI协会(BCI Society)将在中国上海合作举办BCI Society & Chen Institute Joint BCI Meeting,标志这一球脑机接口领域顶级学术会议创办25年后首次落户亚洲。大会将汇聚全球顶尖学者与前沿研究团队,打造一个开放创新的学术交流平台,推动跨学科合作和全球对话,致力于脑机接口技术的科学突破与产业化应用,加强亚洲与全球研究人员的深度互动,共同构建一个真正的全球脑机接口社区。

大会议程涵盖从基础研究到临床应用、从技术开发到产业转化的多个层次,展示脑机接口领域的前沿突破和未来发展趋势。会议期间将举办十余场主题演讲,邀请来自全球知名学术与研究机构的重要学术领袖与行业先锋,共同分享他们在神经技术、脑科学与人工智能的最新研究成果,深入探讨脑机接口在神经康复、临床医疗与人类能力增强等方面的应用前景与挑战。

大会特别重视跨学科合作与行业交流,设置了丰富的社交环节,为与会者提供轻松的交流平台,促进来自不同领域专家之间的深度对话、经验分享与合作机会的探索。

 2024国际脑机接口大会将以丰富的内容和多样的形式,带来关于脑机接口未来的深度探讨与前瞻性思考。我们诚邀全球神经科学、临床、人工智能、数据、新材料、集成电路等相关领域的科学家及产业界的同行参会,共同见证脑机接口技术如何重新定义未来,推动人类认知边界的进一步拓展。

 

演讲嘉宾

Jocelyne Bloch, MD

Neurosurgeon, Lausanne University Hospital

Full Professor at Faculty of Life Science at Ecole Polytechnique Fédérale de Lausanne (EPFL)

Founder and Director of Defitech Center at Defitech Center for Interventional Neurotherapies

Chief Scientific Officer at ONWARD Medical N. V.

Grégoire Courtine, Ph.D.

Full Professor at Faculty of Life Science at Ecole Polytechnique Fédérale de Lausanne (EPFL)

Adjunct Professor at Neurosurgical Department at University Hospital Lausanne

Founder and Director of Defitech Center at Defitech Center for Interventional Neurotherapies

Chief Scientific Officer at ONWARD Medical N. V.

Jocelyne Bloch is neurosurgeon at the University Hospital Lausanne (CHUV) where she leads the functional neurosurgery unit, while Grégoire Courtine is a neuroscientist with a background in physics. Jocelyne and Grégoire are Professor within the NeuroX institute of the Ecole Polytechnique Fédérale de Lausanne (EPFL), within the neurosurgical department of CHUV, and at the Faculty of Medicine of the University of Lausanne (UNIL). Together, they founded the Defitech Center for Interventional Neurotherapies, named .NeuroRestore, which develop bioengineering strategies involving neurosurgical interventions to restore neurological functions. In 2014, they also co-founded ONWARD Medical (Euronext: ONWD) with the aim to translate the neurotherapies developed at .NeuroRestore into clinical treatments. Jocelyne and Grégoire are known worldwide for the conception of neuroprosthetic implants that restored walking in people with chronic paralysis.

 

Robert Gaunt, Ph.D.

Associate Professor and Engineering Director, Rehab Neural Engineering Labs, 

Department of Physical Medicine and Rehabilitation at the University of Pittsburgh

Associate Professor, Physical Medicine and Rehabilitation, Department of Biomedical Engineering at Carnegie Mellon University

Robert Gaunt is an Associate Professor of Physical Medicine and Rehabilitation at the University of Pittsburgh. He holds a BEng in Mechanical Engineering from the University of Victoria and a PhD in Biomedical Engineering from the University of Alberta. Dr. Gaunt’s research focuses on neuroprosthetics for sensorimotor control of the hand and bladder, aiming to restore function after injury or disease. His work includes developing brain-computer interfaces to enable movement and sensation for those with upper-limb paralysis and creating neural interfaces to regulate bladder function. Recognized by the National Academy of Engineering, his research has been featured widely in media, and he holds multiple patents.

 

Bo Hong, Ph.D.

Professor, Department of Biomedical Engineering, Tsinghua University

Principle Investigator, McGovern Institute for Brain Research

Dr. Bo Hong received his Ph.D. degree of Biomedical Engineering from Tsinghua University in 2001. From 2004 to 2005, he was a visiting scientist in the Department of Biomedical Engineering and the Center for Neural Engineering at Johns Hopkins University, USA. He is now full professor with School of Biomedical Engineering, Tsinghua University, and an investigator of McGovern Institute for Brain Research at Tsinghua. His main research interests are brain computer interface and language network in human brain. His team designed and developed minimally invasive BCI – NEO system and conducted the first-in-human clinical trial successfully in 2023. He has co-authored more than 80 papers on Nature Neuroscience, PNAS, Nature Communications, NeuroImage, Journal of Neuroscience etc. He has served as the Associate Editor of IEEE Transactions on Biomedical Engineering and IEEE Transactions on Neural Systems and Rehabilitation Engineering.

 

Cory Inman, Ph.D.

Professor, Department of Biomedical Engineering, Tsinghua University

Principle Investigator, McGovern Institute for Brain Research

Director, Immersive Neuromodulation and Neuroimaging (INMAN) Laboratory in the Psychology

Department at the University of Utah

Dr. Cory Inman is the director of the Immersive Neuromodulation and Neuroimaging (INMAN) Laboratory in the Psychology Department at the University of Utah. He received his BA in Psychology from Georgia State University and his PhD from Emory University. He completed his first postdoctoral fellowship in the Neurosurgery department at Emory University and a second postdoctoral fellowship at UCLA. He has broad interests in helping to establish approaches that push our understanding of emotion and memory from the laboratory into the wild, real world. The INMAN Lab is currently undertaking studies examining the use of direct brain stimulation to the human amygdala for episodic memory enhancement and direct brain recordings of deep memory structures, like the hippocampus during, navigation and autobiographical memory encoding in real-world settings. Outside the lab, he spends his time playing with his 2 kids, playing guitar, playing basketball, and exploring as many outdoor adventures as possible including white water rafting, snowboarding, hiking, and climbing.

 

Camille Jeunet-Kelway, Ph.D.

