Lightnews — Scholar-powered news

Paul Vicol @paulvicol.bsky.social · 15d

🚀 All this and more in our paper!

arXiv: arxiv.org/abs/2509.20328

By
@thwiedemer.bsky.social, Yuxuan Li, @paulvicol.bsky.social, @nmatares.bsky.social, Been Kim, Kevin Swersky, Priyank Jaini, and Robert Geirhos.

1 2

Paul Vicol @paulvicol.bsky.social · 15d

Veo 3 understands material properties, including buoyancy, reflections, flammability, and soft body physics.

1 2

Paul Vicol @paulvicol.bsky.social · 15d

Veo 3 can understand which objects fit in a backpack, and can categorize objects (putting all the toys in a bucket). Veo 3 can also draw, and understands the effects of gravity and air resistance on falling objects.

1 2

Paul Vicol @paulvicol.bsky.social · 15d

Veo 3 can reason, filling in the next element in a sequence of images.

1 2

Paul Vicol @paulvicol.bsky.social · 15d

🔥Veo 3 has emergent zero-shot learning and reasoning capabilities!

This multitalented model can do a huge range of interesting tasks.

It understands physical properties, can manipulate objects, and can even reason.

arxiv.org/abs/2509.20328
video-zero-shot.github.io

Examples in this thread!

1 3 9

Paul Vicol @paulvicol.bsky.social · Dec 19

⚙️ Tong Chen presented "Generative Adapter: Contextualizing Language Models in Parameters with a Single Forward Pass"

openreview.net/forum?id=fk5...

Tong Chen, Hao Fang, Patrick Xia, Xiaodong Liu, Benjamin Van Durme, @lukezettlemoyer.bsky.social, Jianfeng Gao, Hao Cheng

#NeurIPS2024

1

Paul Vicol @paulvicol.bsky.social · Dec 19

📰 Amelia Hui Dai presented "Are LLMs Prescient? A Continuous Evaluation using Daily News as the Oracle"

openreview.net/forum?id=cgi...

Hui Dai, Ryan Teehan, Mengye Ren

#NeurIPS2024 #AdaptiveFoundationModels

1

Paul Vicol @paulvicol.bsky.social · Dec 19

👀 Wen-Tse Chen presented "Fine-tuning LLM Agents with Retrospective In-Context Online Learning"

openreview.net/forum?id=dSZ...

Wen-Tse Chen, Jiayu Chen, Fahim Tajwar, Hao Zhu, Xintong Duan, @rsalakhu.bsky.social, Jeff Schneider

#NeurIPS2024 #AdaptiveFoundationModels

1

Paul Vicol @paulvicol.bsky.social · Dec 19

🏃 Zhepei Wei presented "Fast and Accurate Language Model Decoding via Parallel Token Processing"

openreview.net/forum?id=coZ...

@zhepei.bsky.social @wlchen.bsky.social, Xinyu Zhu, Yu Meng

#NeurIPS2024 #AdaptiveFoundationModels

1 2 3

Paul Vicol @paulvicol.bsky.social · Dec 19

▶️ Yue Wu presented "Self-Play Preference Optimization for Language Model Alignment"

openreview.net/forum?id=Z1P...

Yue Wu, Zhiqing Sun, @angelahzyuan.bsky.social, Kaixuan Ji, Yiming Yang, @quanquangu.bsky.social

#NeurIPS2024 #AdaptiveFoundationModels

1 1

Paul Vicol @paulvicol.bsky.social · Dec 19

🎾 Jennifer Hsia presented "RAGGED: Towards Informed Design of Retrieval Augmented Generation Systems"

openreview.net/forum?id=SX1...

Jennifer Hsia, Afreen Shaikh, Zhiruo Wang, @gneubig.bsky.social

#NeurIPS2024 #AdaptiveFoundationModels

1

Paul Vicol @paulvicol.bsky.social · Dec 19

🥣 Seungone Kim presented "Personalized Soups"

openreview.net/forum?id=EMr...

Joel Jang, Seungone Kim, @billyuchenlin.bsky.social @yizhongw.bsky.social @jmhessel.bsky.social @lukezettlemoyer.bsky.social @hanna-nlp.bsky.social @yejinchoinka.bsky.social @rajammanabrolu.bsky.social

#NeurIPS2024

1 1

Paul Vicol @paulvicol.bsky.social · Dec 19

💎 Lukas Hauzenberger presented "One Initialization to Rule them All: Fine-tuning via Explained Variance Adaptation"

openreview.net/forum?id=tjZ...

