Walter Simoncini

Computer Vision & Efficiency
PhD Student @ University of Amsterdam

[Dheeraj Varghese @dhevarghese]

Ever wondered whether AI agents could play thousands of video games on a supercomputer? We did. 🎮

Introducing PlaySuite: 5,734 games, 14 open models, and a large-scale benchmark for interactive visual intelligence.

Here's what we built, and what we learned. 🧵👇

[Pedro @pmpcurvo]

♾️ LiFT: Loop Flow Transformers 🧵👇

Diffusion models usually spend more inference compute by taking more sampling steps. What if we could also spend compute inside each step?

LiFT turns part of a DiT into a recurrent core, letting us scale depth at inference with only a small architectural change. On ImageNet, it beats dense DiT-XL/2 with 3.34 lower FID, while using 60% fewer parameters, 32% fewer training FLOPs, and 52% fewer inference FLOPs.

Jointly led with Mohammad Mahdi Derakhshani 👑(@mmderakhshani)

And amazing supervisors Cees Snoek (@cgmsnoek), Jan-Willem van de Meent (@jwvdm) and Gertjan J. Burghouts (@gjburghouts)

[Floor Eijkelboom @FEijkelboom]

The flow-based language model agenda continues 🌊

Amazing work by @ayhozade showing how models can learn to reason through iterative denoising.

[Yuki @y_m_asano]

[🔁🚨PhD opening]

We have a new opening for a PhD student at our @FunAILab at the University of Technology Nuremberg on combining Vision-Language-Action Models with 3D Geometry.

This PhD position is in close collaboration with @HelsingAI with the amazing researcher @TomUnterthiner and will be a unique opportunity to conduct research and have impact.

Please see below for more information:
https://t.co/9FU8ryPTPN

We particularly encourage applications from people of diverse and underrepresented backgrounds. If you have any doubt if you should apply or not, send me a message.

[Daniil Cherniavskii @Cherniavskii_D]

New preprint: Context-weighted Discrete Flow Matching.

Paper: https://t.co/iHokRakAtG

We study how local context relates to token predictability in discrete flow and diffusion models -- and how it can be used to improve sampling.

🧵

[Ioana @ioanasimioni]

(1/n) What if you could teach a vision model about 3D structure without a single 3D label and just a tiny fraction of its pretraining compute? 👁️ That’s 3DPoV. ✨ Heading to #ICML2026 in Seoul 🇰🇷 to share it 👇

[Yuki @y_m_asano]

🎉[openings] I’m hiring postdoctoral researchers to join our @FunAILab at UTN through the Alexander von Humboldt Research Fellowship (@AvHStiftung), via the Henriette Herz Scouting Programme.

As a Henriette Herz Scout, I can nominate outstanding international researchers for this fellowship route. I’m especially keen to hear from candidates working on multimodal learning, video and image pretraining, and post-training.

Fellows would be hosted in our lab at UTN and work closely with us on these topics.

Key requirements:
* finished your doctoral studies less than 4 years ago or will finish in the next 6 months
* did not live/work in Germany in the last 10 years
* applications from female, trans* and/or non-binary candidates are highly encouraged!

Interested? Please send a short note with your CV, PhD year, current affiliation, 2–3 key publications, and a few lines on how your work connects.

Please share! 🔀

[Jona Ruthardt @jonaruthardt]

Ever asked yourself: 𝙃𝙤𝙬 𝙙𝙤 𝙩𝙚𝙭𝙩-𝙤𝙣𝙡𝙮 𝙇𝙇𝙈𝙨 "𝙨𝙚𝙚" 𝙩𝙝𝙚 𝙫𝙞𝙨𝙪𝙖𝙡 𝙬𝙤𝙧𝙡𝙙?

Our new TMLR paper finds that language capability predicts visual alignment of LLM features (r=0.768).

This means: LLM progress boosts vision(-language) models as well.

🧵⬇️

[Ivona Najdenkoska @ivonajdenkoska]

I’m excited to announce that our workshop “Multimodal Intelligence: Next Token Prediction & Beyond” will take place at ICLR 2026!

📅 April 2026
📍Rio de Janeiro, Brazil
🔗https://t.co/rk77hbL1dd

[Yuki @y_m_asano]

[new paper]: MLLMs can understand both images and text, but how consistent are they when given the *same* content?
We introduce REST and REST+: benchmarks for evaluating cross-modal inconsistency in multimodal LLMs (MLLMs). 1/

[Walter Simoncini @iWSimoncini]

We’ll be presenting our paper today at 4:30 PST, come say hi at poster 4709!

[Walter Simoncini @iWSimoncini]

This week I'll be at #NeurIPS presenting our latest work "Elastic ViTs from Pretrained Models without Retraining"!

TL;DR: You can make pretrained ViTs elastic in under 5 minutes on a single A100 via structured pruning, without re-training!

Photo: in-context semantic segmentation with DINOv1 ViT-B/16 models pruned from 0 to 50% sparsity.

1/🧵

[Andrei Bursuc @abursuc]

Hey #NeurIPS2025! We have a fun deck of works with the team and friends to present this year. Check them out!

[Walter Simoncini @iWSimoncini]

@TheMattGacek Let’s do it! DM me and we’ll find a spot and time :)

[Walter Simoncini @iWSimoncini]

This week I'll be at #NeurIPS presenting our latest work "Elastic ViTs from Pretrained Models without Retraining"!

TL;DR: You can make pretrained ViTs elastic in under 5 minutes on a single A100 via structured pruning, without re-training!

Photo: in-context semantic segmentation with DINOv1 ViT-B/16 models pruned from 0 to 50% sparsity.

1/🧵

[Piyush Bagad @bagad_piyush]

Excited to present our work @NeurIPSConf in San Diego! LiFT is a simple method to obtain time-aware⌛️video embeddings that can distinguish between “chiral actions”, e.g., opening a door vs. closing it, or climbing up vs down a ladder. Visit our poster on 3 Dec 4:30-7:30 p.m.

[Walter Simoncini @iWSimoncini]

I'll be presenting this work in Exhibit Hall C,D,E between 4:30 p.m. and 7:30 p.m. PST on Thursday, the 4th. Look for poster 4709!

/🧵

[Walter Simoncini @iWSimoncini]

This was a super fun collaboration with @mdorkenw (equal contribution), @TiRune , @cgmsnoek, and @y_m_asano!

Website: https://t.co/0PFDKL7oi1
Paper: https://t.co/qF8JYW9EK8
Code: https://t.co/sZICDOaOZW

12/🧵