CNRS Research Scientist

INCIA, University of Bordeaux & CNRS, France

Deputy head of the SMART platform, University of Bordeaux

Camille Jeunet-Kelway received her PhD in cognitive sciences in 2016 at the University of Bordeaux, France. After a post-doctoral fellowship in Inria (Rennes, France) and EPFL (Geneva, Switzerland), she was recruited as a tenured CNRS Research Scientist. In 2021, she has joined the institute for cognitive and integrative neurosciences (INCIA) in Bordeaux, where she leads interdisciplinary research on the use of EEG-BCIs to improve or restore cognitive and motor abilities, both for clinical (stroke patients and patients with Parkinson disease) and non-clinical (athletes) populations. She is particularly interested in studying the learning mechanisms underlying neurofeedback training as well as the acceptability of neurofeedback procedures and BCI technologies. Camille Jeunet-Kelway has received 3 PhD awards, the European Label as well as 5 national fundings from the French research agency for her research. In 2022 she was awarded the Early Career Award in neuroscience from the BCI Society. Since 2024, she is deputy head of the SMART reasearch and innovation plateform, an “intelligent, connected gym”, dedicated to sports and movement sciences.

 

Bao-Liang Lu, Ph.D.

Director for the Center for Brain-Like Computing and Machine Intelligence

The Key Laboratory of Shanghai Education Commission Intelligent Interaction

Cognitive Engineering, and Ruijin-Mihoyo Laboratory

Bao-Liang Lu received his PhD in electrical engineering from Kyoto University, Kyoto, Japan, in 1994. Since August 2002, he has been a full professor at the Department of Computer Science and Engineering, Shanghai Jiao Tong University, China. He currently serves as the Director for the Center for Brain-Like Computing and Machine Intelligence, the Key Laboratory of Shanghai Education Commission Intelligent Interaction and Cognitive Engineering, and Ruijin-Mihoyo Laboratory. He received the 2018 IEEE Transactions on Autonomous Mental Development Outstanding Paper Award and the 2021 Best of IEEE Transactions on Affective Computing Paper Collection. He is also the Associate Editor of the IEEE Transactions on Affective Computing and Journal of Neural Engineering and an IEEE Fellow. His research interests include deep learning, large EEG model, emotion artificial intelligence, and affective brain-computer interface.

 

Natalie Mrachacz-Kersting, Ph.D.

Director, Laboratory for Brain-Computer Interfaces and Neurorehabilitation, Albert-Ludwigs-University Freiburg

Chair of Neuroscience in Sport and Movement, Institute for Sport and Sport Science 

VP of MAC, IEEE Engineering in Medicine and Biology Society

Deputy Editor-in-Chief, IEEE TNSRE

Prof. Dr. Natalie Mrachacz-Kersting, a member of IEEE, received her Ph.D. degree in biomedical engineering from Aalborg University in 2005. She currently holds the Chair for Neuroscience and Neuroscience in Sport at the Albert-Ludwigs University of Freiburg and is a member of the BrainLinks-BrainTools Cluster of Excellence at IMBIT, Albert-Ludwigs University of Freiburg. Dr. Mrachacz-Kersting serves on the Executive Committee of the IEEE Engineering in Medicine and Biology Society (EMBS) as the Vice President-elect for Member and Student Activities. She is also the Chair of the IEEE Women in Biomedical Engineering (WI(BM)E), member of the Steering Committee of IEEE Brain Technical Community, and the Deputy Editor-in Chief of the IEEE Transactions on Neural Systems and Rehabilitation Engineering Journal. She has previously held positions at Aalborg University in Denmark, FH Dortmund in Germany, and the University of Auckland in New Zealand. Dr. Mrachacz Kersting’s research focuses on medical technology, biomedical engineering, and neuroscience. She has authored over 80 peer-reviewed journal articles, more than 130 conference papers and abstracts, ten book chapters, and one book. Her current projects primarily involve Brain-Computer Interfaces (BCIs) for patient populations, including those suffering from stroke or ALS. Dr. Mrachacz-Kersting received several awards including the international BCI award in 2017.

 

Mu-ming Poo, Ph.D.

Scientific Director, Institute of Neuroscience, Chinese Academy of Sciences (CAS)

Director, Shanghai Center for Brain Science and Brain-Inspired Technology

Paul Licht Distinguished Professor in Biology Emeritus, University of California, Berkeley

Mu-ming Poo is the Scientific Director of Institute of Neuroscience, Chinese Academy of Sciences (CAS), Director of Shanghai Center for Brain Science and Brain-Inspired Technology, and Paul Licht Distinguished Professor in Biology Emeritus at University of California, Berkeley. He studied physics at Tsinghua University in Taiwan and received PhD in biophysics from Johns Hopkins University in 1974. During 1976-2012, He had served on the faculty of UC Irvine, Yale, Columbia, and UCSD, and UC Berkeley. He was the founding director of Institute of Neuroscience, CAS (1999-2019), and a member of Chinese Academy of Science, Academia Sinica, and Hong Kong Academy of Science, and an international member of US National Academy of Science. He was awarded Ameritec Prize, International Science & Technology Cooperation Award of P. R. China, and Gruber Neuroscience Prize. Poo’s research interest includes axon growth, synaptic plasticity, and the use non-human primates to study higher cognitive functions and human brain disorders. He is the Executive Editor-in-Chief of National Science Review and the editorial board member for many journals, including Neuron and Progress in Neurobiology.

 

Mao Ying, MD

President of Huashan Hospital at Fudan University

Dr. Ying Mao is a highly respected neurosurgeon at Huashan Hospital in Shanghai, China. He now serves as the director of the National Center for Neurological Disorders, the chairman-designate of the Neurosurgery Section of the Chinese Medical Association and the president of Huashan Hospital. With years of experience under his belt, Dr. Mao is known for his expertise in treating complex neurological conditions and performing advanced brain surgeries. Patients appreciate his compassionate approach and his dedication to staying at the forefront of medical technology and techniques. Beyond the operating room, Dr. Mao is an avid researcher, contributing significantly to the field of neurosurgery with numerous publications and studies. He’s also a beloved mentor, guiding the next generation of neurosurgeons with his wealth of knowledge and experience. When he’s not working, Dr. Mao enjoys spending time with his family, exploring new culinary delights, and indulging in a good book. His commitment to his patients, coupled with his warm personality, makes him a standout figure in the medical community.

 

Alexander von Lühmann, Ph.D.

Head of Independent Research Group

BIFOLD-ML | Machine Learning Department

Technische Universität Berlin

Alexander von Lühmann is currently head of the “Intelligent Biomedical Sensing” research group at TU Berlin’s Machine Learning department and BIFOLD. He is also a visiting researcher at the Neurophotonics Center of Boston University (BU NPC) and the Lead Technology Advisor at NIRx Medical Technologies. Before this, he was the Chief Science Officer and R&D Director at NIRx for 2,5 years, a post-doc at Boston University, a visiting researcher at Harvard Medical School, and the Chief Technology

Officer at Crely, a healthcare startup based in the US and Singapore. He is a member of the ISO/DIN, SfNIRS, OPTICA, OHBM and VDI. His contributions to the field have been recognized by the fNIRS Society (Early Investigator Award 2022), the TU Berlin BIMOS graduate school (PhD Award 2019), the German Society for Biomedical Engineering (DGBMT Klee Award 2018) and the IEEE Biomedical Engineering Society (Special Feature & Cover Article 2017). He received his PhD (Dr.-Ing.) with distinction in 2018 from TU Berlin, and his M.Sc. and B.Sc. degrees in Electrical Engineering from Karlsruhe Institute of Technology in 2014/11.