Fabian Paischer, Lukas Hauzenberger, Thomas Schmied, Benedikt Alkin, Marc Peter Deisenroth, Sepp Hochreiter

#NeurIPS2024 #AdaptiveFoundationModels

1

Paul Vicol @paulvicol.bsky.social · Dec 19

🖨️ We also had two exciting poster sessions.

👕 Some presenters already put on their Adaptive Foundation Models T-shirts! Thanks Jianan Zhao Zhaocheng Zhu!

🧵 Oral paper presentations next!

#NeurIPS2024 #AdaptiveFoundationModels

1

Paul Vicol @paulvicol.bsky.social · Dec 19

☁️ Tianqi Chen from CMU (@scsatcmu.bsky.social) gave a very interesting talk on Enabling Large Language Model Deployment Across Cloud and Edge with ML Compilation.

Timestamp 7:48:30 in neurips.cc/virtual/2024...

#NeurIPS2024 #AdaptiveFoundationModels

1

Paul Vicol @paulvicol.bsky.social · Dec 19

🪐 Bing Liu from the University of Illinois at Chicago gave a great talk on Continual Learning with Large Language Models.

Timestamp 7:08:17 in neurips.cc/virtual/2024...

#NeurIPS2024 #AdaptiveFoundationModels

1 1

Paul Vicol @paulvicol.bsky.social · Dec 19

⚡ Minjoon Seo from KAIST AI gave a great talk On the Knowledge Adaptability of Language Models.

Timestamp 5:42:00 in neurips.cc/virtual/2024...

#NeurIPS2024 #AdaptiveFoundationModels

1 1

Paul Vicol @paulvicol.bsky.social · Dec 19

♾️ Matthias Bethge from University of Tuebingen (@unituebingen.bsky.social) had a talk on Continual Foundation Model Learning, presented by his amazing PhD student, Vishaal Udandarao (@vishaalurao.bsky.social).

Timestamp 5:06:39 in neurips.cc/virtual/2024...

#NeurIPS2024 #AdaptiveFoundationModels

1

Paul Vicol @paulvicol.bsky.social · Dec 19

🔑 Kate Saenko from Meta and Boston University (@bostonu.bsky.social) gave a very interesting talk titled "Is pretraining the key to successful domain generalization?"

Timestamp 1:55:39 in neurips.cc/virtual/2024...

📎 arxiv.org/abs/2412.02856

#NeurIPS2024 #AdaptiveFoundationModels

1 1

Paul Vicol @paulvicol.bsky.social · Dec 19

💡 Sander Dieleman (@sedielem.bsky.social) from Google DeepMind gave a thought-provoking talk on Multimodal Iterative Refinement.

Timestamp 1:10:45 in neurips.cc/virtual/2024...

#NeurIPS2024 #AdaptiveFoundationModels

1

Paul Vicol @paulvicol.bsky.social · Dec 19

🌲 Ruslan Salakhutdinov (@rsalakhu.bsky.social) from CMU (@scsatcmu.bsky.social) opened the workshop with a talk on Tree Search for Language Model Agents.

Timestamp 36:20 in neurips.cc/virtual/2024...

📎 arxiv.org/abs/2407.01476

#NeurIPS2024 #AdaptiveFoundationModels

1 1 1

Paul Vicol @paulvicol.bsky.social · Dec 19

Our workshop in numbers:
🖇️ 128 Papers
💬 8 Orals
🖋️ 564 Authors
✅ 40 Reviewers
🔊 7 Invited Speakers
👕 100 T-Shirts

🔥 Organizers: Paul Vicol, Mengye Ren, Renjie Liao, Naila Murray, Wei-Chiu Ma, Beidi Chen

#NeurIPS2024 #AdaptiveFoundationModels

1 1 1

Paul Vicol @paulvicol.bsky.social · Dec 19

🎉 Had fun at #NeurIPS2024 Workshop on #AdaptiveFoundationModels!

🚀 Speakers: @rsalakhu.bsky.social @sedielem.bsky.social Kate Saenko, Matthias Bethge / @vishaalurao.bsky.social Minjoon Seo, Bing Liu, Tianqi Chen

🌐Posters: adaptive-foundation-models.org/papers

🎬 neurips.cc/virtual/2024...

🧵Recap!

1 2 9

Reposted by Paul Vicol

Sander Dieleman @sedielem.bsky.social · Dec 18

The recording of my #NeurIPS2024 workshop talk on multimodal iterative refinement is now available to everyone who registered: neurips.cc/virtual/2024...

My talk starts at 1:10:45 into the recording.

I believe this will be made publicly available eventually, but I'm not sure when exactly!

1 4 36

Paul Vicol @paulvicol.bsky.social · Dec 15

Afternoon poster session at the Adaptive Foundation Models workshop!

2