Emil Carlsson

Reinforcement Learning and Statistics.

[T3.com @T3dotcom]

Sleep Cycle just gave its app an AI upgrade that could drastically improve your sleep https://t.co/ckFkNpU844

[Emil Carlsson @ecarlsn]

Snart 24h och vi lever fortfarande!

https://t.co/P1A6O35vCU

[Dr. Mwai Newton @Mwai_Newton]

Over the past five years, I’ve been focused on my PhD in machine learning, studying how to leverage historical data for interactive personalization, at Chalmers University in Sweden. For more on my work and other AI endeavours, visit https://t.co/lBqzZMzRin!

[Emil Carlsson @ecarlsn]

Work together with Herman Bergström, @frejohk and Devdatt Dubhashi

[Emil Carlsson @ecarlsn]

Attending NeurIPS and interested in preference learning?
Visit our poster today to learn about GURO, an active learning algorithm that greedily minimizes a bound on the ordering error to determine the next query.

📍 Session: West Ballroom, Booth 6604
🕚 Time: 11:00 AM - 2:00 PM

[Emil Carlsson @ecarlsn]

#NeurIPS2024
https://t.co/jz176XaXp7

[Emil Carlsson @ecarlsn]

I’m at NeurIPS this week presenting joint work with Herman Bergström, @frejohk and Devdatt Dubhashi on how to use context for better preference learning.

Feel free to reach out if you want to discuss the paper!

https://t.co/jz176Xapzz

[Fredrik Johansson @frejohk]

I'm at NeurIPS, eager to talk about the two papers from my group!

The first, by Herman Bergström, Emil Carlsson, Devdatt Dubhashi, and me, explores how important context is to active preference learning for ordering items from pairwise comparisons: https://t.co/MIMPQDm9Oe

[Emil Carlsson @ecarlsn]

I would also like to express my gratitude towards Kenny Smith, who served as the opponent during my public defense, as well as to the members of my grading committee: Karl Åström, @VerhoefTessa , and @ssshanest.

[Emil Carlsson @ecarlsn]

I successfully defended my PhD thesis earlier this week! 😃👨‍🎓 It feels a bit weird not to be a student anymore😅

I am very grateful for all the support I've received from my supervisor, Devdatt Dubhashi, and my co-supervisor, @frejohk , throughout these years.

[Emil Carlsson @ecarlsn]

🥲

[Anthropic @AnthropicAI]

Recommendation #1: Use the Central Limit Theorem

Evaluation scores are usually mean scores across a large number of questions. This is a perfect fit for the Central Limit Theorem. In the paper, we suggest reporting the CLT-based standard errors alongside mean scores.

[Emil Carlsson @ecarlsn]

A bit late to the party but our paper on active energy optimization with preference dynamics got accepted to the biggest Physics conference in Vancouver this year.

Work together with Herman Bergström, @frejohk and Devdatt Dubhashi.

[IF Elfsborg @IFElfsborg1904]

SEGER!

#IFE | 1-0 | #ASR

[Christian Wolf (🦋🦋🦋) @chriswolfvision]

#CVPR2025 changes:

"If a reviewer is flagged by an Area Chair as “highly irresponsible”, their paper submissions will be desk rejected per the discretion of the PCs"

👏👏👏

https://t.co/rGqqOsIBxd

[Emil Carlsson @ecarlsn]

Stop the count

[Erik Mohlin @karlerikmohlin]

Pretty strong evidence of negative mental health effects of doing a PhD.
Recent working paper by
@EvaRanehill, @annahsandberg, Sanna Bergvall, and Clara Fernström.
Paper link: https://t.co/7X4uEjzPCC

[Fredrik Johansson @frejohk]

Searching for optimal decision-making policies without constraints typically results in universal recommendations: Always do A, not B! Not so if you have constraints on budget, diversity, side effects, etc. [1/2]

[Emtiyaz Khan @EmtiyazKhan]

Call for papers for AISTATS 2025 is out now
https://t.co/k2svn9hJZT

You can also (self-)nominate for reviewer/AC position
https://t.co/rprkN2X1HS

We are (trying to) switch to OpenReview so submission details will be available a bit later.

@StephanMandt @liyzhen2 @ashipra
1/

[Debabrota Basu @BasuDebabrota]

As bandits serve as the cornerstone of decision making with incomplete info & sequential interactions (e.g. recommenders), privacy becomes a pressing concern. We show #DifferentialPrivacy for (some) bandits is free. Is it true for contextual bandits also? #COLT2024 #openproblem

[Emil Carlsson @ecarlsn]

“Borderline reject”