 

青年科学家

Enming Song, Ph.D.

Associate Professor at Institute of Optoelectronics of Fudan University

Enming Song is an Associate Professor at the Institute of Optoelectronics, Fudan University. Previously, he was a postdoctoral fellow at Northwestern University’s Simpson Querrey Institute and an Adjunct Research Assistant Professor at the University of Illinois at UrbanaChampaign. He earned his Ph.D. and B.S. from Fudan University’s Department of Materials Science. His research focuses on soft electronic materials for biomedical engineering, particularly flexible bioelectronic systems for neural interfaces. Over the past five years, he has published 21 scientific papers and holds a U.S. patent. His contributions earned him honors like MIT Technology Review’s Innovators Under 35 Asia Pacific, Shanghai Science & Technology 35 Under 35, and a Global Innovation Award from UNIDO.

 

Minpeng Xu, Ph.D.

Associate Dean of Academy of Medical Engineering and Translational Medicine and Chair Professor of

Department of Biomedical Engineering at Tianjin University

Minpeng Xu received the B. S. and Ph. D. degrees in biomedical engineering from Tianjin University in 2010 and 2015, respectively. He visited the Tzyy-Ping Jung’s lab at the Institute for Neural Computation (INC) of University of California, San Diego (UCSD) from 2014 to 2015. He is currently the Associate Dean of Academy of Medical Engineering and Translational Medicine and Chair Professor of Department of Biomedical Engineering at Tianjin University. His research interests include brain-computer interface, neural signal processing and neuromodulation. He has published more than 80 academic papers as the (co-)first or (co-)corresponding author, and some of them were selected as ESI highly cited papers, IEEE TBME cover story, and JNE highlight.

 

Yuxiao Yang, Ph.D.

Assistant Professor, MOE Frontier Science Center for Brain Science and Brain-Machine Integration and

the State Key Laboratory of Brain-Machine Intelligence, Zhejiang University

Yuxiao Yang is an Assistant Professor at the MOE Frontier Science Center for Brain Science and Brain-machine Integration and the State Key Laboratory of Brain-machine Intelligence, Zhejiang University. Prior to joining Zhejiang University, he was an Assistant Professor of Electrical and Computer Engineering (ECE) at the University of Central Florida (UCF).  He received a Ph.D. degree in ECE from University of Southern California in 2019 and received a B.S. degree in Electronics Engineering from Tsinghua University in 2013.His research centers on designing closed-loop brain-machine interface systems for neural decoding and control, aiming to provide new therapies for neurological and neuropsychiatric disorders. He has published in prestigious neural engineering journals, including cover articles in Nature Biotechnology and Nature Biomedical Engineering. He received the Annual Brain Computer Interface Award in 2019 and the IEEE EMBS Best Student Paper Award in 2015.

 

项目委员会及主持人

Cuntai Guan, Ph.D.

Program Committee and Moderator

President’s Chair in Computer Science and Engineering

Deputy Dean, College of Computing and Data Science

Nanyang Technological University, Singapore

Professor Cuntai Guan is a President’s Chair in Computer Science and Engineering and Deputy Dean of the College of Computing & Data Science at the Nanyang Technological University, Singapore. He is the Director of the Centre for BrainComputing Research, Director of the Artificial Intelligence Research Institute, CoDirector of S-Lab for Advanced Intelligence, and Co-Director of the Rehabilitation Research Institute of Singapore. His research interests include Brain-Computer Interfaces (BCI), Machine Learning, Neural Signal & Image Processing, and Artificial Intelligence. He is a recipient of several awards owing to his contributions to Brain Computer Interface research.

 

Christian Herff, Ph.D.

Program Committee and Moderator

Assistant professor, Mental Health and Neuroscience Research Institute

Maastricht University, the Netherlands

Dr. Christian Herff is an assistant professor in the Mental Health and Neuroscience Research Institute at Maastricht University where he leads the Neural Interfacing Lab. His research interest lays in the application of machine learning technology to neurophysiological data for Brain-Computer Interfaces and neuroscience research. With a particular focus on the decoding of speech processes from intracranial data, he tries to improve the lives of severely paralyzed patients while simultaneously improving our understanding of complex higher order cognition. He emphasizes the ability to achieve interpretable results based on computational models. In particular, visualization of complex dynamic models, such as deep neural networks, is of interest to him.

 

Dean J. Krusienski, Ph.D.

Program Committee and Moderator

Professor and Graduate Program Director

Department of Biomedical Engineering

Virginia Commonwealth University, USA

Dean Krusienski is a Professor and Graduate Program Director of Biomedical Engineering at Virginia Commonwealth University in Richmond, Virginia, where

he also directs the Advanced Signal Processing in Engineering and Neuroscience (ASPEN) Laboratory. He has co-authored over 100 peer-reviewed publications related to advancing brain-computer interface and neural signal analysis techniques, which have collectively received over 11,900 citations. His lab’s work has been funded by NSF, NIH, NASA/NIA and DoD, including current projects on intracranial speech decoding and synthesis; closed-loop DBS; biomarkers of hippocampal and sleep pathologies; user-state estimation; visual, auditory, and memory processing; and virtual reality applications.

 

Yan Li, Ph.D.

Program Committee and Moderator

Executive Director of Scientific Programs

Dr. Yan Li serves as the Executive Director of Scientific Programs at the Tianqiao Chen & Chrissy Luo Institute (TCCI®). She oversees the daily management of the Institute’s initiatives and fosters strong communications with beneficiary organizations and individuals. In addition to collaborating with cornerstone partners, Yan is spearheading the development of programs that empower researchers at all stages of their careers, advancing the overarching mission of the foundation to drive innovation and discovery. Dr. Li received her Ph.D. in Neuroscience at the Institute of Neuroscience, Chinese Academy of Sciences at Shanghai. She spent six years as a postdoctoral fellow at Profession Fred H. Gage’s lab at the Salk Institute and then moved to Stanford where she worked with Professor Thomas Sudhof and Marius Wernig for two years. Prior to joining Chen Institute, Dr. Li worked at a startup company at Menlo Park, California.

 

Mariska Vansteensel, Ph.D.

