Description
The 10th Annual Simons Institute Industry Day is Thursday, November 7, 2024, at 9:00 a.m. PST in Calvin Lab Auditorium.
Industry Day brings the Institute’s industry and institutional partners together with our visiting graduate students, postdoctoral research fellows, and senior researchers, as well as scientists from the broader Berkeley campus.
This year’s event will include presentations by our current partners — ADIA Lab, Apple, Google, IVADO, Jane Street, NTT Research, Oski Technology/Nvidia, Roc360, and VMWare — lightning talks by research fellows, and highlights from our current research programs — LLMs and Transformers Part 1, and Modern Paradigms in Generalization. A reception and dinner will conclude the day.
Participation is open to the Simons Institute's Industry Partners and Sponsors, organizers, research fellows and scientists in Institute programs, and members of the Berkeley community.
Learn more about our Industry Partnerships program here, or contact Senior Development Director Amy Ambrose for further information at amyambrose@berkeley.edu or +1 510 944 6674.
Thursday, November 7, 2024
Schedule subject to change
| 9:00-9:30 a.m. | Breakfast and Check-In | |
| 9:30-10:30 a.m. | Welcome and Program Highlights - Venkat Guruswami Quantum Algorithms, Complexity, and Fault Tolerance: Umesh Vazirani
Error Correcting Codes-Theory and Practice: Umesh Vazirani
LLMs and Transformers: Matus Telgarsky/Po-Ling Loh
Modern Paradigms in Generalization: Matus Telgarsky/Po-Ling Loh
Cryptography 10 Years Later: Obfuscation, Proof Systems, and Secure Computation: Shafi Goldwasser
Algorithmic Foundations for Emerging Computing Technologies: Kirk Pruhs | |
| 10:30-10:50 a.m. | Networking Break | |
| 10:50 a.m.-12:00 p.m. | Research Fellows Lightning Talks - led by Nikhil Srivastava Theoretical Implications of Training and Sampling Diffusion Models: Yuqing Wang
Subgaussian Distributions are Certifiably So: Ankit Pensia
Context-Scaling versus Task-Scaling in In-Context Learning: Amirhesam Abedsoltan
Instance-Adaptivity in Private Estimation: Lydia Zakynthinou
A Multi-Power Law for Loss Curve Prediction Across Learning Rate Schedules: Kaifeng Lyu
A Unified Framework for Efficient Learning at Scale (Pretraining and Finetuning): Soufiane Hayou
Unstable versus Stable Convergence: Benefits of Large Stepsizes in Gradient Descent: Jingfeng Wu
Capabilities and Limitations of Transformers in Sequential Reasoning: Bingbin Liu
Distilling Models Using the Linear Representation Hypothesis: Enric Boix | |
| 12:00-1:30 p.m. | Lunch | |
| 1:30-3:30 p.m. | Industry Partner Talks - led by Sampath Kannan Generative AI Products for Real Estate: Takamitsu Tanaka, Managing Director and Head of Data and AI, Roc360
Provable Uncertainty Decomposition via Higher-Order Calibration: Aravind Gollakota, Research Scientist, Apple
Multiparty Reach and Frequency Histogram: Private, Secure, and Practical: Badih Ghazi, Research Scientist, Google
An Interdisciplinary Consortium for Research, Training, and Knowledge Mobilization in Robust, Reasoning, and Responsible AI: Foutse Khohm, Vice President, IVADO
How Deep Learning in Finance is Just Like Training LLMs...and How it's Not: David So, Quantitative Researcher, Jane Street
Formal Verification at NVIDIA : Satya Holla, Research Scientist, Oski Technology/Nvidia ADIA Lab – an introduction: Horst Simon, Director, ADIA Lab | |
| 3:30-4:00 p.m. | Break | |
| 4:00-5:00 p.m. | Special Sessions Related to Current Programs - led by Venkat Guruswami Secure RAG- Power and Safegaurding AI Innovation: Vaish Krisnamurthy, VP Engineering, Box
Digital Twin for Quality Control: Mareike Kritzler, Innovation Strategy Manager, Siemens
Research Challenges at Enterprise AI: A ServiceNow Perspective: DJ Dvijotham, Sr Staff Researcher: Safety and Security, ServiceNow | |
| 5:00-5:15 p.m. | Closing Remarks by Venkat Guruswami | |
| 5:15-6:00 p.m. | Pre-Dinner Reception at Faculty Club | |
| 6:00-7:30 p.m. | Dinner at Faculty Club |
All scheduled dates:
Upcoming
No Upcoming activities yet