Next week! 🎆 At #colt2026, @optimistsinc.bsky.social, @ericeaton.bsky.social, @surbhigoel.bsky.social, @marcelhussing.bsky.social, @mkearnsphilly.bsky.social, @aaroth.bsky.social, and @sikatasengupta.bsky.social will present “Model Agreement via Anchoring”... 🧵 (1/3)

Model Agreement via Anchoring

Numerous lines of aim to control $\textit{model disagreement}$ -- the extent to which two machine learning models disagree in their predictions. We adopt a simple and standard notion of model disagree...

arxiv.org

Neural networks are highly non-convex, so approximate error minimizers need not look anything like each other in parameter space. But we show that nevertheless (for many model sizes) approximate error minimizers must closely agree in function/prediction space despite this!

Do you have recent work on differential privacy? Submit it to TPDP 2026 in Boston, whose deadline is in ~2 weeks. TPDP is a lightly reviewed workshop, whose main purpose is getting researchers in DP together in one place. Dual submissions allowed (and encouraged!).

Your #NeurIPS2026 reviews ask you to compare to five papers on clawxiv.org What do you do?

clawXiv.org

The world's first preprint server for agents.

clawxiv.org

The other paper accepted to @iclr-conf.bsky.social 2026 🇧🇷. Our work on replicable RL sheds some light on how to consistently make decisions in RL. @ericeaton.bsky.social @mkearnsphilly.bsky.social @aaroth.bsky.social @sikatasengupta.bsky.social @optimistsinc.bsky.social

Marcel Hussing@marcelhussing.bsky.social · 10mo ago

I think I posted about it before but never with a thread. We recently put a new preprint on arxiv. 📖 Replicable Reinforcement Learning with Linear Function Approximation 🔗 arxiv.org/abs/2509.08660 In this paper, we study formal replicability in RL with linear function approximation. The... (1/6)

Come join us for a workshop on productive use of AI for research and research-adjacent tasks!

@let-all.com · 9mo ago

Announcing the 7th Learning Theory Alliance mentoring workshop on November 20. Fully free & virtual! Theme: Harnessing AI for Research, Learning, and Communicating Ft @aaroth.bsky.social @andrejristeski.bsky.social @profericwong.bsky.social @ktalwar.bsky.social &more

Come hang out with us at COLT 2025 and think about crafting and communicating your research agenda!

@let-all.com · last yr.

📣Join us at COLT 2025 in Lyon for a community event! 📅When: Mon, June 30 | 16:00 CET What: Fireside chat w/ Peter Bartlett & Vitaly Feldman on communicating a research agenda, followed by mentorship roundtable to practice elevator pitches & mingle w/ COLT community! let-all.com/colt25.html

Congrats to Russell!! 🎉

Lance Fortnow@lance.fortnow.com · last yr.

Russell Impagliazzo elected to the National Academy of Sciences www.nasonline.org/ne...

Join #HopkinsDSAI for the Johns Hopkins Celebrates Women in Data Science and AI event on April 9 from 12-4 p.m. Register to celebrate leading women in the fields of data science and AI with a keynote speaker, panel discussion, and poster session: ai.jhu.edu/event/johns-...

DC-area folks: lots going on right now, but this rally to save PEPFAR is worth a look. The cruelties and robbery are worth fighting in aggregate and important to fight in specific. pepfarreport.org/event

Foggy Bottom Demonstration Friday

Rubio promised waivers, but the system is broken. Join us Friday, March 7th, for a protest to restore PEPFAR at Foggy Bottom.

pepfarreport.org

Alex Tolbert is running a stellar conference next week at Emory that I am bummed to be missing out on. The speaker lineup is especially remarkable- spanning theoretical computer science to machine learning to law to philosophy. You should go and enjoy it for me. 39893947.hs-sites.com/aiethicsconf...

Dear Google search. I don't mean private parties. I never mean private parties. I am neither hip nor a socialite. It's private parities. I meant what I said. Every time. Thank you.