Program Committee and Moderator

Assistant Professor, University Medical Center Utrecht, the Netherlands

President, International BCI Society

Mariska Vansteensel is an Assistant Professor at the UMC Utrecht Brain Center in Utrecht. Her main research goal is to use the wealth of neuroscientific knowledge directly for the benefit of people with disease or disability. The main focus of her research since 2007 has been the development and validation of implantable electrocorticography (ECoG)-based Brain-Computer Interfaces (BCI) for communication in individuals with severe motor and speech impairment. She has conducted research on the proof of concept, working with epilepsy patients who receive ECoG electrodes for diagnostic purposes and on the first worldwide investigation of the use of fully implantable BCIs in settings of daily living of people with severe motor impairment.

 

Theresa M. Vaughan, BA

Program Committee and Moderator

Research Scientist

National Center for Adaptive Neurotechnologies

Albany Stratton VA Medical Center, USA

Theresa Vaughan is a research scientist with thirty years of experience in clinical research studies, 25 years focused specifically on brain-computer interfaces (BCIs) as new communication channels for people with severe motor disabilities. Her laboratory supervised the first-ever large-scale trial of independent home use of a BCI by people with amyotrophic lateral sclerosis (ALS). This work has become the foundation of the National Center for Adaptive Neurotechnologies (NCAN) translational service project program. To accomplish this, they have developed and tested hardware; modified the BCI2000 software; developed a range of Windows-based applications; produced training tools for therapists, caregivers, and users; and designed software.

 

Tao Hu, Ph.D

Moderator

Deputy Director, Shanghai Institute of Microssystem 

Information Technology, Chinese Academy of Sciences

Tao Hu is a researcher and deputy director at the Shanghai Institute of Microsystemand Information Technology, Chinese Academy of Sciences, and director of the StateKey Laboratory of Transducer Technology. He chairs the Brain-Computer Interfaceand Interaction Branch of the Chinese Society for Neuroscience and oversees theBCI industrial roadmap for the China Association for Science and Technology. He hasreceived numerous awards, including the Shanghai Young Scientists OutstandingContribution Award (2020), CAS Young Scientist Award (2021), and Best Paper Awardat IEEE MEMS 2022. Recognized as a “Highly Cited Chinese Scholar” and among the“Top 2% Scientists in the World,” he is also a celebrated mentor and public sciencecommunicator.

 

Jianjun Meng, Ph.D.

Moderator

Associate Professor Institute of  Robotics,  

Shanghai Jiao Tong University

Jianjun Meng is an associate professor at Shanghai Jiao Tong University (SJTU,2019–present). He earned his Bachelor’s and Ph.D. in Mechanical Engineeringat SJTU and conducted postdoctoral research at the University of Minnesota andCarnegie Mellon University (2014–2019). His research focuses on noninvasive brain-computer interface (BCI), neural prosthetics, biomedical signal processing, andneural engineering. He has over 40 SCI-indexed publications in journals like ScienceRobotics and NeuroImage and co-authored a book chapter in Neural Engineering.A senior IEEE member, he is an associate editor for IEEE biomedical journals. Hereceived China’s Ministry of Education First Prize in Natural Science and is part ofthe Shanghai Pujiang Talent Program.

 

大会议程

 

会议时间:

2024年12月6日-7日

会议地点:

上海斯格威铂尔曼大酒店

(上海市黄浦区打浦路15号,近徐家汇路,金玉兰广场内)

参与方式:

  • 报名注册免费,食宿差旅请参会者自理;

  • 会议席位有限,以主办方发送的参会确认短信为准;

  • 报名注册以及了解更多信息,请扫描文末二维码。

扫码注册参会

 

关于BCI协会

BCI协会(The Brain-Computer Interface Society,BCI Society)是全球最具影响力的脑机接口学术机构,其旗舰活动International BCI Meeting自1999年在美国纽约首次举办以来,已成为该领域的标杆性学术会议。2023年第10届会议在比利时布鲁塞尔举办,吸引了来自全球39个国家和地区的237所实验室的近500位科学家。

 

关于天桥脑科学研究院

天桥脑科学研究院(Tianqiao and  Chrissy Chen Institute, TCCl)是由陈天桥、雒芊芊夫妇出资10亿美元创建的世界最大私人脑科学研究机构之一,围绕全球化、跨学科和青年科学家三大重点,支持脑科学研究,造福人类。

TCCI与华山医院、上海市精神卫生中心设立了应用神经技术前沿实验室、人工智能与精神健康前沿实验室;与加州理工学院合作成立了TCCI加州理工神经科学研究院。

TCCI建成了支持脑科学研究的生态系统,项目遍布欧美、亚洲和大洋洲,包括学术会议和交流、夏校培训AI驱动科学大奖、科研型临床医生奖励计划、特殊病例社区、中文媒体追问等。

21

2024-11

你是否渴望用人工智能推动科学,重塑未来的科研模式?

是否也曾好奇,如果智能体拥有长期记忆,世界将变成怎样?

天桥脑科学研究院“人工智能人才计划”正是为有志向的你而设!不论是加入我们与普林斯顿大学王梦迪教授联合发起的AI for Science实习项目、角逐AI驱动科学大奖,还是拿下超级Offer成为我们的一员,这里为你提供了一个广阔的舞台。你将有机会开拓科研边界,与全球顶尖科学家和优秀同学并肩作战,共同推动AI在科学领域的深度应用,成为引领未来的AI科学先锋!

破解AI的记忆密码,普林斯顿大学王梦迪教授携手天桥脑科学研究院推出联合实习计划

在人工智能领域,长期记忆(Long Term Memory)研究正成为AI能力涌现的新焦点。长期记忆是AI实现自我进化的关键。它不仅关系到AI的个性化发展,更是构建真正智能系统的基础。由天桥脑科学研究院、普林斯顿大学、清华大学、上海交通大学和盛大集团联合发表的论文《Long Term Memory: The Foundation of AI Self-Evolution》深入探讨了这一议题,并提出了创新解决方案:通过模仿人类的“长期记忆”,帮助AI超越LLM的局限。

普林斯顿大学王梦迪团队与天桥脑科学研究院正面向全球发起联合实习计划,招募优秀的学生参与这一前沿科技领域的研究。普林斯顿大学电子与计算工程系、计算机系终身教授,Donald Eckman奖得主王梦迪,是这一研究的领军人物。她表示:“利用长期记忆进行 AI 自我进化以及解决现实世界问题拥有巨大潜力。我们正在寻找有志于在这一领域做出贡献的青年人才,共同拓展人工智能的边界。”

据了解,这项联合实习计划主要面向大模型和AIGC的算法实习生,在读本科和硕博研究生均可申请。申请者需具备深度学习基础、良好的编码能力、对高质量数据的理解,以及LLM常用技术,如SFT、RL算法等,能够快速提升业务场景中的应用效果。逻辑思维严谨、积极主动的团队合作能力也是申请的加分项。

申请者还将享有以下机会:

