Artificial Intelligence #112

Hey, in this issue: the state of AI in 9 charts; a new PyTorch library to train models faster; restoring and attributing ancient texts; new research on deep learning for tabular data; a survey on efficient transformers; injecting fairness into machine-learning models; bootstrapping labels via supervision and human-in-the-loop; and more.

This issue is sponsored by Scale and Run:AI.

• [Stanford] The state of AI in 9 charts (and the entire 2022 AI Index Report)
• AI critical to business success in post-pandemic world, study finds
• [Project] Composer: a new PyTorch library to train models ~2-4x faster with better algorithms
• [Project] Restoring and attributing ancient texts using deep neural networks (paper, code, demo, video)
• [Sponsored] Leveraging Data Science – From Fintech to TradFi

From startups to traditional financial institutions, what are the biggest challenges in fintech and where is data science most impactful? Don't miss VP of Product and Data Science at FIS, Christine Hurtubise with Jaclyn Rice Nelson, co-founder of Tribe AI. [Thursday, March 24th, 2022 at 3PM ET / 12PM PT]

• [ArXiv] On embeddings for numerical features in tabular deep learning (code)
• [ArXiv] Efficient transformers: a survey
• [Sponsored] Guide to Inference Workloads in the Cloud: How to squeeze more out of cloud resources, meet SLAs & reduce compute costs.
• [MIT] Injecting fairness into machine-learning models
• [The Gradient] Bootstrapping labels via supervision and human-in-the-loop

The newsletter is now read by more than 570,000 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