Part 2: So why DO Random Forests work?! On this, I’ll have to disagree with Elements of Statistical Learning (my first time ever 💔) EoSL says the success of forests should be understood as a consequence of variance reduction *alone*, but I think that’s not a good intuition 1/n

cant wait for another year of the machine learning research experience in 2024

when reviewer 2 deeply misunderstands your paper but recommends "weak accept" anyway

The reproducibility crisis might be a lot better if we *didn’t* publish papers that claim obviously false conclusions. Like, of course if e.g. Yann LeCun tweets a paper out it will cause more citations. This doesn’t even need a study 🤦‍♂️

Excited to be back at #icml!☀️Find me floating around or come chat to me & Alan Jeffares about our integrated attempt at understanding deep double descent, grokking, linear mode connectivity & differences between gradient boosting and neural nets on Friday at the HiLD workshop!🤓

Great news for Alicia Curth!

Tomorrow at ICML (13:30 poster #2415) we will present our work that “relaxes” the symmetric noise assumption of quantile regression, proposing a simple, effective alternative. Come say hi!!

I generally find ML review quality pretty poor so I appreciate the recognition of my efforts trying to improve things

Excellent work by TPouplin and Alan Jeffares for ICML Conference! Fantastic to see my students confidently presenting their cutting-edge research! Well done! 👏 Read their paper here: arxiv.org/abs/2406.03258

Want to hear about the next stop on our journey into understanding modern deep learning phenomena? Come find Alan Jeffares & myself in the poster sessions at 10:00 and 15:30 at the workshop on high-dimensional learning dynamics ICML Conference in Straus 2 tomorrow! 🙌🏻 #ICML2024

So the human brain is Mixture of Experts with n=2?

that post NeurIPS rebuttal feeling

(1/3) Tired of Reflection-70B controversy? We’d like to share our recent work on LLM Benchmark Cheating and Detecting. (Reflection-70B included in Fig. 3!) Title: “Data Contamination Can Cross Language Barriers” Paper: arxiv.org/pdf/2406.13236 Code: github.com/shangdatalab/D…