  • 构建人工智能前沿认知调研,深入了解智能体功能和市场需求,掌握产业数智化进展。
  • 提升智能体驾驭能力,参与智能体的设计、开发、测试和运营,积累产品全生命周期的经验。
  • 深度参与普林斯顿王梦迪教授与天桥脑科学研究院的领先创新项目,积累宝贵的交叉领域工作经验。

 

投递简历: aitalents@cheninstitute.org

AI驱动科学大奖申报倒计时,面向全球寻找AI for Science未来领袖

天桥脑科学研究院(Tianqiao and Chrissy Chen Institute)与《科学》杂志 “AI驱动科学大奖”,旨在表彰全球范围内用人工智能技术(如机器学习、深度学习、强化学习等)在自然科学(包括生命科学和物质科学)研究领域解决关键问题、推动科学研究加速发展的青年科学家。

申请者需提交一篇1000字的研究论文,描述他们最重要的研究发现,主题必须是利用人工智能相关技术,在其它科研领域做出了根本性的突破,而这些突破在未使用人工智能技术前是难以实现的。

该奖项目前已进入申报倒计时,报名截止时间为2024年12月13日,《科学》杂志将组成专家评审委员会评比,于2025年7月公布获奖名单。大奖(The Grand Prize)得主将获得3万美元奖金以及《科学》杂志5年数字版订阅权,获奖论文将发表在《科学》杂志的印刷版和在线版上。此外还设有最多两名优胜奖得主,各获得1万美元奖金和《科学》杂志5年的数字版订阅权,他们的论文将在《科学》杂志在线版上发表。

 

 

加入我们,敲开AI技术创新的大门

同时,我们持续在上海/新加坡/美国三地寻找具备技术热情、创新精神和团队合作精神的AI技术实习生/全职员工,共同致力于利用人工智能解决现实世界中的各种挑战和难题。欢迎所有人工智能方向的优秀人才投递简历:

岗位方向

 

    后端开发与集成:

  • 设计、开发和维护支持AI模型的后端系统和服务
  • 集成各种机器学习模型和算法到后端系统,解决业务问题,如推荐系统、预测分析等
  • 优化后端系统,提高AI模型的性能和效率
  • 设计和实施API和微服务,以使AI功能可通过RESTful接口访问

 

 

    算法研究与开发:

  • 探索研究最先进的大语言模型相关技术,持续跟踪论文和开源社区产品化落地的最新进展
  • 根据业务场景需求,提出算法方案,构建算法原型,验证算法效果
  • 构建高质量的算法服务,跟踪解决线上出现的相关问题

 

 

岗位要求:

  • 对人工智能技术充满热情,有志于在该领域深入学习和发展
  • 具备扎实的编程基础,熟练掌握至少一种编程语言,如Python、Golang等
  • 对后端开发、数据库系统、Web开发、API设计等有一定了解
  • 对机器学习、深度学习等人工智能领域有浓厚兴趣,有相关项目经验者优先考虑
  • 良好的沟通能力和团队合作精神

 

投递简历: aitalents@cheninstitute.org

我们提供国际化的工作环境和广阔的个人发展空间,极具竞争力的薪酬福利。加入我们,你将站在科学与技术的交汇点,挑战最前沿的AI课题,与全球一流的科研团队共同推动AI创新。一起探索人类智能的未来!

更多相关新闻

12

2024-11

由天桥脑科学研究院(Tianqiao and Chrissy Chen Institute)、《科学》(Science)杂志和上海市精神卫生中心(国家精神疾病医学中心)联合主办的“人工智能与精神健康”研讨会于11月7日至8日在上海宛平南路600号上海精中徐汇院区举办,吸引了近400位科学家、临床医生、产业界人士参会,共同探讨人工智能在精神健康领域的应用与未来发展。

 

天桥脑科学研究院创始人雒芊芊说:“人工智能和精神健康是我和天桥寄予厚望的领域,希望能够尽早有所突破,造福人类。这是我们连续第四年与《科学》杂志合办国际学术讲座,很高兴首次落地中国,期待吸引各方人才特别是AI人才加入我们,用人工智能技术,更好地让中国人不焦虑,不抑郁,更幸福。”

《科学》系列期刊出版人Bill Moran在会议现场致辞中表示:“《科学》一直处于新兴科学的前沿。我们将继续联手天桥脑科学研究院,与全球科学家和研究人员密切合作,释放人工智能在精神健康治疗和管理方面的巨大潜力。”

 

 

上海市精神卫生中心院长、国家精神疾病医学中心主任赵敏教授说:”全球心理健康问题日益严峻,人工智能与精神疾病诊疗的结合正是解决这一问题的良方。我们期待与全球科学家进行深度交流,共同推动这一领域的发展。”

 

 

在《科学》高级编辑Peter Stern、上海交通大学心理学院院长傅小兰教授、上海市精神卫生中心副院长王振教授的主持下,会议围绕众多前沿话题展开深入讨论,包括人工智能在精神疾病诊断、个性化治疗方案、心理健康预测模型以及计算精神病学等领域的应用。来自中国、美国、德国、英国、澳大利亚的专家们分享了在各自领域的最新研究成果和实践经验。

德国耶拿大学医院转化精神病学特聘教授Nils Opel的研究团队开发了创新远程监测应用程序,通过收集患者的行为和语音数据来评估其心理健康状况。他领导建立的大规模数字队列研究DigiHero,正在监测德国人口的心理健康,尤其是受新冠疫情的影响。这些技术在临床实践中显示出巨大潜力,可用于患者的细分和进一步个性化治疗。

 

 

 

上海精中赵敏教授展示了AI在精神健康诊疗中的多种临床应用,这些应用正推动精神健康服务向智能化、个性化方向发展。AI辅助诊断系统能够通过分析MRI扫描结果,准确识别重度精神疾病患者。利用虚拟现实和AI技术创建沉浸式治疗环境,显著改善了恐高症等精神疾病患者的症状。他们正在开发基于AI的心理治疗机器人,通过分析面部表情来辅助焦虑症的筛查和诊断。

美国耶鲁大学精神病学系副教授Philip Corlett分享了如何将计算精神病学引入临床诊疗的实践。他的团队开发了一种”条件性幻觉”模型,利用AI和机器学习技术来研究幻觉的形成机制。此外,他们还应用分层高斯滤波器模型,发现偏执症患者倾向于将世界视为更加不稳定的形态。这些计算模型能帮助我们更好地理解精神疾病的机制。

 

 

澳大利亚纽卡斯尔大学神经科学与精神病学教授Michael Breakspear围绕人工智能解码面部表情、大语言模型模拟生成自然和疾病下的言语、生成式人工智能量化大脑异常以及以人为中心的数字未来四个方面,全面地讨论了如何使用大脑的生成式模型来优化AI在精神健康领域的应用。他尝试用模型模拟人类精神世界,在他看来,人类是一种活跃的主体,动态地嵌入自身构建的人际、社会、文化和历史体系中;而深度神经网络是其核心,是执行随机插值的静态矩阵。

