kempnerinstitute.bsky.social
The Kempner Institute for the Study of Natural and Artificial Intelligence at Harvard University.
108 posts
399 followers
42 following
Prolific Poster
Active Commenter
comment in response to
post
4/26 at 3pm:
'HARDMath: A Benchmark Dataset for Challenging Problems in Applied Mathematics'
Jingxuan Fan · Sarah Martinson · Erik Wang · Kaylie Hausknecht · Jonah Brenner · Danxian Liu · Nianli Peng · Corey Wang · Michael Brenner
Submission: openreview.net/forum?id=nDT...
comment in response to
post
4/26 at 3pm:
'Deconstructing What Makes a Good Optimizer for Autoregressive Language Models'
@rosieyzh.bsky.social · Depen Morwani · David Brandfonbrener · Nikhil Vyas · Sham Kakade
Submission: openreview.net/forum?id=zfe...
comment in response to
post
4/26 at 3pm:
'Recite, Reconstruct, Recollect: Memorization in LMs as a Multifaceted Phenomenon'
USVSN Sai Prashanth · @nsaphra.bsky.social et al
Submission: openreview.net/forum?id=3E8...
comment in response to
post
4/26 at 3pm:
'Lie Algebra Canonicalization: Equivariant Neural Operators under arbitrary Lie Groups'
Zakhar Shumaylov · Peter Zaika · James Rowbottom · Ferdia Sherry · @mweber.bsky.social · Carola-Bibiane Schönlieb
Submission: openreview.net/forum?id=7PL...
comment in response to
post
4/26 at 3pm:
'The Optimization Landscape of SGD Across the Feature Learning Strength'
Alexander Atanasov · Alexandru Meterez · James Simon · @cpehlevan.bsky.social
Submission: openreview.net/forum?id=iEf...
comment in response to
post
4/26 at 10am:
'MLPs Learn In-Context on Regression and Classification Tasks'
William Tong · @cpehlevan.bsky.social
Submission: openreview.net/forum?id=MbX...
comment in response to
post
4/26 at 10am:
'SOAP: Improving and Stabilizing Shampoo using Adam for Language Modeling'
Nikhil Vyas · Depen Morwani · @rosieyzh.bsky.social · Itai Shapira · David Brandfonbrener · Lucas Janson · Sham Kakade
Submission: openreview.net/forum?id=IDx...
comment in response to
post
4/26 at 10am:
'Mind the Gap: Examining the Self-Improvement Capabilities of Large Language Models'
Yuda Song · Hanlin Zhang · Carson Eisenach · Sham Kakade · Dean Foster · Udaya Ghai
Submission: openreview.net/forum?id=mtJ...
comment in response to
post
4/25 at 3pm:
'When narrower is better: the narrow width limit of Bayesian parallel branching neural networks'
Zechen Zhang · Haim Sompolinsky
Submission: openreview.net/forum?id=CkU...
comment in response to
post
4/25 at 3pm:
'PolyPythias: Stability and Outliers across Fifty Language Model Pre-Training Runs'
Oskar van der Wal · Pietro Lesci · Max Müller-Eberstein · @nsaphra.bsky.social · Hailey Schoelkopf · Willem Zuidema · Stella R Biderman
Submission: openreview.net/forum?id=bmr...
comment in response to
post
4/25 at 10am:
'KGARevion: An AI Agent for Knowledge-Intensive Biomedical QA'
Xiaorui Su · Yibo Wang · Shanghua Gao · Xiaolong Liu · Valentina Giunchiglia · Djork-Arné Clevert · @marinkazitnik.bsky.social
Submission: openreview.net/forum?id=tnB...
comment in response to
post
4/25 at 10am:
'Generalization through variance: how noise shapes inductive biases in diffusion models'
John Vastola
Submission: openreview.net/forum?id=7lU...
comment in response to
post
4/25 at 10am:
'How Does Critical Batch Size Scale in Pre-training?'
Hanlin Zhang · Depen Morwani · Nikhil Vyas · Jingfeng Wu · Difan Zou · Udaya Ghai · Dean Foster · Sham Kakade
Submission: openreview.net/forum?id=JCi...
comment in response to
post
Apr 24 at 3pm:
'Flash Inference: Near Linear Time Inference for Long Convolution Sequence Models and Beyond'
Costin-Andrei Oncescu · Sanket Jayant Purandare · Stratos Idreos · Sham Kakade
Submission: openreview.net/forum?id=cZW...
#ICLR2025
comment in response to
post
Apr 24 at 3pm:
'Mixture of Parrots: Experts improve memorization more than reasoning'
Samy Jelassi · Clara Mohri · David Brandfonbrener · Alex Gu · Nikhil Vyas · Nikhil Anand · David Alvarez-Melis · Yuanzhi Li · Sham Kakade · Eran Malach
openreview.net/forum?id=9XE...
#ICLR2025
comment in response to
post
Featuring work from @nsaphra.bsky.social @marinkazitnik.bsky.social @rosieyzh.bsky.social @mweber.bsky.social @mgkumar138.bsky.social @noorsajidt.bsky.social
comment in response to
post
Featuring: @nsaphra.bsky.social @rosieyzh.bsky.social @mweber.bsky.social @mgkumar138.bsky.social @noorsajidt.bsky.social
comment in response to
post
And for a broad perspective on how this framework helps untangle multiplexed information in neural recordings, check out our piece by @dryohanjohn.bsky.social.
bit.ly/KempnerDUNL2
comment in response to
post
Great work from @btolooshams.bsky.social @saramatias.bsky.social, Hao Wu, Simona Temereanca, @naoshigeuchida.bsky.social, @neurovenki.bsky.social, @paulmasset.bsky.social, and Demba Ba!
comment in response to
post
Featuring: @kanakarajanphd.bsky.social @ryanpaulbadman1.bsky.social @gershbrain.bsky.social @mgkumar138.bsky.social @frostedblakess.bsky.social @cpehlevan.bsky.social @chingfang.bsky.social @andykeller.bsky.social @gordfishell.bsky.social @jdrugowitsch.bsky.social, @naoshigeuchida.bsky.social