The Future of Decentralization, AI, and Computing Summit, hosted by Berkeley RDI, aims to bring together researchers, innovators, thought leaders, and builders to showcase and discuss the future of decentralization, AI, and computing.

In particular, the event will include topics on how AI/ML and decentralization technology can help each other, and how to build a full decentralized, open source stack for AI/ML, including 1) decentralized ML infra such as decentralized training, inference, provenance, integrity guarantees, 2) open source data, model, and tooling, 3) personalized AI with privacy and trustworthiness; 4) democratic and decentralized process for AI governance and alignment. We hope this summit will provide an opportunity for researchers from diverse backgrounds including AI/ML, Security & Privacy, game theory, economics, decentralization technology, to get together and cross pollinate to help seed the initial discussions in this important space and make advancement in responsible AI with decentralization.

Date:   August 27, 2023

Venue:   Pauley Ballroom, Martin Luther King Jr. Building, UC Berkeley

Program Chair

Dawn Song
Professor, UC Berkeley
Co-director of Berkeley Center for Responsible,
Decentralized Intelligence

Keynote Speakers

Shafi Goldwasser Michael Jordan
Director Pehong Chen
Distinguished Professor
Simons Institute, UC Berkeley UC Berkeley

Program Schedule

8:45am Door open; refreshment; coffee
9:15am Opening remark: Dawn Song (Professor, UC Berkeley)   [Slides] [Video]
Session I: Open source LLMs
9:30-10:30am Taming the Wild West of LLMs   [Slides] [Video]
Nazneen Rajani (Research Lead, Hugging Face)
Building and Studying Instruction-following Models   [Slides] [Video]
Tatsunori Hashimoto (Assistant Professor, Stanford)
DecodingTrust: Assessing Trustworthiness and Risks of Generative Models   [Slides] [Video]
Bo Li (Associate Professor, University of Chicago)
Rethinking LLM Evaluation   [Slides] [Video]
Ion Stoica (Professor, UC Berkeley)
Panel: Open Source vs. Closed Source: Will Open Source win? Challenges & Future in Open Source LLM   [Video]
Moderator: Joseph Gonzalez (Associate Professor, UC Berkeley)
Session II: Decentralized and Distributed Machine Learning Infrastructure
10:30-11:20am Open Dataset and Decentralized Compute for LLMs   [Slides] [Video]
Tri Dao (Assistant Professor, Princeton)
Megatron-LM   [Slides] [Video]
Bryan Catanzaro (VP of Applied Deep Learning Research, NVIDIA)
Federated Learning of Gboard Language Models with Differential Privacy   [Slides] [Video]
Zheng Xu (Research Scientist, Google Research)
Panel (20 minutes): Future of Decentralized and Distributed ML Infrastructure: Challenges & Opportunities   [Video]
Moderator: Dawn Song (Professor, UC Berkeley)
11:20-11:40am Break
Session III: Cryptography & ML; Personalization
11:40am-12:30pm Keynote: Constructing Trust (or confirming Distrust) using Cryptographic Paradigms   [Slides] [Video]
Shafi Goldwasser (Director of the Simons Institute for the Theory of Computing, Professor; UC Berkeley)
Personal Digital Data: Challenges and Opportunities   [Slides] [Video]
Alon Halevy (Director, Meta)
Lightning talks [Video]
12:30-1:30pm Lunch
Session IV: Open-source LLMs tools & ecosystem
1:30-2:15pm LangChain for LLM Application Development   [Slides] [Video]
Harrison Chase (CEO, Langchain)
From RAG to Agents: Building LLM Apps over Your Data with LlamaIndex   [Slides] [Video]
Jerry Liu (CEO, LlamaIndex)
ReplitLM: Using Open-source from Training to Production for a Code Completion LLM   [Slides] [Video]
Michele Catasta (VP of AI, Replit)
Panel (20 minutes): Future of Open Source LLM tooling & ecosystem; how to build successful Open Source LLM ecosystem   [Video]
Moderator: Matei Zaharia (Associate Professor, UC Berkeley)
Session V: Multi-agent Systems & Economics
2:15-3:00pm Learning to Cooperate and Compete via Self Play   [Slides] [Video]
Noam Brown (Research Scientist, OpenAI)
Keynote: An Alternative View on AI: Collaborative Learning, Incentives, and Social Welfare   [Slides] [Video]
Mike Jordan (Professor, UC Berkeley)
3:00-3:20pm Break
Session VI: Future Directions in AI + Decentralization
3:20-4:30pm Part I: AI/ML infrastructure & Services with Decentralized Trust [Video]
    Pramod Viswanath (Professor, Princeton)
    Chaoyang He (Co-founder and CTO, FedML)
    Ryan Cao (CTO, Modulus Lab)
    Ben Fielding (Co-founder, Gensyn)
Part II: Cooperative AI with Decentralized Trust [Video]
    Gordon Liao (Chief Economist, Circle)
    Matheus Venturyne Xavier Ferreira* (Assistant Professor, University of Virginia)
    Matt Stephenson* (Head of Cryptoeconomics, Pantera Capital)
    Xinyuan Sun* (Flashbots)
Part III: AI for Decentralization & Security [Video]
    Arthur Gervais (Associate Professor, UCL)
    Victor Fang (CEO, AnChain.AI)
Lightning Talks, Startup Spotlights, Xcelerator Demo Day, and Reception
4:30pm Lightning Talks [video] and Startup Spotlights [video]

The summit is hosted by Berkeley RDI, a multi-disciplinary campus-wide center, focusing on advancing the science, technology and education of web3, decentralization and empowering of a responsible digital economy.

Join RDI Telegram for announcements and mailing list for all RDI news!

Following the Summit, join and hear from Top Innovators via the Xcelerator Demo Day.

For Ecosystems and Projects, Apply to Recruit Talent & Meet Devs throughout the day.

Home | Research | People | Academics | Events | Metaverse