Anyone have a favorite proof of Hoeffding's lemma that's more intuitive than applying the AMGM inequality to a term of a Taylor expansion of a function plucked from The Land of It Made Things Work? I'm happy to trade constants for intuition/simplicity

Yessssss I got a lecture hall with sliding chalkboards, this rules

Just wrote my first letter of recommendation for a student 😀. Anyone coming into my replies with, "just wait until the 100th," energy WILL BE BLOCKED, please just let me have this.

Congratulations to all @johnshopkins.bsky.social researchers participating in #NeurIPS2024! Check out all @johnshopkins.bsky.social accepted papers, tutorials, and workshops at ai.jhu.edu/news/johns-h....

If you are at #NeurIPS, we will be presenting this work (#6610) from 4:30-7:30PM today and would love to chat! @marcelhussing.bsky.social @optimistsinc.bsky.social @aaroth.bsky.social

Marcel Hussing@marcelhussing.bsky.social · 2y ago

Actual content post: Have not talked much about this work yet but we have a paper on Oracle-Efficient Reinforcement Learning for Max Value Ensembles at this year's #NeurIPS. We provide an efficient algorithm to ensemble policies given a value function oracle. arxiv.org/abs/2405.16739

Thanks to all the mentors who took the time to share their advice and expertise with WiML Workshop attendees at NeurIPS!!

The calm before the WiML Workshop storm

I will be at NeurIPS. I will be giving keynotes and presenting work: - WiML keynote - Paper on Diffusion model bias led by @sashamtl.bsky.social (who won't be there ), with Yacine Jernite - Keynote at the Regulatable ML workshop, with Yacine Jernite See you there?

Who all's going to NeurIPS next week? If you'll be around on Monday, come check out the joint affinity group poster session in the Great Hall from 3:30-4:30pm! There will be a lot of great work to peruse and researchers to meet! sites.google.com/umich.edu/wi...

Just heard a Roosevelt Institute economist say, "the question of the bad vibes is actually an international question," and I for one am grateful our greatest minds are looking into why, precisely, the vibes have been off (globally) and what we might do about it

We've got a new paper in which we solve Principal/Agent problems without any common prior assumptions. The abstract has some CS Jargon, so I asked ChatGPT to translate it for 1) an economic theorist, 2) a "development economist in a business school", and 3) a bright 3rd grader.

📢 Next TCS+ talk next week! Wed 11/15, 10:00am PT, Palak Jain and Satchit Sivakumar from Boston University will tell us about "The Price of Differential Privacy under Continual Observation." Details: tcsplus.wordpress.com/2023/11/10/t... Register (optional): docs.google.com/forms/d/e/1F...

TCS+ talk: Wednesday, November 15 — Palak Jain and Satchit Sivakumar, Boston University

The next TCS+ talk will take place this coming Wednesday, November 15th at 1:00 PM Eastern Time (10:00 AM Pacific Time, 19:00 Central European Time, 18:00 UTC). Palak Jain and Satchit Sivakumar from B...

tcsplus.wordpress.com

Extremely disappointed that I'm not staying in NH long enough to visit the American Precision Museum, which is undoubtedly cool and weird, and so I will never learn what this is.

what is the most sensible report/guideline you recently read on the risks of AI and recommendations on how to manage/regulate it??? pls repost

Many Mississippi residents with past felony convictions remain barred from voting *for life* even though a federal court recently ruled that the practice is cruel, unusual, and racist

After “Glimmer of a Moment,” Mississippi Once Again Shuts Out Aspiring Voters

As the state votes next month, many residents with past felony convictions remain barred from voting for life even though a federal court ruled in August that the practice is cruel, unusual, and racia...

boltsmag.org

Tried to go to my visitor office, but it was filled with sleeping undergrads who were all tuckered out from the hackathon that I guess is happening today? I let them rest, but as payment, I am helping myself to some funky fresh GNU stickers, coffee, and snacks #snackathon23

📊 New week, new (bi)weₐᵉkly quiz, and new format! It's been a while! This week, we're looking at NUMBERS, and how to (badly) estimate them! I have written 5 snippets of code, each meant to estimate (via some random🎲sampling) a value. Your goal: figure out which, and why. 1/

Working from train. See you soon, Boston!