Lightnews — Scholar-powered news

Reposted by Rabiraj Banerjee

Subbarao Kambhampati (కంభంపాటి సుబ్బారావు) @rao2z.bsky.social · 24d

In the year since LRMs ("reasoning models") hit the scene, we have been trying to understand, analyze and demystify them.. Here are our efforts to date--conveniently all in one place..👇

www.linkedin.com/posts/subbar...

In the year since LRMs ("reasoning models") hit the scene, we have been trying to understand, analyze and demystify them.. Here are our efforts to date--conveniently all in one… | Subbarao K...

In the year since LRMs ("reasoning models") hit the scene, we have been trying to understand, analyze and demystify them.. Here are our efforts to date--conveniently all in one place.. (𝗙𝗶𝗿𝘀𝘁..) 𝗘𝘃𝗮𝗹...

www.linkedin.com

1 4

Reposted by Rabiraj Banerjee

Sung Kim @sungkim.bsky.social · 27d

A Survey of Reinforcement Learning for Large Reasoning Models

Five sections:

- Foundational Components
- Foundational Problems
- Training Resources
- Applications
- Future Directions

1 3 18

Reposted by Rabiraj Banerjee

Yoav Goldberg @yoavgo.bsky.social · Aug 27

When reading AI reasoning text (aka CoT), we (humans) form a narrative about the underlying computation process, which we take as a transparent explanation of model behavior. But what if our narratives are wrong? We measure that and find it usually is.

Now on arXiv: arxiv.org/abs/2508.16599

Humans Perceive Wrong Narratives from AI Reasoning Texts

A new generation of AI models generates step-by-step reasoning text before producing an answer. This text appears to offer a human-readable window into their computation process, and is increasingly r...

arxiv.org

4 22 85

Reposted by Rabiraj Banerjee

Johan Ugander @jugander.bsky.social · Aug 24

Great interview with @stevenstrogatz.com with a lot of discussion of research advising. Parts reminded me of @eegilbert.org and @informor.bsky.social's (excellent) guides to PhD mentorship, with a big focus on ideation.
Eric's: docs.google.com/document/d/1...
Mor's: s.tech.cornell.edu/phd-syllabus/

8 49

Reposted by Rabiraj Banerjee

Nikhil Garg @nkgarg.bsky.social · Aug 20

We try to avoid self-promoting too much, but we (with @sjgreenwood.bsky.social) built a personalized feed with posts about papers from your network. Many people say it's the closest they can get to old academic twitter, and I hope you enjoy it and share with others too!

bsky.app/profile/pape...

Paper Skygest Team @paper-feed.bsky.social · Aug 19

**Please repost** If you're enjoying Paper Skygest -- our personalized feed of academic content on Bluesky -- we'd appreciate you reposting this! We’ve found that the most effective way for us to reach new users and communities is through users sharing it with their network

1 5 19

Reposted by Rabiraj Banerjee

Jacob Eisenstein is at CoLM @jacobeisenstein.bsky.social · Aug 19

1 6

Reposted by Rabiraj Banerjee

Margaret Mitchell @mmitchell.bsky.social · Aug 13

🤖 But wait! There's more! You can check out @shiraamitchell.bsky.social 's most recent update on the details of Calibration, posted yesterday! statmodeling.stat.columbia.edu/2025/08/12/s...

3 7

Reposted by Rabiraj Banerjee

brendan o’connor (going to COLM) @brenocon.bsky.social · Jul 28

#acl2025 anyone get a good quote of phil resnik's last comment?

context: (some?all?) panelists & him agree the field needs more deep, careful research on smaller models to do better science. everyone is frustrated with impossibility of large-scale pretraining experiments

1 1 7

Rabiraj Banerjee @rabirajb.bsky.social · Jul 24

@kennyjoseph.bsky.social , Kenny check this thread out

Maria Antoniak @mariaa.bsky.social · Jul 23

What are your favorite recent papers on using LMs for annotation (especially in a loop with human annotators), synthetic data for task-specific prediction, active learning, and similar?

Looking for practical methods for settings where human annotations are costly.

A few examples in thread ↴

2

Rabiraj Banerjee @rabirajb.bsky.social · Jul 24

aclanthology.org/2023.emnlp-m..., for Active Learning I really liked this paper, uses LLMs as annotator for knowledge distillation for small LMs

aclanthology.org

1

Reposted by Rabiraj Banerjee

Maria Antoniak @mariaa.bsky.social · Jul 23

What are your favorite recent papers on using LMs for annotation (especially in a loop with human annotators), synthetic data for task-specific prediction, active learning, and similar?

Looking for practical methods for settings where human annotations are costly.

A few examples in thread ↴

13 23 76

Rabiraj Banerjee @rabirajb.bsky.social · Jul 21

This is so mean !!!

