Artificial Intelligence #145

Hey, in this issue: who is working to end the threat of AI-generated deepfakes, and why is it so difficult?; reincarnating reinforcement learning; synthetic data can offer real performance improvements; generating human-level text with contrastive search in transformers; a tutorial on building an end-2-end active learning pipeline; and more.

This issue is sponsored by AtScale and Data Science Salon.

• Who is working to end the threat of AI-generated deepfakes, and why is it so difficult?
• [Google] Beyond tabula rasa: reincarnating reinforcement learning (code and more)
• [MIT] In machine learning, synthetic data can offer real performance improvements
• [Google] ReAct: synergizing reasoning and acting in language models
• [Sponsored] Practical panel with data leaders from UPS, efood, & Saks

Join this webinar panel to get practical advice from best-in-class data orgs. Learn democratization initiatives to scale data analytics access and business intelligence across your company. Register now to save your spot.

• [HuggingFace] Generating human-level text with contrastive search in transformers
• New Go-playing trick defeats world-class Go AI—but loses to human amateurs
• [Sponsored] Join Data Science Salon Austin on Feb 21-22 for 35+ sessions and panels covering AI and ML applications in the enterprise!
• Study assesses the quality of AI literary translations by comparing them with human translations
• [Tutorial] Build an end-2-end active learning pipeline

More than 600,000 subscribers are reading this newsletter. If you are building an AI or a data product or service, you can become a sponsor of one of the future newsletter issues and join the list of?these great companies. Feel free to reach out to?[email protected]?for more details on sponsorships.

Enjoy the newsletter? Help us make it bigger and better by sharing it with your colleagues and friends.

Have a nice weekend! See you next week. — Andriy