BROADCASTS.com

Ishan Misra is a research scientist at FAIR working on self-supervised visual learning. Please support this podcast by checking out our sponsors:
\n- Onnit: https://lexfridman.com/onnit to get up to 10% off
\n- The Information: https://theinformation.com/lex to get 75% off first month
\n- Grammarly: https://grammarly.com/lex to get 20% off premium
\n- Athletic Greens: https://athleticgreens.com/lex and use code LEX to get 1 month of fish oil
\n
\nEPISODE LINKS:
\nIshan's twitter: https://twitter.com/imisra_
\nIshan's website: https://imisra.github.io
\nIshan's FAIR page: https://ai.facebook.com/people/ishan-misra/
\n
\nPODCAST INFO:
\nPodcast website: https://lexfridman.com/podcast
\nApple Podcasts: https://apple.co/2lwqZIr
\nSpotify: https://spoti.fi/2nEwCF8
\nRSS: https://lexfridman.com/feed/podcast/
\nYouTube Full Episodes: https://youtube.com/lexfridman
\nYouTube Clips: https://youtube.com/lexclips
\n
\nSUPPORT & CONNECT:
\n- Check out the sponsors above, it's the best way to support this podcast
\n- Support on Patreon: https://www.patreon.com/lexfridman
\n- Twitter: https://twitter.com/lexfridman
\n- Instagram: https://www.instagram.com/lexfridman
\n- LinkedIn: https://www.linkedin.com/in/lexfridman
\n- Facebook: https://www.facebook.com/lexfridman
\n- Medium: https://medium.com/@lexfridman
\n
\nOUTLINE:
\nHere's the timestamps for the episode. On some podcast players you should be able to click the timestamp to jump to that time.
\n(00:00) - Introduction
\n(07:49) - Self-supervised learning
\n(16:24) - Self-supervised learning is the dark matter of intelligence
\n(20:17) - Categorization
\n(28:50) - Is computer vision still really hard?
\n(32:35) - Understanding Language
\n(42:14) - Harder to solve: vision or language
\n(48:59) - Contrastive learning & energy-based models
\n(52:59) - Data augmentation
\n(57:19) - Fixed audio spike by lowering sound with pen tool
\n(1:05:33) - Real data vs. augmented data
\n(1:09:16) - Non-contrastive learning energy based self supervised learning methods
\n(1:12:54) - Unsupervised learning (SwAV)
\n(1:15:37) - Self-supervised Pretraining (SEER)
\n(1:20:44) - Self-supervised learning (SSL) architectures
\n(1:26:43) - VISSL pytorch-based SSL library
\n(1:29:38) - Multi-modal
\n(1:37:06) - Active learning
\n(1:42:45) - Autonomous driving
\n(1:54:12) - Limits of deep learning
\n(1:58:19) - Difference between learning and reasoning
\n(2:03:26) - Building super-human AI
\n(2:11:14) - Most beautiful idea in self-supervised learning
\n(2:15:02) - Simulation for training AI
\n(2:18:27) - Video games replacing reality
\n(2:19:40) - How to write a good research paper
\n(2:24:08) - Best programming language for beginners
\n(2:25:01) - PyTorch vs TensorFlow
\n(2:28:26) - Advice for getting into machine learning
\n(2:30:31) - Advice for young people
\n(2:32:58) - Meaning of life