Lightnews — Scholar-powered news

Quentin Berthet @qberthet.bsky.social · 26d

Being a rejected paper seems like a good criteria (possible issues, but very easy to implement)

Quentin Berthet @qberthet.bsky.social · 26d

I'm surprised, there must have been an additional round of changes.

All the emails I saw as an SAC were "help your ACs write new meta-reviews".

1

Quentin Berthet @qberthet.bsky.social · 27d

I think that they are hard to bootstrap.

While I agree that general chairs, program chairs, local chairs etc do a lot of work (seeing a glimpse of that sometimes myself with ALT/AISTATS), once you start having a bit of money, it can get easier with using conference organising services

1

Quentin Berthet @qberthet.bsky.social · 27d

How do-able would it be to organize a conference with its deadline a week after NeurIPS, held a week before, that would always take place physically in Europe (but welcome contributions from anywhere)?

1

Quentin Berthet @qberthet.bsky.social · 27d

I'm very happy that EurIPS is happening, and planning to attend.

But as a NeurIPS SAC I must say that the last definition should not exist - ACs are supposed to change the decisions and update their meta-review themselves.

2 1

Reposted by Quentin Berthet

Claire Vernade @claireve.bsky.social · Aug 29

🚨AISTATS workshop proposals open!

Big News: For the first time, there will be a day of workshops at #AISTATS 2026, in Tangier, Morocco 🌴🇲🇦

Quentin Berthet @qberthet.bsky.social and I are workshop chairs.

virtual.aistats.org/Conferences/...
Deadline: Oct 17, AOE

Call for Workshops

virtual.aistats.org

12 28

Quentin Berthet @qberthet.bsky.social · Apr 17

At #AISTATS2025, I will be giving an "Oral" presentation of our work on "Implicit Diffusion"

arxiv.org/abs/2402.05468

1 8

Quentin Berthet @qberthet.bsky.social · Apr 17

I will be attending #ICLR2025 in Singapore and #AISTATS2025 in Mai Khao over the next two weeks.

Looking forward to meeting new people and learning about new things. Feel free to reach out if you want to talk about Google DeepMind.

3 4 18

Quentin Berthet @qberthet.bsky.social · Mar 25

ICML reviews are out

3

Quentin Berthet @qberthet.bsky.social · Feb 24

Very proud to have contributed to this work, and very happy to have learnt a lot about votes, choices, and agents!

Marc Lanctot @sharky6000.bsky.social · Feb 24

Looking for a principled evaluation method for ranking of *general* agents or models, i.e. that get evaluated across a myriad of different tasks?

I’m delighted to tell you about our new paper, Soft Condorcet Optimization (SCO) for Ranking of General Agents, to be presented at AAMAS 2025! 🧵 1/N

7

Quentin Berthet @qberthet.bsky.social · Feb 22

Il y a plus de 200 notes ?

1

Quentin Berthet @qberthet.bsky.social · Feb 21

🇨🇦

3

Reposted by Quentin Berthet

Samuel Vaiter @samuelvaiter.com · Feb 20

📣 New preprint 📣

Learning Theory for Kernel Bilevel Optimization

w/ @fareselkhoury.bsky.social E. Pauwels @michael-arbel.bsky.social

We provide generalization error bounds for bilevel optimization problems where the inner objective is minimized over a RKHS.

arxiv.org/abs/2502.08457

1 6 20

Quentin Berthet @qberthet.bsky.social · Feb 10

These findings are quite consistent

Our end-to-end method captures a regression and a classification objective, as well as the autoencoder loss.

We see it as "building a bridge" between these different problems.

8/8

3

Quentin Berthet @qberthet.bsky.social · Feb 10

This is a simple generalization from previous binning approaches, the main difference being that we learn the encoding.

We compare different training methods, showing up to 25% improvement on the least-squares baseline error for our full end-to-end method, over 8 datasets.

7/8

1 2

Quentin Berthet @qberthet.bsky.social · Feb 10

You can then train a classification model π_θ (green, top) on the encoded targets, with a KL/cross-entropy objective!

At inference time, you use the same decoder μ to perform your prediction!

6/8

1 1

Quentin Berthet @qberthet.bsky.social · Feb 10

First, the architecture - we use a general target encoder Ψ_w (red, bottom) that transforms the target y in a distribution over k classes

e.g. use softmax(dist) to k different centers

The encoder and the associated decoder μ (in blue) can be trained on an autoencoder loss

5/8

1 2

Quentin Berthet @qberthet.bsky.social · Feb 10

In this work, we propose improvements for "regression as classification"

- Soft-binning: encode the target as a probability, not just a one-hot.
- Learnt target encoders: Instead of designing this transformation by hand, learn it from data.
- Train everything jointly!

4/8

1 2

Quentin Berthet @qberthet.bsky.social · Feb 10

The idea is to transform each value of the target into a class one-hot, and to train a classification model to predict the value of the target. (here with y in 2D, with a grid)

It seems strange, but it's been shown to work well in many settings, even for RL applications.

3/8

1 1

Quentin Berthet @qberthet.bsky.social · Feb 10

An intriguing observation:

In many tasks with a continuous target (price, rating, pitch..), instead of training on a regression objective with least-squares [which seems super natural!] - people have been instead training their models using classification!

2/8

1 1

Quentin Berthet @qberthet.bsky.social · Feb 10

Check out our paper, with Lawrence Stewart and @bachfrancis.bsky.social

Link: arxiv.org/abs/2502.02996

1/8

Building Bridges between Regression, Clustering, and Classification

Regression, the task of predicting a continuous scalar target y based on some features x is one of the most fundamental tasks in machine learning and statistics. It has been observed and...

arxiv.org

1 2 9

Quentin Berthet @qberthet.bsky.social · Feb 10

🚨 New paper on regression and classification!

Adding to the discussion on using least-squares or cross-entropy, regression or classification formulations of supervised problems!

A thread on how to bridge these problems:

4 8 49

Reposted by Quentin Berthet

Tom Phillips @tom-phillips.com · Feb 5

I've got to say, the guy who accidentally became the Director of the FBI does 100% look like the guy who accidentally becomes the Director of the FBI in a mid-2000s comedy about a guy who accidentally becomes the Director of the FBI

Brian Driscoll, the acting director of the FBI, sporting curly hair, a kinda rakish jazz dirtbag moustache and beard combo, a faintly amused look in his eyes and a wry "whaddaya gonna do" smirk

570 2.7K 21K

Reposted by Quentin Berthet

Oussama Zekri @ozekri.bsky.social · Feb 4

🚀 Policy gradient methods like DeepSeek’s GRPO are great for finetuning LLMs via RLHF.

But what happens when we swap autoregressive generation for discrete diffusion, a rising architecture promising faster & more controllable LLMs?

Introducing SEPO !

📑 arxiv.org/pdf/2502.01384

🧵👇

1 2 5

Quentin Berthet @qberthet.bsky.social · Feb 4

Congratulations!

2