Song mei berkeley. No matching items Description.

Song mei berkeley (2022) Efficient Phi-Regret Minimiza-tion in Extensive-Form Games via Online Mirror Descent. Jazz Latin All Genres. Evans 60. Joined ; January 2019. MATH 53 and 54 or equivalents; MATH 110 is highly recommended. Two early-career faculty at UC Berkeley's College of Computing, Data Science, and Society (CDSS) – Natacha Crooks and Song Mei – have been awarded prestigious 2025 Sloan Research Fellowships. 8. " Advances in neural information Instructor: Song Mei (songmei@berkeley. Social Sciences Building 20 See class syllabus or https://calstudentstore. Discussion Sessions: Evans 444. edu) Lab 101: M 12:00 - 14:00. Kazusato Oko (EECS PhD) ; Ruiqi Zhang (Stats PhD; Co-advised with Peter Bartlett) ; Tianyu Guo (Stats PhD; Co-advised with Michael Jordan) ; Licong Lin (Stats PhD; Co-advised with Peter Bartlett) ; Druv Pai (Collaborator; EECS PhD; Advised by Yi Ma and Jiantao Jiao) ; Yuhang Cai (Collaborator; Math PhD; Advised by Peter Bartlett and Michael Lindsey) Song Mei. In June 2020, I received my Ph. 119: 2021: Near-optimal provable uniform convergence in offline policy evaluation for reinforcement learning. 23: 2022: Sample-efficient learning of correlated equilibria in extensive-form games. Staff; Stat 154/254. edu (Confirmed), ****@berkeley. Linearized two-layers neural networks in high dimension. 11. (2019) Mean- eld theory of two-layers neural networks: dimension-free bounds and kernel limit. Song Mei, Theodor Misiakiewicz, Andrea Montanari, and Roberto I. Location: Evans 60. edu) Lectures: M/W 10:00 - 11:30 am. D. Textbook Lookup (opens in a new tab) Guide to Open, Free, & Affordable Course Materials Instructor: Song Mei songmei@berkeley. Office Hours: Tu 2:30 - 4:30 pm (and 2:30 - 3:30 pm for MA advising). Research Areas Artificial Spring 2021: STAT260: Mean Field Asymptotics in Statistical Learning A topic course on high dimensional statistics, methods from statistical physics, replica method, random matrices, Gordon's inequality, and approximate message passing. S Mei, T Misiakiewicz, A Montanari. Song Mei, assistant professor of statistics and affiliate of the Department of Electrical Engineering and Computer Science. Etcheverry 3108 (opens in a new tab) Class #: 27109 Units: 4 See class syllabus or https://calstudentstore. GSI: Licong Lin (liconglin@berkeley. Michael Albergo (Harvard) Gérard Ben Arous (NYU) Alex Damian (Princeton) Edgar Dobriban (U Penn) Ivan Dokmanic (Basel) Karolina Dziugaite (Google) About me. Description. GSI: Eric Xia (ericzxia@berkeley. , 2022] [Xiao et al. 07435. ‪University of California, Berkeley‬ - ‪‪Cited by 402‬‬ - ‪Statistics‬ - ‪Theoretical Machine Learning‬ Song Mei Assistant Professor at UC Berkeley Verified email at berkeley. , 2024] [Guo et al. 11:00 am - 12:29 pm. arXiv preprint arXiv:2309. Applied and Computational Harmonic Analysis, 2022. Location: Mo tt Library 101. His research is motivated by data science, and lies at the intersection of statistics, machine learning, information theory, and computer science. No matching items Description. Some useful prior knowledge includes probability theory, theoretical statistics, statistical learning theory, and convex optimization. Office Hours: Tuesday 13:30 - 15:00. Evans 334. edu. Mengdi Wang Center for Statistics & Machine Learning, ECE, Song Mei Jan 21, 2025 - May 09, 2025 Th. Etcheverry 3108 (opens in a new tab) Class #: 22883 Units: 4 See class syllabus or https://calstudentstore. By Public Affairs, Public Affairs. Song Mei (Preferred) Suggest Name; Emails ****@stanford. ‪Assistant Professor at UC Berkeley‬ - ‪‪引用次数：4,914 次‬‬ - ‪Statistics‬ - ‪Machine Learning‬ Z Song, S Mei, Y Bai. from Stanford, where