 

 

 

 

英国剑桥大学计算认知神经科学教授Zoe Kourtzi主持开发了一种用于早期预测和诊断痴呆症的AI工具。它能够在症状出现前10-15年就识别出潜在的痴呆症患者,准确率高达91%。在实际应用中,这一AI工具比传统临床诊断方法更精确地预测了轻度认知障碍患者转变为阿尔茨海默病的可能性,其准确度提高了3倍。她提出,未来有可能用AI更广泛地追踪大脑的健康轨迹。

 

 

上海精中陈剑华教授介绍,天桥脑科学研究院人工智能与精神健康前沿实验室,正在与上海精中合作开展一项名为”灵溪”的项目。基于精神疾病是“唯一可以靠对话诊疗的疾病“这个特点,建设真实、高质量的抑郁焦虑患者与医生的问诊对话集,提供给AI大语言模型进行训练。项目已通过伦理审批,采集了5000多例对话,时长1000个小时,且数量正在持续增加。项目计划在严格合规的前提下向研究者和AI团队开放合作。

天桥脑科学研究院科学计划执行主任李艳博士介绍了研究院积极推进AI赋能脑科学的举措,包括与《科学》杂志合作发起全球AI驱动科学大奖,与世界名校共同举办AI跨学科夏校培训,今年已在全球主办、资助了近50场AI+脑科学高端国际会议。

 

 

波士顿大学计算和计算科学与工程哈里里研究所所长Ioannis Paschalidis、清华大学副教授眭亚楠、中科院深圳先进技术研究院先进计算与数字工程研究所所长李烨、伦敦大学学院计算精神病学和衰老研究中心副主任Quentin Huys围绕将AI运用于大脑疾病的早期检测和预防、用AI构建心理健康基础模型、AI在抑郁症治疗中的应用等话题,分享了在各自领域的前沿探索和实践。

会议还举办了青年科学家闭门讨论,10多位青年科学家与海内外嘉宾进行了深入交流。

在这两天里,与会专家们展示了AI在精神健康领域的巨大应用潜力。展望未来,人工智能与精神健康的结合将继续深化,并有望在以下几个方面取得突破:

  • AI辅助精神疾病诊疗的精确度和普及度将进一步提高;

  • 个性化治疗方案的制定将更加智能化;

  • AI的发展将帮助我们更深入地理解精神疾病的机制,为新的治疗方法提供理论基础;

  • 通过AI技术,我们将可能实现对大脑健康轨迹的长期追踪,从而更好地预防和管理精神健康问题;

  • AI技术在精神疾病诊疗领域的研究需要特别加强伦理治理问题。

 

 

08

2024-11

 

你是否渴望用人工智能推动科学,重塑未来的科研模式?

是否也曾好奇,如果智能体拥有长期记忆,世界将变成怎样?

天桥脑科学研究院“人工智能人才计划”正是为有志向的你而设!不论是加入我们与普林斯顿大学王梦迪教授联合发起的AI for Science实习项目、角逐AI驱动科学大奖,还是拿下超级Offer成为我们的一员,这里为你提供了一个广阔的舞台。你将有机会开拓科研边界,与全球顶尖科学家和优秀同学并肩作战,共同推动AI在科学领域的深度应用,成为引领未来的AI科学先锋!

 

01

破解AI的记忆密码

普林斯顿大学王梦迪教授携手天桥脑科学研究院推出联合实习计划

在人工智能领域,长期记忆(Long Term Memory)研究正成为AI能力涌现的新焦点。长期记忆是AI实现自我进化的关键。它不仅关系到AI的个性化发展,更是构建真正智能系统的基础。由天桥脑科学研究院、普林斯顿大学、清华大学、上海交通大学和盛大集团联合发表的论文《Long Term Memory: The Foundation of AI Self-Evolution》深入探讨了这一议题,并提出了创新解决方案:通过模仿人类的“长期记忆”,帮助AI超越LLM的局限。

 

 

相关研究发表在arxiv上.图片来源:arxiv

 

普林斯顿大学王梦迪团队与天桥脑科学研究院正面向全球发起联合实习计划,招募优秀的学生参与这一前沿科技领域的研究。普林斯顿大学电子与计算工程系、计算机系终身教授,Donald Eckman奖得主王梦迪,是这一研究的领军人物。她表示:“利用长期记忆进行 AI 自我进化以及解决现实世界问题拥有巨大潜力。我们正在寻找有志于在这一领域做出贡献的青年人才,共同拓展人工智能的边界。”

据了解,这项联合实习计划主要面向大模型和AIGC的算法实习生,在读本科和硕博研究生均可申请。申请者需具备深度学习基础、良好的编码能力、对高质量数据的理解,以及LLM常用技术,如SFT、RL算法等,能够快速提升业务场景中的应用效果。逻辑思维严谨、积极主动的团队合作能力也是申请的加分项。

申请者还将享有以下机会:

  1. 构建人工智能前沿认知调研,深入了解智能体功能和市场需求,掌握产业数智化进展。

  2. 提升智能体驾驭能力,参与智能体的设计、开发、测试和运营,积累产品全生命周期的经验。

  3. 深度参与普林斯顿王梦迪教授与天桥脑科学研究院的领先创新项目,积累宝贵的交叉领域工作经验。

投递简历:

aitalents@cheninstitute.org

 

02

AI驱动科学大奖申报倒计时

面向全球寻找AI for Science未来领袖

天桥脑科学研究院(Tianqiao and Chrissy Chen Institute)与《科学》杂志 “AI驱动科学大奖”,旨在表彰全球范围内用人工智能技术(如机器学习、深度学习、强化学习等)在自然科学(包括生命科学和物质科学)研究领域解决关键问题、推动科学研究加速发展的青年科学家。

申请者需提交一篇1000字的研究论文,描述他们最重要的研究发现,主题必须是利用人工智能相关技术,在其它科研领域做出了根本性的突破,而这些突破在未使用人工智能技术前是难以实现的。

该奖项目前已进入申报倒计时,报名截止时间为2024年12月13日,《科学》杂志将组成专家评审委员会评比,于2025年7月公布获奖名单。大奖(The Grand Prize)得主将获得3万美元奖金以及《科学》杂志5年数字版订阅权,获奖论文将发表在《科学》杂志的印刷版和在线版上。此外还设有最多两名优胜奖得主,各获得1万美元奖金和《科学》杂志5年的数字版订阅权,他们的论文将在《科学》杂志在线版上发表。

天桥脑科学研究院官方网站:

https://www.cheninstitute.org/prize

 