Reposted by Rabiraj Banerjee

Kenny Joseph @kennyjoseph.bsky.social · Jul 17

UB's new Department of AI and Society is hiring faculty across ranks (Assistant, Associate, Full Professor). We’re looking for transdisciplinary scholars interested in building AI by society, for society. Start dates begin Fall 2025.

More info: www.ubjobs.buffalo.edu/postings/57734

Assistant, Associate or Full Professor, AI & Society

The Department of AI and Society (AIS) at the University at Buffalo (UB) invites candidates to apply for multiple positions as Assistant Professor, Associate Professor, or Full Professor. The new AIS ...

www.ubjobs.buffalo.edu

9 10

Reposted by Rabiraj Banerjee

Alexander Hoyle @alexanderhoyle.bsky.social · Jul 16

Took me a second, but I knew I'd seen something related to this recently:

arxiv.org/abs/2505.120...

Do different prompting methods yield a common task representation in language models?

Demonstrations and instructions are two primary approaches for prompting language models to perform in-context learning (ICL) tasks. Do identical tasks elicited in different ways result in similar rep...

arxiv.org

1 2 10

Reposted by Rabiraj Banerjee

Yanai Elazar @yanai.bsky.social · Jul 1

Check out our take on Chain-of-Thought.
I really like this paper as a survey on the current literature on what CoT is, but more importantly on what it's not.
It also serves as a cautionary tale to the (apparently quite common) misuse of CoT as an interpretable method.

Fazl Barez @fbarez.bsky.social · Jul 1

Excited to share our paper: "Chain-of-Thought Is Not Explainability"! We unpack a critical misconception in AI: models explaining their steps (CoT) aren't necessarily revealing their true reasoning. Spoiler: the transparency can be an illusion. (1/9) 🧵

1 4 14

Rabiraj Banerjee @rabirajb.bsky.social · Jul 1

These are just battle scars of doing Data Science and ML Engineering in the industry!!

Reposted by Rabiraj Banerjee

Mark Riedl @markriedl.bsky.social · Jul 1

Hi everyone. I'm excited to announce that I will be organizing a 2nd Summit on Responsible Computing, AI, and Society rcais.github.io October 27-29, 2025.

We will explore the future of computing for health, sustainability, human-centered AI, and policy.

Please consider submitting a 1-page abstract

Screenshot of website
The Georgia Tech School of Interactive Computing is hosting the 2025 Summit on Responsible Computing, AI, and Society, October 27-29, 2025.

Overview

The Summit on Responsible Computing, AI, and Society aims to explore the future of computing for health, sustainability, human-centered AI, and policy. The summit will bring together luminary researchers in computing for health, sustainability, human-centered AI, and tech policy to lay out the frontiers of these critical fields, and to plot out how they must evolve.

1 4 17

Rabiraj Banerjee @rabirajb.bsky.social · Jun 27

Send him a note of appreciation 😊

Rabiraj Banerjee @rabirajb.bsky.social · Jun 26

To @pcarragher.bsky.social @lleibm.bsky.social , @jmendelsohn2.bsky.social , Evan and Catherine, and others for some really fruitful and nice convos, hope to see you all soon.

2

Rabiraj Banerjee @rabirajb.bsky.social · Jun 26

A huge shoutout to the organizing team, and to the web chair
@andersgiovanni.com
for updating the schedule in such an easy to follow manner, hope you get some well deserved rest (as an ex web chair I know the pain)

2 2

Rabiraj Banerjee @rabirajb.bsky.social · Jun 26

So ICWSM concluded today and it was a blast, was a great honor to attend @icwsm.bsky.social at Copenhagen and present my work with @kennyjoseph.bsky.social and other colleagues. The paper link is here :
ojs.aaai.org/index.php/IC...,

View of Measuring Dimensions of Self-Presentation in Twitter Bios and their Links to Misinformation Sharing

ojs.aaai.org

1 2 9

Reposted by Rabiraj Banerjee

Shubhendu Trivedi @shubhendu.bsky.social · Jun 26

This is great! The idea is somewhat obvious (good!), and I'm sure many have toyed with the connection to learning-to-rank. However, no work had developed it. This should be relevant for constructing valid PIs from just preferential feedback. openreview.net/pdf?id=ENJd3...

1 2

Rabiraj Banerjee @rabirajb.bsky.social · Jun 26

Me we were in the same session :) (Session 8)

1

Reposted by Rabiraj Banerjee

Naomi Saphra @nsaphra.bsky.social · Jun 24

🚨 New preprint! 🚨
Phase transitions! We love to see them during LM training. Syntactic attention structure, induction heads, grokking; they seem to suggest the model has learned a discrete, interpretable concept. Unfortunately, they’re pretty rare—or are they?

3 10 48

Reposted by Rabiraj Banerjee

Hanna Wallach @hannawallach.bsky.social · Jun 16

Generative language systems are everywhere, and many of them stereotype, demean, or erase particular social groups.

1 2 9