Beyond the basics: w/ a huge & dense literature, papers building on top of each other, it's hard to jump into research. Accessible technical material of more advanced topics is helpful. With @let4all we solicit and curate learning theory blog posts: https://
let-all.com/blog/ 4/n
RESEARCH
-
Making Advanced Learning Theory Research More Accessible Through Blog Posts
By
–
-
Importance of Open Educational Materials for Machine Learning Newcomers
By
–
One very important technical tool is good open introductory educational material: textbooks, lecture notes, videos. O/W it's very difficult for newcomers w/o connections to learn the basics. See eg the classic @shai_s_shwartz @shaibendavid5 book https://
cs.huji.ac.il/~shais/Underst
andingMachineLearning/
… 2/n -
Solved AI but Schmidhuber Already Did It
By
–
The other day I solved AI, but why bother to publish it? Somehow, somewhere Juergen Schmidhuber has already done it.
-
Infinity Non-Existence Challenges Turing Machine Hierarchy Theory
By
–
Because infinity does not exist, there is no difference between a Turing machine and a finite state machine, and the Chomsky hierarchy is vacuous.
-
Opinion: RLHF not necessary for stop tokens
By
–
You don't need RLHF for stop tokens; GPT‑3 had that long before RLHF shipped, SFT works fine, and here base Llama 3.1 gets it (presumably) just from Llama 2 prompts floating around online. RLHF does a lot but it's not that deep; I doubt it's adding new concepts.
-
India AI Innovation Centre Launches Indigenous Large Multimodal Models
By
–
The pillar IndiaAI Innovation centre of the INDIAai mission is dedicated to developing and deploying indigenous Large Multimodal Models (LMMs) and domain-specific foundational models in critical sectors.
— IndiaAI (@OfficialINDIAai) 29 août 2024
Visit https://t.co/PTPrQmCzor to know more!@GoI_MeitY @abhish18… pic.twitter.com/qSyREig8e7The pillar IndiaAI Innovation centre of the INDIAai mission is dedicated to developing and deploying indigenous Large Multimodal Models (LMMs) and domain-specific foundational models in critical sectors. Visit https://
indiaai.gov.in to know more! @GoI_MeitY @abhish18 -
Stanford Med Conference Speakers Downplay Pandemic Policy Failures
By
–
@StanfordMed is hosting a pandemic policy conference in October. Many of the featured speakers would like everyone to forget that they advocated for rapidly infecting all young people and incorrectly predicted that this would end the spread in months.
-
Llama 3.1 base — AI categories and topics
By
–
Also see my previous thread demonstrating Llama 3.1 base:
-
Unabridged Llama 3.1 base outputs show psychosis-like spontaneity
By
–
"Hey what's up." — An unabridged, minimally prompted conversation with Llama 3.1 405B base bf16. Responses exhibit naturalistic, emotional prose and psychosis-like spontaneity, both common in under-prompted base Llama and vividly unlike the output of widely used chat LLMs.
-
ReMamba: Enhancing Mamba Architecture for Long-Sequence Modeling
By
–
ReMamba Equip Mamba with Effective Long-Sequence Modeling discuss: https://
huggingface.co/papers/2408.15
496
… While the Mamba architecture demonstrates superior inference efficiency and competitive performance on short-context natural language processing (NLP) tasks, empirical evidence suggests