《科学》杂志官方网站(申报通道):

https://www.science.org/content/page/how-enter-chen-institute-science-prize-ai-accelerated-research

 

03

加入我们

敲开AI技术创新的大门

同时,我们持续在上海/新加坡/美国三地寻找具备技术热情、创新精神和团队合作精神的AI技术实习生/全职员工,共同致力于利用人工智能解决现实世界中的各种挑战和难题。欢迎所有人工智能方向的优秀人才投递简历:

岗位方向

后端开发与集成

  1. 设计、开发和维护支持AI模型的后端系统和服务。

  2. 集成各种机器学习模型和算法到后端系统,解决业务问题,如推荐系统、预测分析等。

  3. 优化后端系统,提高AI模型的性能和效率。

  4. 设计和实施API和微服务,以使AI功能可通过RESTful接口访问。

算法研究与开发:

  1. 探索研究最先进的大语言模型相关技术,持续跟踪论文和开源社区产品化落地的最新进展。

  2. 根据业务场景需求,提出算法方案,构建算法原型,验证算法效果。

  3. 构建高质量的算法服务,跟踪解决线上出现的相关问题。

岗位要求

  1. 对人工智能技术充满热情,有志于在该领域深入学习和发展。

  2. 具备扎实的编程基础,熟练掌握至少一种编程语言,如Python、Golang等。

  3. 对后端开发、数据库系统、Web开发、API设计等有一定了解。

  4. 对机器学习、深度学习等人工智能领域有浓厚兴趣,有相关项目经验者优先考虑。

  5. 良好的沟通能力和团队合作精神。

我们提供国际化的工作环境和广阔的个人发展空间,极具竞争力的薪酬福利。加入我们,你将站在科学与技术的交汇点,挑战最前沿的AI课题,与全球一流的科研团队共同推动AI创新。一起探索人类智能的未来!

最后,我们诚挚邀请有志于人工智能领域的青年学者前来参加本次“人工智能人才计划”高校宣讲会,深入了解项目详情及申请要求。宣讲会将为大家提供直接与项目导师和研究人员互动的机会,解答大家对于项目的疑问,并分享前沿AI和长期记忆研究的最新动态。具体时间安排如下:

城市

地点

日期

杭州

浙江大学

11月12日

南京

南京大学

11月14日

合肥

中国科学技术大学

11月19日

北京

清华大学

11月26日

北京

北京大学

11月27日

欢迎感兴趣的青年学者前来参加,期待与各位青年才俊在宣讲会上见面!

 

投递简历

aitalents@cheninstitute.org

扫描二维码添加研究院助理

 

关于天桥脑科学研究院

天桥脑科学研究院(Tianqiao and Chrissy Chen Institute, TCCl)是由陈天桥、雒芊芊夫妇出资10亿美元创建的世界最大私人脑科学研究机构之一,围绕全球化、跨学科和青年科学家三大重点,支持脑科学研究,造福人类。

TCCI与华山医院、上海市精神卫生中心设立了应用神经技术前沿实验室、人工智能与精神健康前沿实验室;与加州理工学院合作成立了TCCI加州理工神经科学研究院。

TCCI建成了支持脑科学和人工智能领域研究的生态系统,项目遍布欧美、亚洲和大洋洲,包括学术会议和交流夏校培训AI驱动科学大奖、科研型临床医生奖励计划、特殊病例社区、中文媒体追问等。

07

2024-11

概括

全球心理健康危机使得超过三分之二的受影响者未能得到治疗。人工智能技术正逐渐成为解决这一问题的变革性工具。在辛辛那提儿童医院,人工智能通过分析语言模式,以90%的准确率识别自杀倾向,并能比传统方法更早几个月检测到焦虑症状。布朗大学的弗雷德里克·佩茨施纳博士强调,人工智能具有提供个性化护理和改进治疗预测的潜力。先进的平台如Woebot和Happify利用人工智能驱动的认知疗法和正念练习来支持心理健康。然而,构建健全的数据集仍然是一个挑战,因为精神病学缺乏明确的诊断生物标志物,这限制了人工智能全面融入心理健康护理解决方案的能力。

06

2024-11

金秋十月,枫叶如火。在这个收获的季节,天桥脑科学研究院(Tianqiao and Chrissy Chen Insititue, TCCI)首届Chen Scholars Retreat,于10月17日在波士顿公共图书馆成功举办。

波士顿公共图书馆是世界上第一个向所有公民免费开放借阅的图书馆,素有“人民的宫殿”之称。同样,天桥脑科学研究院也一直提倡开放共享、造福人类。多年来天桥脑科学研究院致力于推动脑科学与人工智能的融合发展,通过搭建平台、汇聚资源、培养人才等一系列举措,为这一领域的进步贡献着自己的力量。

回顾发展历程,天桥脑科学研究院创始人雒芊芊女士感慨道:

自2016年天桥脑科学研究院创立之初,我们就怀揣着一个梦想——构建一个生态系统,赋能卓越科学家,以创造足以改写人类历史的突破性成果。我们深知,探索大脑挑战重重,但正是这份挑战,激发了我们不断前行的动力。

 

天桥脑科学研究院创始人雒芊芊女士致辞

 

01深度对话,启迪新知

此次Chen Scholars Retreat上,来自世界各地的脑科学、人工智能以及医学领域的陈氏学者齐聚一堂,共同探索脑科学与人工智能的未来。

 

天桥脑科学研究院前沿实验室成果汇报,线上接入,幻灯片演示中

 

此次会议专门设立的Chen Frontier Lab(前沿实验室)报告环节上,华山医院陈亮教授上海精神卫生中心陈剑华教授以及天桥脑科学研究院耿海洋博士,代表天桥脑科学研究院在中国建立的两间前沿实验室——应用神经技术前沿实验室人工智能和精神健康前沿实验室,向全球陈氏学者介绍了在BCI技术研发和新一代AI模型方面取得的重要成果。

陈亮教授介绍了应用神经技术前沿实验室在脑深部电刺激(DBS)的研究范式方面的创新工作,通过植入NAc和aIC双脑区,以改善情绪调节和认知功能。此外,他们与上海精神卫生中心联合开展的临床研究,两年内共为28位严重精神疾病患者植入DBS,随访观察到患者接受DBS后疾病症状改善效果明显。陈氏学者们表示期待DBS用于严重精神疾病治疗的更多成果。

陈剑华教授分享了人工智能和精神健康前沿实验室在大语言模型 (LLMs) 在精神疾病研究中的应用方面的重要进展。基于上海精神卫生中心的真实临床咨询数据,他们正在构建一个辅助提高抑郁症和焦虑症诊断准确性的大模型,项目现已收集了4500组高质量原始数据。讨论环节中,陈剑华教授与陈氏学者们友好探讨了“利用LLM评估抗抑郁药治疗效果并指导个性化用药”的研究潜力。

