Fast Inference of Mixture-of-Experts Language Models with Offloading Artyom Eliseev, Denis Mazur : https://
arxiv.org/abs/2312.17238 #ArtificialIntelligence #DeepLearning #MachineLearning
RESEARCH
-

Fast Inference of Mixture-of-Experts Language Models with Offloading
By
–
-
2024 AI Predictions Thread: Share Your Forecasts
By
–
Okay, 2024 prediction thread for AI. Leave your predictions below! I'll check this tweet next year to see who had the best predictions. 🙂
-
Named Tensors Utility Without Complete Coverage Requirements
By
–
yes its here to stay, as it doesn't need complete coverage to be useful (unlike named tensors)
-
PyTorch Discontinues Named Tensors for Tensor Dims
By
–
we actively stopped developing named tensors in favor of tensor dims
-
Lex Fridman Podcast Quality Critique and Alternative Recommendation
By
–
on podcasts: lex fridman brings on a wonderful selection of guests (mostly) but doesn’t research them beforehand and resorts to asking them things like “do u think aliens are real” for people asking for a replacement, I think @dwarkesh_sp is great and has spectacular guests too
-
AI Discovers Novel Drugs, Now Testing Their Efficacy
By
–
AI is dreaming up drugs that no one has ever seen. Now we've got to see if they work. | MIT Technology Review https://
technologyreview.com/2023/02/15/106
7904/ai-automation-drug-development/
… -
tDCS and TMS: Emerging Neurotechnology Applications
By
–
very early but tDCS is pretty interesting overall. another interesting one is TMS (which uses targeted magnets)
-
GPT-4 Capability to Recognize Knowledge Limits
By
–
was recently thinking about how good gpt-4 seems to be at knowing when it doesn't know something
-
EEG Sensor Conductance Challenges in Brain Computer Interfaces
By
–
yeah, and hold your breath / don't blink! I saw a mention of muse which was one I've tried a few years ago. form factor is nice but has similar issues. most eeg research uses tight sensor caps with gel / saline because conductance is another issue with dry sensors
-
EEG Signal Processing Challenges in Brain-Computer Interface Development
By
–
there have been a bunch of attempts to do this with eeg (tried building my own with openbci too) but there are still big signal processing issues that make it primarily a novelty right now (sensitive to all movement, etcetera)