#iclr2025 — Public Fediverse posts

🎉🤖 "Behold, the future of AI: #Discrete #Distribution Networks! 🤯 Now accepted at ICLR 2025! 🎓✨ Watch in awe as GIF animations show off density estimations in shapes like spirals and QR codes, because that’s exactly what the world needs more of. 🎆🔍 Meanwhile, the masses eagerly await the life-altering impacts of blurcircles. 🙄🔄"
https://discrete-distribution-networks.github.io/ #AI #Networks #ICLR2025 #DensityEstimations #FutureOfAI #HackerNews #ngated

🎉🤖 "Behold, the future of AI: #Discrete #Distribution Networks! 🤯 Now accepted at ICLR 2025! 🎓✨ Watch in awe as GIF animations show off density estimations in shapes like spirals and QR codes, because that’s exactly what the world needs more of. 🎆🔍 Meanwhile, the masses eagerly await the life-altering impacts of blurcircles. 🙄🔄"
https://discrete-distribution-networks.github.io/ #AI #Networks #ICLR2025 #DensityEstimations #FutureOfAI #HackerNews #ngated

1/ Controlling LLMs with steering vectors is unreliable, but why?  Our paper, "Understanding (Un)Reliability of Steering Vectors in Language Models," at the #ICLR2025 Workshop on Foundation Models in the Wild investigates this! What did we find?

The Data Science and Machine Learning Unit at OIST is working to improve the algorithms of the information age and how we interact with AI💻⚙ They had 5 papers – 4 written by interns – accepted at the recent #ICLR2025. Learn more about the unit 👇 www.oist.jp/news-center/...

Learning at peak efficiency

An illustrated guide to automatic sparse differentiation

https://iclr-blogposts.github.io/2025/blog/sparse-autodiff/

#HackerNews #automaticdifferentiation #sparseautodiff #illustration #machinelearning #ICLR2025

An illustrated guide to automatic sparse differentiation

https://iclr-blogposts.github.io/2025/blog/sparse-autodiff/

#HackerNews #automaticdifferentiation #sparseautodiff #illustration #machinelearning #ICLR2025

Friends met at #ICLR2025 and then #AABI2025 😊

There is a Learning Theory Day on Tuesday, April 29 (the day after #ICLR2025) at NTU in Singapore. Featuring a great lineup of speakers! sites.google.com/view/learnin...

At #ICLR2025 main conference! (with updated information)

At #ICLR2025 main conference! (with updated information)

Attending ICLR this week, happy to meet for coffee/tea/etc.

On Saturday we have our poster for Expected Return Symmetries, see https://iclr.cc/virtual/2025/poster/27862

On Friday I'll be at the Microsoft booth for AI for Science coffee chat, see https://www.microsoft.com/en-us/research/event/microsoft-at-iclr-2025/booth-schedule/

See you there!

#ICLR2025

Attending ICLR this week, happy to meet for coffee/tea/etc.

On Saturday we have our poster for Expected Return Symmetries, see https://iclr.cc/virtual/2025/poster/27862

On Friday I'll be at the Microsoft booth for AI for Science coffee chat, see https://www.microsoft.com/en-us/research/event/microsoft-at-iclr-2025/booth-schedule/

See you there!

#ICLR2025

Consider following the authors Indraneil Paul, Haoyi Yang, Goran Glavaš, Kristian Kersting (@kerstingAIML) & Iryna Gurevych if you are interested in more information or an exchange of ideas. (7/7)

See you this week in Singapore! #ICLR2025

Consider following the authors Indraneil Paul, Haoyi Yang, Goran Glavaš, Kristian Kersting (@kerstingAIML) & Iryna Gurevych if you are interested in more information or an exchange of ideas. (7/7)

See you this week in Singapore! #ICLR2025

And consider following the authors
Nico Daheim, Clara Meister, Thomas Möllenhoff & Iryna Gurevych if you are interested in more information or an exchange of ideas. (5/5)

See you in Singapore! #ICLR2025

And consider following the authors
Nico Daheim, Clara Meister, Thomas Möllenhoff & Iryna Gurevych if you are interested in more information or an exchange of ideas. (5/5)

See you in Singapore! #ICLR2025

🔍 𝗟𝗟𝗠𝘀 𝗰𝗮𝗻 𝗽𝗿𝗼𝗱𝘂𝗰𝗲 𝗲𝗿𝗿𝗼𝗻𝗲𝗼𝘂𝘀 𝗼𝘂𝘁𝗽𝘂𝘁𝘀 𝗱𝘂𝗲 𝘁𝗼 (𝗲𝗽𝗶𝘀𝘁𝗲𝗺𝗶𝗰) 𝘂𝗻𝗰𝗲𝗿𝘁𝗮𝗶𝗻𝘁𝘆
🚀 To address this, we present decoding algorithms that consider uncertainty over model weights to improve LLM decoding using a Minimum Bayes Risk approach. #ICLR2025

(1/🧵)

🔍 𝗟𝗟𝗠𝘀 𝗰𝗮𝗻 𝗽𝗿𝗼𝗱𝘂𝗰𝗲 𝗲𝗿𝗿𝗼𝗻𝗲𝗼𝘂𝘀 𝗼𝘂𝘁𝗽𝘂𝘁𝘀 𝗱𝘂𝗲 𝘁𝗼 (𝗲𝗽𝗶𝘀𝘁𝗲𝗺𝗶𝗰) 𝘂𝗻𝗰𝗲𝗿𝘁𝗮𝗶𝗻𝘁𝘆
🚀 To address this, we present decoding algorithms that consider uncertainty over model weights to improve LLM decoding using a Minimum Bayes Risk approach. #ICLR2025

(1/🧵)

This year, our team is excited to share that four papers have been accepted to the Main Conference Proceedings of #NAACL2025 in Albuquerque 🌵 and three to #ICLR2025 in Singapore 🌏, with additional contributions featured in Findings at both conferences! 🚀📚

(1/🧵)

This year, our team is excited to share that four papers have been accepted to the Main Conference Proceedings of #NAACL2025 in Albuquerque 🌵 and three to #ICLR2025 in Singapore 🌏, with additional contributions featured in Findings at both conferences! 🚀📚

(1/🧵)

#TopoNets: High-performing vision and language models with brain-like topography toponets.github.io by @apurvaratan.bsky.social et al. #TopoLoss #ICLR2025 #neuroscience More information in the Twitter/X threads x.com/MainakDeb19/... and x.com/apurvaratan/...

#TopoNets: High-performing vision and language models with brain-like topography https://toponets.github.io by @apurvaratan et al. #TopoLoss #ICLR2025 #neuroscience

More information in the Twitter/X threads https://x.com/MainakDeb19/status/1884661749653397634 and https://x.com/apurvaratan/status/1884805634400330002

#TopoNets: High-performing vision and language models with brain-like topography https://toponets.github.io by @apurvaratan et al. #TopoLoss #ICLR2025 #neuroscience

More information in the Twitter/X threads https://x.com/MainakDeb19/status/1884661749653397634 and https://x.com/apurvaratan/status/1884805634400330002