The TWIML AI Podcast (formerly This Week in Machine Learning & Artificial Intelligence) Fighting Fake News and Deep Fakes with Machine Learning w/ Delip Rao - TWiML Talk #260
May 3, 2019
Delip Rao, Vice President of Research at the AI Foundation and co-author of Natural Language Processing with PyTorch, dives into the complex world of fake news and deep fakes. He discusses innovative strategies for detecting artificial content across various media, including text and audio. Rao highlights the role of GANs in both creating and combating misinformation. He also shares insights from his journey in AI, emphasizing the collaboration needed among researchers and journalists to tackle these pressing challenges.
AI Snips
Chapters
Books
Transcript
Episode notes
Path to AI
- Delip Rao's initial interest was in distributed computing, but a professor's sabbatical led him to AI.
- Reading books by Nils Nilsson and Peter Norvig sparked his passion for AI and later NLP.
Academia vs. Reality
- At Twitter, Delip Rao learned that real-world products require more than just good models.
- Simple rule-based systems sometimes outperform complex models due to cost and engineering overhead.
AI Foundation's Approach
- The AI Foundation uses a for-profit/non-profit model to tackle synthetic content.
- The for-profit arm focuses on generation, while the non-profit focuses on detection, creating a synergistic relationship.
