PDF (book website): t.co/v8Imf3FJGt
Starbucks cost index (i.e., cost as the number of granda cappuccino caps):
Amazon: 27 ☕️
PDF: 0 ☕️ 🦄
Thanks to the authors for making this book available for free online! 🙏🏼
We've added Regression and Other Stories PDF to the book website t.co/BgKvE90xQv. The PDF is free to view and download for personal use only. Not for re-distribution, re-sale or use in derivative works. W/ permission from Cambridge University Press t.co/Im2Z6mzccUt.co/7bCNhrymoQ
Fourth Workshop on Image Matching: Local Features & Beyond
Image Matching Workshop #CVPR2022 Call for papers:
Topics: anything related to image matching, reloc, etc.
Deadline: March 28, 2022.
Notifications: April 4, 2022.
Submissions must contain novel work and will be indexed in IEEE Xplore/CVF.
Job Application for Research Engineer, Language & RL at OpenAI
On that note, we're building an “applied alignment” team to make sure that the models OpenAI deploys are aligned. It’s ML research, engineering, and social science. Our current job posting is for RE (t.co/Y4eCV36PZw), but if you’re interested please send me a message! ❤️
We trained InstructGPT to follow instructions, making it much more aligned with human intent.
InstructGPT, now in our API, is safer and more functional than previous models.
See example about the moon landing!
One small step for alignment research...
Our API is now powered by InstructGPT, language models trained to follow instructions rather than follow the pattern set by example (which is how GPT-3 works).
Much easier to use, and much more aligned with our vision of how safe AI models should operate: t.co/v9pmjxuhget.co/3Ykp6IDtyy
History of neural network potential methods(predict potential energy from the atomic structure by NN). Incorporating all invariances was challenging but now are solved. The remaining issues are non-local interactions and training data generation. t.co/wmjMHOCSho
This is a great read: how Amazon Prime used Rust, WebAssembly and egui to cut video latency by ~half: t.co/RYz0Dza0xa
If you want to join to the team making this happen, please get in touch. #tech…t.co/w34kWGbwkkt.co/cckdo5DgwS
OpenAI GPT-3 Text Embeddings - Really a new state-of-the-art in dense text ...
An extensive comparison of OpenAI's GPT-3 embeddings with a number of (smaller) freely-available open models on a number of NLP benchmarks.
Bottom line: don't be hypnotized by hype and size.
Blog post: t.co/RSiqJIJY6Gt.co/XcGv1LUU0p
Less than a month before the @AIESConf paper submission deadline! CFP below. All the usual (and new) topics are in scope, but we also have interaction of humans and AI as a theme.
Neural Architecture Search (NAS): basic principles and different approaches...
Neural Architecture Search (NAS) is the process of automating the design of neural networks’ topology.
In this article, we discuss its basic principles and the most popular approaches. Pytorch implementation is also included at the end.