Artificial Intelligence #133

Hey, in this issue: on using natural language in robotics; an 8-bit inference method that saves 2x memory and does not degrade performance for 175B models; a library for representing Python programs as graphs for machine learning; are designers even needed anymore?; tutorials on accelerating BERT inference and on supervised clustering; GPT language model spells out new proteins; and more.

This issue is sponsored by AtScale and AI Chambers .

• [Google] Towards helpful robots: grounding language in robotic affordances
• LLM.int8(): an 8-bit inference method that saves 2x memory and does not degrade performance for 175B models?(paper , tutorial )
• A library for representing Python programs as graphs for machine learning (report , code , package )
• Machine learning: are designers even needed anymore ?
• [Sponsored] New semantic layer whitepapers

Check out this bundle of Semantic Layer whitepapers by best-selling authors - download here .?You'll learn the key value propositions to implement a semantic layer and best practices for analytics success with one.

• [Tutorial] Accelerate BERT inference with DeepSpeed-inference on GPUs
• Deep learning alone isn’t getting us to human-like AI
• [Sponsored] AI Chambers – Ask an experienced AI Patent Attorney via 1-on-1 Zoom meeting, powered by Ridout & Maybee. Meet monthly.
• [IEEE] GPT language model spells out new proteins
• [Tutorial] Supervised clustering: how to use SHAP values for better cluster analysis

More than 600,000 subscribers are reading the newsletter. 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