#ICLR2023 reviews are out… and seem incredibly brutal for a couple of papers that I really like! (by others, not by me) Good luck and happy weekend to everyone…
RESEARCH
-
Wikipedia and SCP Foundation: Internet’s Most Successful Social Networks
By
–
The only large, long-term, productive social network to come out of the first 40 years of Internet may be Wikipedia. And possibly the SCP Foundation shared story universe.
-
Living Through History: The AI Revolution Unfolds Daily
By
–
when you are living through history, it feels like just another day
-
Geoengineering and AI: A Solution for Climate Warming
By
–
A new UN Climate Change report estimates we might be on track to 2.5°C of warming. It's time to seriously consider cooling down Earth by geoengineering using stratospheric aerosol injection, a solution in which AI will play a critical role. See why.
-
Improving Fine-Tuning Capabilities for Better AI Customization
By
–
still have a lot to figure out, but we definitely want to let people do more and better fine-tuning
-
Investment and Machine Learning Specific Sessions Overview
By
–
no–this is an investment and some ML-specific sessions, not an accelerator-like thing
-
Rodney Brooks Compiles Archive of IEEE Spectrum Articles
By
–
I have collected links to everything I have written for @IEEESpectrum over the last decade and more. Here at my blog: https://
rodneybrooks.com/my-ieee-spectr
um-columns-and-articles/
… -
Memepocalypse: exponential meme evolution and cultural implications
By
–
In the end, we will be outnumbered by our own memes. Meme types are doubling every 6 months. And they are evolving more visual complexity over time. The memepocalypse is inevitable. https://
nature.com/articles/s4159
8-021-99468-6
… -
Research on AI Example Benefits: Output Syntax and Label Space Learning
By
–
I think what Min et al. shows is that much of the benefit of examples is learning the output syntax and label space. Results aren’t as clean as this anecdote suggests by itself, e.g. OOD labels dramatically hurt performance.
-
Few-shot Prompt Setup Notes for Scaling Law Experiments
By
–
Some setup notes (cc @EthanJPerez
) – We used the exact 2-shot prompt for Quote Repetition, which is already U-shaped for Gopher/Chinchilla – We used fewer shots for Hindsight – We did few-shot instead of 0-shot for Negation QA – We also showed inverse scaling up to PaLM 62B