Here’s a quick rundown of the most popular (RAG) search approaches out there, plus tips on when they shine: • Keyword Search (e.g., BM25, BM42): Finds direct matches—great for straightforward data or well-defined keywords. • Embedding Search: Captures deeper semantics,
RESEARCH
-
Stanford’s Robotic Boot Reduces Energy Use While Improving Mobility
By
–
Stanford's new robotic boot proves tech isn't just hype.
— Pascal Bornet (@pascal_bornet) 3 janvier 2025
It makes you feel 30 pounds lighter, helps you walk 9% faster while using 17% less energy, and adapts to your unique stride.
This isn't just cool tech—it's life-changing mobility for those who need it most.
By choosing… pic.twitter.com/jKM6iKl61RStanford's new robotic boot proves tech isn't just hype. It makes you feel 30 pounds lighter, helps you walk 9% faster while using 17% less energy, and adapts to your unique stride. This isn't just cool tech—it's life-changing mobility for those who need it most. By choosing
-

LLM Fine-Tuning Notebooks with OpenAI and Hugging Face
By
–
check out some of my latest notebooks containing llm fine-tuning using OAI and HF fine-tuners https://
github.com/avikumart/LLM-
GenAI-Transformers-Notebooks/tree/main/TMLC_LLM_projects
… -
LLMs Don’t Read Books Sequentially During Training
By
–
From the replies, some people appear to think that LLMs read books in order. They don’t; they take interleaves batches samples from the entire training set.
-

AI Decodes Animal Communication Patterns with Machine Learning
By
–
AI decodes the calls of the wild https://
go.nature.com/4gp4X0f
#AI #MachineLearning #DeepLearning #LLMs #DataScience -
AI Experts in News and Media Research
By
–
If you want a startup founder focused on AI-assisted news, @fpmarconi
’s great. For someone researching AI’s impact on news, try @_FelixSimon_ or @rasmus_kleis
. -
AI Solving Cancer Treatment and Recycling Challenges
By
–
From treating cancer to revolutionizing recycling, #AI can help solve some of humanity's most pressing and complex problems.
-

Tribute to Felix Hill’s Influential Deep Learning NLP Lectures
By
–
I have not personally met @FelixHill84
, but this is incredibly sad to hear. I have watched his lectures on deep learning for NLP several times, they were my reference materials for long time, and still valuable today. May his family, friends, and colleagues stay strong -
Tracking Per-Token Loss to Measure Document Contribution in LLM Training
By
–
It would be interesting if LLM training tracked the per-token loss back to the source material — it would be an objective measure of how much each specific book / document contributed to the training. Might say something useful for human learning!

