Artificial Intelligence #118

Hey, in this issue: Meta is releasing a 175B parameter language model; a library to inspect intermediate layers of PyTorch models; how to use massive AI models in your startup; an easier way to teach robots new skills; attacks that involve poisoning training data; generative models for drug discovery; and more.

This issue is sponsored by Tecton and SearchBlox.

• [Meta] is releasing a 175B parameter language model
• [Project] surgeon-pytorch – a small library to inspect intermediate layers of PyTorch models
• How to use massive AI models in your startup
• [MIT] An easier way to teach robots new skills
• [Sponsored] Free registration for apply(): the conference for ML data engineering

Join us at apply(), the free, online event for ML data engineering. Users and thought leaders share their experiences and best practices on ML data engineering. Speakers include practitioners from Stripe, Uber, Facebook, Instacart, Snowflake, Databricks, and more. Register for free at applyconf.com.

• The risks of attacks that involve poisoning training data for machine learning models
• [DeepMind] Specification gaming: the flip side of AI ingenuity
• [Sponsored] Using Elasticsearch for your NLP solution? Let’s show you what SearchBlox can do using PreText and Hugging Face.
• Compact word vectors with Bloom embeddings
• [Microsoft] MoLeR: generative models for drug discovery

The newsletter is now read by more than 580,000 subscribers. If you are building an AI or a data product or service, you are invited to join the list of these great companies by becoming a sponsor of one of the future newsletter issues. 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