ruchiradatta.bsky.social
Mathematician and computer scientist. Have done research in game theory, AI/ML for NLP, and computational biology (phylogenomics, cancer, and immunology). Now working on making energy-efficient AI/ML. Vegetarian.
34 posts
485 followers
1,931 following
Prolific Poster
Conversation Starter
comment in response to
post
advertising.amazon.com/terms
comment in response to
post
📌
comment in response to
post
Some people have recommended the paid search engine kagi. I don't know if it fits the bill, I haven't tried it yet.
comment in response to
post
A classic book that was recommended to me is Colin Ware's _Information Visualization: Perception for Design_. I haven't read it yet, so am not sure how deep it goes into what you're seeking.
comment in response to
post
Congratulations!
comment in response to
post
Awaiting
comment in response to
post
Congratulations!
comment in response to
post
On Deezer the cool kids come and go, rapping for Adam the Kotsko.
comment in response to
post
The public post I saw shared on Facebook was by one Roger Tang.
comment in response to
post
That's also how I found you.
comment in response to
post
For more about Project Phasor:
Homepage: www.project-phasor.org
Discord: discord.gg/qV5QAEkU
#projectphasor #neuroai #neuromorphic
comment in response to
post
I have a lot more experiments from working on Syllabus so I’ll share more of those over the next few weeks. Now is probably a good time to mention I’m also looking for industry or postdoc positions starting in Fall 2024, so if you’re working on anything RL-related let me know!
comment in response to
post
Syllabus opens up a ton of low hanging fruit in CL. I’m still working on this and actively using it for my research, so if you’re interested in contributing, please feel free to reach out!
Paper: arxiv.org/abs/2411.11318
Github: github.com/RyanNavillus...
comment in response to
post
We have implementations of Prioritized Level Replay, a learning progress curriculum, and Prioritized Fictitious Self Play, plus several tools for manually designing curricula like simulated annealing and sequential curricula. Stay tuned for more methods in the very near future!
comment in response to
post
These portable implementations of CL methods work with nearly any RL library, meaning that you only need to implement the method once to guarantee that the same CL code is being used in every project. This minimizes the risk of implementation errors and promotes reproducibility.
comment in response to
post
Most importantly, it’s extremely easy to use! You add a synchronization wrapper to your environments and your curriculum, plus a little more configuration, and it just works. For most methods, you don’t need to make any changes to the actual training logic.
comment in response to
post
Syllabus helps researchers study CL in complex, open-ended environments without having to write new multiprocessing infrastructure. It uses a separate multiprocessing channel between the curriculum and environments to directly send new tasks and receive feedback
comment in response to
post
As a result, CL research often focuses on relatively simple environments, despite the existence of challenging benchmarks like NetHack, Minecraft, and Neural MMO. Unsurprisingly, many of the methods developed in simpler environments won’t work as well on more complex domains.