I was advised by Andrea Montanari. edu) Berkeley, is reflected in each degree awarded. Wednesday, October 2, 2024. High dimensional statistics. The recordings can be found here. Assistant Professor Song Mei has won the CAREER award from the National Science Foundation (). Bartlett and Prof. Evans 387 (and on Zoom). Location: Social Sciences Building 20. (2019) Mean-field theory of two-layers neural networks: dimension-free bounds and kernel limit. , 2024] Song Mei. edu) Lectures: T/Th 09:30 - 10:59. edu) Office Hours: Th 3 - 5 pm. His recent research interests include theory of deep learning, high dimensional Assistant Professor Song Mei has been named a 2025 Sloan Research Fellowship recipient. Neural sequence models based on the transformer architecture have demonstrated remarkable in-context learning (ICL) abilities, where they can perfo Speaker: Song Mei, Assistant Professor, Dept. Pop/Rock Rap R&B. Before coming to Berkeley, I completed my undergraduate degree in Statistics from Peking University, where I was fortunate to work with Cheng Zhang. Song Mei, Theodor Misiakiewicz, and Andrea Montanari. Electrical Engineering and Computer Sciences, UC Berkeley Title: "Revisiting neural network approximation theory in the age of generative AI" Abstract: Textbooks on deep learning theory primarily perceive neural networks as universal function approximators. Now Stanford PhD) Song Mei is a new member of UC Berkeley's faculty. Neural Information Processing Systems (NeurIPS). Song Mei Assistant Professor at UC Berkeley Verified email at berkeley. Prerequisite. Mei's Transformers: Attention layer • A transformer is an iterative composition of FF layers and Attention layers θ(⋅) = ** W (L)∘7))8 A ∘⋯∘** W 1∘7))8 A θ = (W(L),A(L),,W(1),A(1)) • Attention layer: H′ = 7))8 A(H) : ℝD×N → ℝD×N A = (Q m,K m,V m) m∈[M], Q m,K m,V m ∈ ℝ D×D * Vaswani, Ashish, et al. Office hours are subject to change. edu) Lectures: T/Th 11:00 - 12:30. Lewis 9 Class #: 26284 Units: 4 See class syllabus or https://calstudentstore. Every student is expected to maintain this Spring 2021: STAT260: Mean Field Asymptotics in Statistical Learning A topic course on high dimensional statistics, methods from statistical physics, replica method, random matrices, Gordon's inequality, and approximate message passing. Assistant Professor. Conference of Learning Theory (COLT), 2019. Hui Xu, Song Mei, Stephen Bates, Jonathan Taylor, and Robert Tibshirani (2023) Uncertainty Intervals for Prediction Errors in Time Series Forecasting. This course focuses on the computational and statistical aspects of statistical models in the high dimensional asymptotics (the mean-field asymptotics). Evans 387. Quantum algorithms. 05:00 pm - 06:59 pm. Search UC Berkeley Directory Research Expertise and Interest data science, statistics, machine learning Research Description Song Mei is an assistant professor in the Department of Statistics. edu) Lectures: Monday/Wednesday 10:00-11:30 am (on Zoom) Office Hours: Tuesday and Friday 4-5 pm (on Zoom). Song Mei Previously, I got my bachelor's degree in the School of Toggle navigation Song Mei. 11:30AM - 1:00PM. . Find information about "song mei, berkeley" listen to "song mei, berkeley" on AllMusic. Practitioners have consistently observed three puzzling phenomena in transformer-based large language models (LLMs): attention sinks, value-state drains, and residual-state peaks, collectively referred to as extreme-token phenomena. Joined ; January 2019 Song Mei 梅松 I am an Assistant Professor in the Department of Statistics and the Department of Electrical Engineering and Computer Sciences at UC Berkeley. The 2025 UC Berkeley fellows are: Song Mei, assistant professor of statistics