elenimi

Program Director at NSF, NLP Researcher at UPenn

[elenimi @elenimi]

Can AI See, or Does It Only Talk About Seeing? https://t.co/IsN7CkKdsR

[elenimi @elenimi]

Can AI See, or Does It Only Talk About Seeing? https://t.co/IsN7CkKdsR

[elenimi @elenimi]

https://t.co/iNXi8aL6hN

[elenimi @elenimi]

When LLMs judge LLMs. Free subscription. https://t.co/3qR0AZs8ev

[elenimi @elenimi]

I started a new blog to share -and discuss- ideas on large language models, neural computation, and how artificial and biological systems learn. Join the conversation if these questions resonate with you. https://t.co/pKpuYrYE7a

[elenimi @elenimi]

Excellent talk. Watch it if you missed it.

[NeurIPS Conference @NeurIPSConf]

Zeynep Tufekci brings her unique perspective on technology and society to NeurIPS. Her work examining the societal implications of AI offers crucial context for our community's research. https://t.co/q8uaeeXFE3 #NeurIPS2025 #NeurIPSanDiego

[Liwei Jiang @liweijianglw]

⚠️Different models. Same thoughts.⚠️

Today’s AI models converge into an 𝐀𝐫𝐭𝐢𝐟𝐢𝐜𝐢𝐚𝐥 𝐇𝐢𝐯𝐞𝐦𝐢𝐧𝐝 🐝, a striking case of mode collapse that persists even across heterogeneous ensembles.

Our #neurips2025 𝐃&𝐁 𝐎𝐫𝐚𝐥 𝐩𝐚𝐩𝐞𝐫 (✨𝐭𝐨𝐩 𝟎.𝟑𝟓%✨) dives deep into this phenomenon, introducing 𝐈𝐧𝐟𝐢𝐧𝐢𝐭𝐲-𝐂𝐡𝐚𝐭, a real-world dataset of 26K real-world open-ended user queries spanning 17 open-ended categories + 31K dense human annotations (𝟐𝟓 𝐢𝐧𝐝𝐞𝐩𝐞𝐧𝐝𝐞𝐧𝐭 𝐚𝐧𝐧𝐨𝐭𝐚𝐭𝐨𝐫𝐬 𝐩𝐞𝐫 𝐞𝐱𝐚𝐦𝐩𝐥𝐞) to push AI’s creative and discovery potential forward.

Now you can build your favorite models to be truly original, diverse, and impactful in the open-ended real world.

📍Paper: https://t.co/b4Ml1ZMJNQ
📍Data: https://t.co/YZgV6ApPjJ

We also systematically reveal Artificial Hivemind across:

💥 𝐆𝐞𝐧𝐞𝐫𝐚𝐭𝐢𝐯𝐞 𝐚𝐛𝐢𝐥𝐢𝐭𝐢𝐞𝐬: not only do individual LLMs repeat themselves, but different models produce strikingly similar content, even when asked fully open-ended questions.

💥 𝐃𝐢𝐬𝐜𝐫𝐢𝐦𝐢𝐧𝐚𝐭𝐢𝐯𝐞 𝐚𝐛𝐢𝐥𝐢𝐭𝐢𝐞𝐬: LLMs, LM judges, and reward models are systematically miscalibrated when rating alternative responses to open-ended queries.

(1/N)

[Gautam Kamath @thegautamkamath]

With nearly 30,000 registrants (including virtual), 64% increase since last year, they're saying that #NeurIPS2025 is the largest conference at San Diego Convention Center

[Justin Angel @JustinAngel]

Very cool resource by @davidbau: NDIF. Hosted models larger than 50B dedicated to explainability research, both through web and API. See layer activation, run traces, change weights, all through NDIF. Neat!

David said this really well: most interpability research is done on far from SoTA models, and this could solve that.

https://t.co/WR0y5w372a

[Alex Dimakis @AlexGDimakis]

This is exactly why it will be the decade of agents, not the year of agents. Getting reliability in tasks that take 10h or 10 days will require environments and tasks with increasing complexity to get those needed nines of reliability. Exactly like self-driving cars.

[elenimi @elenimi]

Congratulations to Julia Hirschberg for the well-deserved ACL 2025 Dragomir Radev Distinguished Service Award! #ACL2025NLP

[elenimi @elenimi]

Congratulations to Kathy McKeown for the well-deserved 2025 ACL Lifetime Achievement Award!

[Greeks in AI @greeksinai]

Panel on Greek AI Ecosystem: Strategy and Vision. 🤖

Panelists: @mdr, @ACM_President, Stefanos Kollias, @pappasg69, and @AgoniGrammi. Moderator: @timossellis.

#greeksinai

[elenimi @elenimi]

@RafaelNadal Very disappointing. Will women be able to play tennis in mini skirts or shorts? Very disappointing. I have no words having been one of your dedicated fans and admirers.

[ICML Conference @icmlconf]

John Schulman's Invited Talk at #ICML2023 has been delayed by 30 minutes to 9:30am. Coffee break in the interim!

[University of Toronto Robotics Institute @UofTRobotics]

Wondering what Semantic SLAM to Semantic Navigation is all about? Find out on October 7 at 3:00 PM from Kostas Daniilidis, Ruth Yalom Stone Professor of Computer & Information Science at the University of Pennsylvania. Live at Myhal or click here https://t.co/YnVjC91Enu

[Kostas Daniilidis @KostasPenn]

#CVPR2022 Motion 3: "Any reviewer who has accepted an invitation to review but violates the reviewing guidelines set forth by the conference will be prohibited from submitting any papers to CVPR for up to two
years." My vote is a resounding NO! 1/n

[Karan Wanchoo @WanchooKaran]

Excellent talk by Professor @KostasPenn at @CVPR on Semantic SLAM and Vision & Language Navigation (VLN) today as the conference unfolded on Day-1! Stay tuned for our *Poster presentation on VLN on Thursday (June 23) afternoon* at the venue!! #cvpr2022

[Karan Wanchoo @WanchooKaran]

Today at #CVPR2022 we presented our poster on Cross-modal Map Learning for Vision and Language Navigation! Multi-modality lies at the intersection of CV, NLP and Robotics and is the key to solving future challenges. Project details can be found here: https://t.co/z4KJR0yifr

[elenimi @elenimi]

I am excited to share the invitation to the next Hellenic University Club of Philadelphia Young Scholar Spotlight speaker. Dr. Artemis Zavaliangos-Petropoulou, a former HUC Scholarship recipient will give a talk on Thursday, June 9th, at 7pm EST on https://t.co/gcfSxUCHEU.