Stanford AI Lab
The Stanford Artificial Intelligence Laboratory (SAIL) has been a center of excellence for Artificial Intelligence research, teaching, theory, and practice since its founding in 1962.
Latest News
Chirpy Cardinal SAIL Student Team Gets 2nd Place in Alexa Prize Socialbot Grand Challenge 4
Congratulations to Chirpy Cardinal! The team, led by Ethan Chi and advised by Chris Manning got 2nd place for the 2nd year running.
Professor Ron Dror and Students Raphael Townshend and Stephan Eismann Featured on the Front Cover of Science
Congratulations to Professor Ron Dror and students Raphael Townshend and Stephan Eismann who were featured on the front cover of Science for their paper “Geometric deep learning of RNA structure”.
AI4ALL Highlighted in The Stanford Daily
The Stanford Daily reports on how Stanford’s AI4ALL program mentors a new generation of diverse leaders in artificial intelligence.
Aditya Grover Wins SIGKDD Dissertation Award
Congratulations to Professor Stefano Ermon advisee Aditya Grover for winning the SIGKDD Dissertation Award!
Dorsa Sadigh Receives MIT Technology Review TR-35 Award
Congratulations to SAIL faculty Dorsa Sadigh on receiving an MIT Tech Review TR-35 award for her work on teaching robots to be better collaborators with people.
Anna Goldie Receives MIT Technology Review TR-35 Award
Congratulations to SAIL PhD student Anna Goldie on receiving an MIT Tech Review TR-35 award for her work on AI layout of computer chips with deep reinforcement learning.
Latest Tweets
"While selective classification improves average accuracy for this task, we find that it does not appreciably improve accuracy on the most clinically relevant subgroup, or subpopulation, of the data..." https://twitter.com/StanfordAILab/status/1448685074115747842
Selective classifiers, which can avoid making uncertain predictions, routinely improve *average* accuracy. But a new post shows they can in fact alarmingly hurt accuracy on data *subpopulations*, and suggests reasons for why this occurs + proposes fixes!
https://ai.stanford.edu/blog/sc-magnifies-disparities/
Our new chat with @SnorkelAI, which started as a research project in the @StanfordAILab.
- How Snorkel Flow helps ML technologists?
- What ar the near-term breakthroughs for automated data labeling?
Find answers here⬇️
https://thesequence.substack.com/p/paroma
Check out our interview with @StanfordAILab professor @chelseabfinn!
We talk about enabling robots and other agents to develop broadly intelligent behavior through learning and interaction, and more! 🤖
Listen to it here
👇
https://thegradientpub.substack.com/p/chelsea-finn-on-meta-learning-and?justPublished=true
Chelsea Finn on Meta Learning & Model Based Reinforcement Learning
An interview Stanford Professor Chelsea Finn, whose research deals with the capability of robots and other agents to...
thegradientpub.substack.com
Selective classifiers, which can avoid making uncertain predictions, routinely improve *average* accuracy. But a new post shows they can in fact alarmingly hurt accuracy on data *subpopulations*, and suggests reasons for why this occurs + proposes fixes!
https://ai.stanford.edu/blog/sc-magnifies-disparities/
@tengyuma @StanfordAILab Nice analysis. We also show that learning from unbalanced data is not necessarily worse than learning from balanced data in the SSL setting: https://arxiv.org/pdf/2103.14005.pdf
Thinking of applying self-supervised learning (SSL) on your uncurated, imbalanced datasets? Good news: we found SSL is more robust to long tails than supervised representations. We also present theoretical and empirical analyses and an improved algorithm. http://arxiv.org/abs/2110.05025
We Are Pleased to Welcome a New Member of Our Faculty
Carlos Guestrin who focuses on Machine Learning, Explainability, Fairness and ML Systems.
Affiliates Program
Stanford AI Lab faculty and students enjoy chances to understand and solve the not-yet-doable pain points of industry. Get a chance to support and interact with SAIL’s brightest minds.
Learn More