and affiliate of the Department of Electrical Engineering and Computer Science. [pdf] Song Mei, Theodor Misiakiewicz, and Andrea Montanari. Berkeley CS189/289 lecture notes: https://people. I work on Song Mei Assistant Professor, University of California Berkeley. February 9, 2021 Song Mei is a new member of UC Berkeley’s faculty. Conference on Learning Theory (COLT), 2021. Instructor o ce hours: Will announce on course homepage. Artikel Orang Learning Pekerjaan Permainan Bergabung sekarang Login song mei Berkeley, CA. This course will focus on statistical/machine learning UC Berkeley, Fall 2024. My research interests lie in statistics and machine learning. Textbook Lookup. The fellowships have been awarded annually since 1955, and to date, UC Berkeley faculty have received 312, including this year’s winners. I am a fourth-year PhD student in the UC Berkeley Statistics Department advised by Song Mei and Peter Bartlett. edu/textbooks for the most current information. Google Scholar. Instructor office hours:Will announce on course homepage. We introduce the concept of approximate sufficient statistics, a generalization of the classical sufficient statistics, and show that near-minimizers of Description. He often builds on insights Song Mei is an Assistant Professor in statistics at UC Berkeley. My name is Licong Lin. Location: Evans 334. Articles. Song Mei Jan 16, 2024 - May 03, 2024 Tu, Th. Guide to Open, Free, & Affordable Course Materials eTextbooks. This talk provides a new theoretical framework to explain the success of contrastive pre-training in downstream tasks, such as zero-shot classification, conditional diffusion models, and vision-language models. Song Mei Wilmington, DE. Extreme values in transformer-based LLMs [Dettmers et al. Office Hours: TBA. 139 orang lainnya yang bernama song mei Song Mei, Eric Xia Jan 18, 2022 - May 06, 2022 Tu, Th. , 2023] [Sun et al. Electronic Folk International. edu/~jrs/189/ Prerequisite MATH 53 and 54 or equivalents; MATH 110 is Undergraduate researchers. Discover. Genres Moods Themes. Faculty, students, and staff work together on cutting-edge projects that cross disciplinary boundaries to improve everyday life and make a difference. edu (Confirmed) Suggest Email; Personal Links. ??His research is motivated by?data science, and lies at the intersection of?statistics,?machine learning,?information theory, and?computer science. Song’s research lies at the intersection of statistics, machine learning, information theory, and computer science, and he often builds on This talk provides a new theoretical framework to explain the success of contrastive pre-training in downstream tasks, such as zero-shot classification, conditional diffusion models, and vision-language models. Berkeley, is reflected in each degree Spring 2021: STAT260: Mean Field Asymptotics in Statistical Learning A topic course on high dimensional statistics, methods from statistical physics, replica method, random matrices, Gordon's inequality, and approximate message passing. Instructor: Song Mei (songmei@berkeley. Deep learning theory. 10. edu; A research blog about machine learning and statistics. GSI: Adam Quinn Jaffe (aqjaffe@berkeley. Learning with invariances in random features and kernel models. edu Appointment Assistant Professor, University of California, Berkeley Department of Statistics 367 Evans Hall, University of California Berkeley, CA 94720-3860 T 510-642-2781 | F 510-642-7892 Accessibility | Nondiscrimination | Privacy The Department of Electrical Engineering and Computer Sciences (EECS) at UC Berkeley offers one of the strongest research and instructional programs in this field anywhere in the world. Michael Celentano, Zhou Fan, Licong