We're sunsetting PodQuest on 2025-07-28. Thank you for your support!
Export Podcast Subscriptions
cover of episode Building Connections Through Open Research: Meta’s Joelle Pineau

Building Connections Through Open Research: Meta’s Joelle Pineau

2024/6/25
logo of podcast Me, Myself, and AI

Me, Myself, and AI
AI Deep Dive AI Chapters Transcript
People
J
Joelle Pineau
Topics
Joelle Pineau：Meta致力于开放式AI研究，认为开放有助于提高研究质量和责任感，促进学术界和工业界的合作，并加快AI技术的进步。开放并不妨碍Meta将研究成果应用于产品，因为产品的成功取决于用户，而非秘密技术。 Joelle Pineau：Meta的AI研究涵盖图像和视频理解、基础模型、语言模型、机器人技术以及核心智能原则等多个领域。她强调了开放研究对吸引和留住优秀研究人员的重要性，并指出开放有助于发现模型中的缺陷，并通过社区贡献改进模型。 Joelle Pineau：学术界和工业界在AI研究方面各有优势，学术界更注重跨学科合作和更广泛的研究问题，而工业界拥有更强大的计算资源。她认为，一个充满活力的生态系统对所有人都有利。 Joelle Pineau：她认为目前AI模型的主要风险在于其可能存在的偏差和对不同群体的潜在危害，而不是所谓的“存在性威胁”。她详细阐述了AI模型偏差的来源，包括有偏差的数据集和模型本身的特性，以及模型部署方式。她强调了进行更严格的分析以避免增强偏差，并理想地利用AI构建更公平公正的社会的重要性。 Joelle Pineau：她还讨论了Meta如何通过开放研究模型来加快改进过程，并分享了在碳捕获等科学发现项目中的经验，以及在数据中心能源效率和元宇宙相关硬件设计中应用AI技术的案例。 Joelle Pineau：她认为AI最大的机遇在于跨语言沟通，这对于促进全球人民之间的理解至关重要。她还指出，人们对AI的最大误解之一是将其视为黑盒，而实际上，我们对AI系统内部的工作原理有相当程度的了解。她还表达了对AI系统之间相互理解和建立社会联系的希望。

Deep Dive

Chapters
Joelle Pineau discusses the benefits of open AI research at Meta, emphasizing the high quality and responsibility it promotes, and how it aids in recruiting top researchers and contributing to the broader AI community.

Shownotes Transcript

Joelle Pineau’s curiosity led her to pursue a doctorate in engineering with a focus on robotics, which she describes as her “gateway into AI.” As vice president of AI research at Meta, Joelle leads a team committed to openness in the service of high-quality research, responsible AI development, and community contribution.

In this episode, Joelle, who is also a professor at McGill University, weighs the advantages industry and academia each have for conducting artificial intelligence research. She also describes specific AI research projects Meta is working on, including scientific discovery initiatives focused on addressing societal problems like carbon capture. Read the episode transcript here).

Guest bio:

Joelle Pineau is vice president of AI research at Meta and a professor at McGill University. Her research focuses primarily on developing new models and algorithms for planning and learning in complex, partially observable domains. She also applies these algorithms to robotics, health care, games, and conversational agents. Pineau serves on the board of the <cite>Journal of Artificial Intelligence Research</cite> and the <cite>Journal of Machine Learning Research</cite>. She has a bachelor’s degree in engineering from the University of Waterloo and master’s degree and doctorate in robotics from Carnegie Mellon University.

Me, Myself, and AI is a collaborative podcast from MIT Sloan Management Review and Boston Consulting Group and is hosted by Sam Ransbotham and Shervin Khodabandeh. Our engineer is David Lishansky, and the coordinating producers are Allison Ryder and Andy Goffin.

Stay in touch with us by joining our LinkedIn group, AI for Leaders at mitsmr.com/AIforLeaders) or by following Me, Myself, and AI on LinkedIn).

We encourage you to rate and review our show. Your comments may be used in *Me, Myself, and AI *materials.