 

Pietro Perona,加州理工学院电子工程学Allen E. Puckett教授

左,Bernardo Sabatini,哈佛大学Kempner研究所联合主任,神经生物学Alice and Rodman W. Moorhead III教授,哈佛医学院Howard Hughes医学研究所研究员;

右,Sham Kakade,哈佛大学Kempner研究所联合主任、计算机科学与统计学Gordon McKay教授

David Anderson,生物学Howard Hughes教授,TCCI加州理工神经科学研究院首席科学家兼主任,哈佛医学院Howard Hughes医学研究所研究员

 

天桥脑科学研究院陈氏学者学术交流现场

 

在陈氏学者学术报告环节,来自哈佛大学医学院麻省总院、梅奥诊所、加州大学旧金山分校、斯坦福大学等知名研究机构的陈氏学者们,聚焦医学领域人工智能应用,分享了对脑科学与人工智能领域未来发展趋势、面临的挑战与机遇的看法。从基础理论研究到临床应用实践,从算法优化到数据分析,不乏一些具有开创性意义的研究成果,这些成果不仅推动了神经科学研究的深入发展,也为AI技术的应用拓展了新的边界。

来自哈佛大学的两位演讲嘉宾——计算机科学与统计学教授Sham Kakade与神经生物学系教授Bernardo Sabatini一同进行了汇报,他们交替分享了他们的学术研究,并发表在人工智能和神经科学领域的深刻见解和前沿思考:

Bernardo Sabatini提到,“大脑中的神经元通信是以非常嘈杂的方式进行的,这与我们训练人工系统时可以指定极高精度的节点间连接形成鲜明对比。这表明大脑的运行方式可能与我们的传统理解存在根本上的差异。”

而Sham Kakade则指出,“在训练大型基础模型时,往往需要数月的时间。随着数据量的增加,我们不希望训练算法的时间也成倍增长。因此,如何提高优化算法的运行效率是一个关键问题。”

 

天桥脑科学研究院创新项目海报展览

 

02携手前行,共创辉煌

展望未来,天桥脑科学研究院将秉持初心,致力于推动神经科学与AI的融合发展“生态系统”。加强与国际顶尖学术机构的合作与交流,搭建更加广阔的学术平台;加大人才培养力度,为神经科学与AI融合领域培养更多优秀的青年才俊;继续深化科研创新,鼓励创新成果转化;在数据共享、伦理规范、隐私保护等方面,积极探索解决方案,为神经科学与AI融合领域的健康发展保驾护航。

雒芊芊女士表示:

一直以来,我们希望通过对大脑如何感知、理解和与外界沟通的了解,促进人们的生活体验。这些年来,天桥脑科学研究院一直为此而做出很多持续不断的努力,希望能够推动脑科学与人工智能的技术发展,造福人类。

在波士顿公共图书馆这座知识的殿堂里,首届Chen Scholars Retreat圆满落幕。这场与脑科学与AI的科学盛宴,将永远留在每一位与会者的心中,激励着他们不断前行,共同创造更加辉煌的未来。

 

天桥脑科学研究院陈氏学者、与会讲者及组织成员集体留影纪念

31

2024-10

近日,来自天桥脑科学研究院(Tianqiao and Chrissy Chen Institute)的AI团队在人工智能技术领域取得了重要进展,其自研的OMNE Multiagent大模型多智能体框架登上Hugging Face发布的GAIA基准测试排行榜榜首,超越了包括微软研究院在内的众多顶尖机构提交的多智能体框架。这项成果借鉴了天桥脑科学研究院多年的大脑研究积累,通过给予Agent长期记忆的能力,让模型能够进行深度慢思考,加强了LLM对复杂问题的决策能力。

这是自去年天桥脑科学研究院创始人、前中国互联网大佬陈天桥宣布All In AI战略以来,旗下AI团队取得的一项重大成果。

 

 

OMNE目前的整体成功率为40.53%,这项结果领先于Meta、微软、Hugging Face、普林斯顿大学、香港大学、英国AI安全研究所、百川等知名机构的提交。而配备插件的GPT-4仅为15%。

GAIA是由Meta AI、Hugging Face和 AutoGPT联合推出的基准测试系统,旨在提供一个涵盖真实世界问题集的测试环境,用于全面评估AI助手的能力,包括推理、大模型多智能体处理、网页浏览和工具使用等核心功能,是当前最具挑战性的多代理智能评估数据集。OMNE框架能够在如此严苛的榜单中拔得头筹,充分体现了AI团队的技术深度与创新能力。

OMNE是一个基于长期记忆(Long Term Memory, LTM)的多智能体协作框架,每个智能体拥有相同且独立的系统结构,能够自主学习和理解完整的世界模型,从而独立理解环境。基于LTM的多智能体协同体系,使AI系统能够实时适应个体行为变化,优化任务规划与执行,推动个性化与高效的自我进化。

本次榜单的的重大突破在于融合了长期记忆的机制,通过长期记忆大幅降低了MCTS的搜索空间,提高了在复杂问题上的决策能力。通过引入更高效的逻辑推理,OMNE不仅提升了单个智能体的智能水平,还通过优化协作机制,显著增强了多智能体系统的整体能力。

这一提升机制正是受到了人类大脑皮层柱状结构研究的启发,皮质柱作为大脑认知和行为功能的基础单元,通过复杂的协作机制实现信息处理。AI模型通过强化单体智能与智能体间的协作,可能逐渐产生认知能力的涌现,构建出内部的表征模型,进而推动系统整体智能的飞跃。

“我们非常自豪地看到OMNE框架登上GAIA 榜首。”天桥脑科学研究院AI团队负责人表示,”这表明了利用LTM进行AI自我进化以及解决现实世界问题的巨大潜力。我们相信,推进长期记忆和AI自进化的研究,对于AI技术的持续发展和实际应用至关重要。”

天桥脑科学研究院由陈天桥、雒芊芊夫妇出资10亿美元创建,是世界上最大的私人脑科学研究机构之一。研究院始终围绕全球化、跨学科和青年科学家三大重点,支持AI和脑科学研究,致力于造福人类。今年以来,天桥脑科学研究院与《Science》杂志合作推出了全球AI驱动科学大奖,并举办和支持了包括”AI+精神健康”在内的各种高水平国际会议和夏校项目,致力于培养跨学科的青年AI人才。

GAIA benchmark链接:

https://huggingface.co/datasets/gaia-benchmark/results_public/viewer/2023/test?sort[column]=score&sort[direction]=desc

天桥脑科学研究院关于AI长期记忆的论文《Long Term Memory: The Foundation of AI Self-Evolution》已发表于arXiv网站:

https://arxiv.org/abs/2410.15665

24

2024-10