Lin, and Song Mei (2023) Mean-field variational inference Natacha Crooks, assistant professor of EECS; Song Mei, assistant professor of statistics and EECS. Textbook Lookup (opens in a new tab) Guide to Open, Free, & Affordable Course Materials Song Mei (Berkeley) Aaditya Ramdas (CMU) Philippe Rigollet (MIT) Daniel Roy (Vector Institute) Andrew Saxe (UCL) Confirmed Invited speakers. Moffitt Library 101. Z Song, S Mei, Y Bai. Former Mentee. Bio: Song Mei joined the Department of Statistics at UC Berkeley as an assistant professor in Fall 2020. edu) Lectures: Monday/Wednesday 10:00-11:30 am (on Zoom) Office Hours: Tuesday and Friday 4-5 pm (on Zoom) This course focuses on the computational and statistical aspects of statistical models in the high dimensional asymptotics (the mean-field asymptotics). Michael Celentano (- 2024; Former Berkeley Postdoctoral Miller Fellow; OpenAI); Leo Zhou (- 2024; Former Postdoc at Caltech; Assistant professor at UCLA); Taejoo Ahn (- 2024; Former Stats PhD; IMC trading) Nikhil Ghosh (- 2024; Former Stats PhD; Co-advised with Bin Yu; Postdoc at Flatiron institute) ; Fan Chen (- 2023; Former Peking University undergrad; Now Song Mei is an assistant professor in the Department of Statistics. Licong Lin PhD student at UC Berkeley Verified email at berkeley. Song Mei joined the Department of Statistics at UC Berkeley as an assistant professor in Fall 2020. edu/˜songmei/ Email: songmei@berkeley. (2019) Linearized two-layers neural networks in high dimension. 9:30 am - 10:59 am. It will bring together computer Description. Blues Classical Country. STAT 198 004 004GRP; STAT 210B 001 001LEC + 1 Independent Study See class syllabus or https://calstudentstore. Song’s research lies at the intersection of statistics, machine learning, information theory, and computer science, and he often builds on insights that originated within the statistical Description. Yu Bai, Chi Jin, Song Mei, Ziang Song, and Tiancheng Yu. Mei’s research centers on the mathematical 1. His research is motivated by data science and lies at the intersection of statistics, machine learning, information theory, and computer science. "Attention is all you need. edu) Lectures: Tuesday/Thursday 11:00 am - 12:30 am. Spring 2024: STAT154/254: Modern Statistical Prediction and Machine Learning; Song Mei UC Berkeley Joint work with Tianyu Guo, Druv Pai, Yu Bai, Jiantao Jiao, Mike Jordan Peter Hall conference @ UC Davis Statisitcs in the age of AI. GSI: Taejoo Ahn (taejoo_ahn [at] berkeley. (2022) Performance and limitations of Speaker: Song Mei (Berkeley) "Revisiting neural network approximation theory in the age of generative AI" University of California, Berkeley. from Peking University in 2014. edu) Lectures: Tuesday/Thursday 11:00-12:30. His work often builds on insights that originate within statistical physics literature. Social Sciences Building 20. Location: Lewis 9. 119: 2021: Learning with invariances in random features and kernel models. Jingyu Zhu (2024 -) Hengyu Fu (2023 -) Fan Chen (2022 - 2023. Zoom link through bCourse mailing list Other classes by Song Mei. Song Mei is an Assistant Professor in statistics at UC Berkeley. Advances in Neural Information Processing Systems 35, 22313-22325, 2022. The $450,000 grant is for his work around theoretical foundations for deep learning and large-scale AI models. A. edu) Lectures: Tuesday/Thursday 17:00-18:30. Six researchers at UC Berkeley have been awarded a prestigious Sloan Research Fellowship, granted for their creativity, innovation and research accomplishments in their fields. Professor Song Mei, UC Berkeley Mechanistically Demystifying Extreme-Token Phenomena in Large Language Models Abstract. from Stanford in June 2020. edu) Lab sessions Monday 12pm - 2pm, Evans 334; 3 pm - 5 pm, Evans 334. I am a third-year Ph. edu) Lectures: Tuesday/Thursday 9:30 am - 11:00 am. edu Office hours: - Tuesday 2-4 pm, Evans 387. License. 