Artificial Intelligence #103

Hey, in this issue: how well do explanation methods for machine-learning models work?; AI that understands speech by looking as well as hearing; using sensors and machine learning to bug-proof crops; the mixture of experts in neural networks; ConvNets vs Transformers; a deep learning toolkit for speech-to-text; and more.

This issue is sponsored by SearchBlox and WeCloudData.

• [MIT News] How well do explanation methods for machine-learning models work?
• [Meta] AI that understands speech by looking as well as hearing
• FarmSense uses sensors and machine learning to bug-proof crops
• [Google] Learning to route by task for efficient inference (mixture of experts in neural networks)
• [Sponsored] PreText? NLP makes Your Content Findable – Automatically

Use SearchBlox PreText NLP to automatically generate titles, metadata, and descriptions for your web content and documents. Use SearchBlox to automate NLP for all types of documents, irrespective of where they are stored. Extract, Process and Store documents securely, all within a single platform.

• [ArXiv] ConvNets vs Transformers (paper, code)
• [Project] EasySynth - Unreal Engine plugin for easy creation of synthetic images
• [Sponsored] Become an excellent AI engineer with WeCloudData's AI Engineering bootcamp and its extensive career support
• [ArXiv] A survey of pre-trained language models based text generation
• [Project] Coqui : a deep learning toolkit for speech-to-text

The newsletter is now read by more than 550,000 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