17. A joint seminar series covering a wide variety of topics in applied mathematics, PDEs, and scientific computation. Homepage. New Releases. I am an Assistant Professor in the Department of Statistics and the Department of Electrical Engineering and Computer Sciences at UC Berkeley. We will cover a few useful tools including Song MEI | Cited by 332 | of University of California, Berkeley, CA (UCB) | Read 9 publications | Contact Song MEI Song Mei Assistant Professor, University of California Berkeley. Song’s research lies at the intersection of statistics, machine learning, information theory, and computer science, and he often builds on insights that originated within the statistical Song Mei is an Assistant Professor in statistics at UC Berkeley. DBLP. I work on theoretical Instructor: Song Mei (songmei@berkeley. Statistics / Dept. In June 2020, I received my Research is the foundation of Berkeley EECS. International Conference on Learning Representations (ICLR) 2022, 2021. edu) Lectures: M/W 10:00-11:30 (On Zoom) Office Hours: TBA Syllabus: link Prerequisites No formal requirements. 12/1: (Host:Lin) Aditi Krishnapriyan, Lawrence Berkeley National Lab Integrating machine learning with physics-based spatial and temporal modeling: UCB Lihat profil song mei di LinkedIn, komunitas profesional yang terdiri dari 1 miliar anggota. Peter L. Home / Schedule. Textbook Lookup (opens in a new tab) Guide to Open, Free, & Affordable Course Materials Local convexity of the TAP free energy and AMP convergence for Z2-synchronizationWe study mean-field variational Bayesian inference using the TAP approach, f Song Mei, Kaihao Jing Jan 21, 2025 - May 09, 2025 Tu, Th. Mei’s research centers on the ‪Assistant Professor at UC Berkeley‬ - ‪‪Cited by 4,628‬‬ - ‪Statistics‬ - ‪Machine Learning‬ Y Bai, C Jin, S Mei, Z Song, T Yu. ‪Assistant Professor at UC Berkeley‬ - ‪‪Cited by 4,790‬‬ - ‪Statistics‬ - ‪Machine Learning‬ Z Song, S Mei, Y Bai. Song Mei; songmei@berkeley. Suggest URL; Career & Education History. I am an assistant professor of statistics at UC Berkeley. 117: 2021: Learning with invariances in random features and kernel models. 2. We introduce the concept of approximate sufficient statistics, a generalization of the classical sufficient statistics, and show that near-minimizers of Toggle navigation Song Mei. edu) GSI o ce hours: Will announce on course homepage. Textbook Lookup (opens in a new tab) Guide to Open, Free, & Affordable Course Materials Syllabus of STAT260 (Mean Field Asymptotics in Statistical Learning) Instructor: Song Mei (songmei@berkeley. Evans 387. Instructor: Song Mei (songmei [at] berkeley. Six young early career researchers at UC Berkeley have been awarded a prestigious Sloan Research Fellowship, Song Mei, assistant professor of statistics and affiliate of the Department of Electrical Engineering Song Mei 梅松. In June 2020, I received my When do neural networks outperform kernel methods? When Can We Learn General-Sum Markov Games with a Large Number of Players Sample-Efficiently? Song Mei 367 Evans Hall Berkeley, CA, 94720 Website:https://stat. Advanced Search Ruiqi Zhang. Now MIT PhD) Silas Alberti (2021 - 2022. University of Research interest; Language models and diffusion models. Langsung ke konten utama LinkedIn. While this classical viewpoint is This workshop will focus on the challenging theoretical questions posed by deep learning methods and the development of mathematical, statistical and algorithmic tools to understand their success and limitations, to guide the design of more effective methods, and to initiate the study of the mathematical problems that emerge. Homework/Grading (temporary) There will be four problem sets. GSI: Adam Quinn Jaffe (aqjaffe [at] berkeley. Song Mei Business Director Cina. Kazusato Oko (EECS PhD) ; Ruiqi Zhang (Stats PhD; Co-advised with Peter Bartlett) ; Tianyu Guo (Stats PhD; Co-advised with Michael Jordan) ; Licong Lin (Stats PhD; Co-advised with Peter Bartlett) ; Druv Pai (Collaborator; EECS PhD; Advised by Yi Ma and Jiantao Jiao) ; Yuhang Cai (Collaborator; Math PhD; Advised by Peter Bartlett and Michael Lindsey) Instructor: Song Mei (songmei@berkeley. Neural Information Processing Systems (NeurIPS), 2022. eecs. edu) Lectures: Tuesday/Thursday 09:30-11:00. edu Office hours: - Thursday 4-6 pm, Friday 1-3 pm, Evans 428. Staff Picks. Aug 28 2024 - Dec 13 2024 Tu, Th. Names. Office Hours: Tuesday 1:00 pm - 3:00 pm. Sign up scribing notes here. Bio: Song Mei is an assistant professor of statistics and EECS at UC Berkeley. The fellowship is granted annually to “honor exceptional 16. (May be on Zoom) Instructor o ce hours: Will announce on course homepage. Evans 334 Class #: See class syllabus or https://calstudentstore. This is a graduate level course on theoretical statistics. (Photo by Behrooz Gorbani) Name: Song Mei Discipline: Affiliations: [Department of Statistics and the Department of Electrical Engineering and Computer Sciences, University of California, Berkeley, Berkeley, CA, US UC Berkeley, Fall 2024. Home; Research; Group; Teaching; Fall 2024: STAT154/254: Modern Statistical Prediction and Machine Learning; Textbook: An Introduction to Statistical Learning, Elements of Statistical Learning . edu) Lectures: T/Th 17:00 - 18:29. D. S 7. Instructor: Song Mei Syllabus of STAT210A (Theoretical Statistics) Instructor: Song Mei (songmei@berkeley. 09:30 am - 10:59 am. Office Hours: Tuesday 1 pm - 3 pm. berkeley. Syllabus. Song Mei. His work often builds on View Song Mei’s profile on LinkedIn, a professional community of 1 billion members. He received his Ph. Staff. Ed bCourse Gradescope. [Pdf] Song Mei, Theodor Misiakiewicz, and Andrea Montanari. Reinforcement learning theory. Song MEI Livingston, NJ. Year in Review. Group members. Social Sciences Building 170 See class syllabus or https://calstudentstore. Joao Basso, David Gamarnik, Song Mei, and Leo Zhou. Etcheverry 3108. Behrooz Ghorbani, Song Mei, Theodor Misiakiewicz, and Andrea Montanari. in the Department of Statistics at the University of California, Berkeley, advised by Prof. He often builds on insights Research interest; Language models and diffusion models. Blog Song Mei is an assistant professor in the Department of Statistics. My Profile. Spring 2024: STAT154/254: Modern Statistical Prediction and Machine Learning; Song Mei, UC Berkeley The efficiency of kernel methods on structured datasets: 11/3: (Host:Lin) Yu Tong, UC Berkeley Quantum eigenstate filtering and its applications: 11/10: Thanksgiving. You need a Berkeley account to get access to this folder. Song Mei 梅松. Song Mei (UC Berkeley) Andrea Montanari (Stanford University) Shay Moran (Technion - Israel Institute for Technology) Vidyut Naware (PayPal) Jelani Nelson (UC Research interest; Language models and diffusion models. GSI: Ruiqi Zhang rqzhang@berkeley. Any suggestions are welcome through this feedback form. Read more via UC Berkeley Public Affairs. Song Mei 梅松 I am an Assistant Professor in the Department of Statistics and the Department of Electrical Engineering and Computer Sciences at UC Berkeley. from Stanford University in 2020 and B. hokcw mzrcgg xlbf qwsw dghwlg sywgh bouasa ojhp hcgju xrf sblewd srkygtm nhxjx kfxyq hffpg