Yael Vinker🎗

715 posts

Yael Vinker🎗 banner
Yael Vinker🎗

Yael Vinker🎗

@YVinker

Postdoctoral Associate at @MIT_csail

Boston เข้าร่วม Temmuz 2021
355 กำลังติดตาม1.5K ผู้ติดตาม
ทวีตที่ปักหมุด
Yael Vinker🎗
Yael Vinker🎗@YVinker·
Excited to share that I officially completed my PhD in Computer Science at Tel Aviv University and have joined Prof. Antonio Torralba’s lab at MIT for my postdoc! Looking forward to researching visual communication, and surviving Boston winters without turning into a popsicle!❄️
Yael Vinker🎗 tweet media
English
46
14
716
36.1K
Yael Vinker🎗 รีทวีตแล้ว
Shivam Duggal
Shivam Duggal@ShivamDuggal4·
Similar thought. Next-token prediction feels statistical: perplexity / shannon-entropy minimization. But creativity / science may require: finding compact generative structures, then exploring in that space. Closer to algorithmic complexity? More Kolmogorov than Shannon.
Andrew Gordon Wilson@andrewgwils

Being good at next word prediction is the opposite of what we want for creativity, for scientific breakthroughs.

English
3
7
68
8.8K
Markus Wulfmeier
Markus Wulfmeier@m_wulfmeier·
Physically embodying AI is its own art form! There's something really cool that happens when it works out of the box (via tool calling) and you don't need to finetune for your model. Model gets better, robot gets better! Here is an initial run from @patricktresset just switching to @GoogleDeepMind #Gemini 3.1. Now I'm still not sure what better means in art or even how to measure it, but the emergency behavior is fascinating!
𝙿𝚊𝚝𝚛𝚒𝚌𝚔 𝚃𝚛𝚎𝚜𝚜𝚎𝚝@patricktresset

A little Saturday morning experiment with my prototyping robot... Greed

English
2
0
8
1.4K
Yael Vinker🎗 รีทวีตแล้ว
MIT CSAIL
MIT CSAIL@MIT_CSAIL·
Reaching scientific goals: expectation vs. reality, v/Florian Aigner.
MIT CSAIL tweet media
English
4
84
439
26.6K
Yael Vinker🎗 รีทวีตแล้ว
Cliff Pickover
Cliff Pickover@pickover·
Physics. Mathematics. Reality. Universe. Stephen Wolfram's "Theory of Everything" is free on line, as a 448-page PDF. Here is a sample figure, to intrigue you. Source: arxiv.org/abs/2004.08210
Cliff Pickover tweet media
English
48
477
2.4K
106.4K
Yoko
Yoko@stuffyokodraws·
The @QuiverAI team is truly epic. They move extremely fast, do frontier research, and most importantly they are incredibly kind people and fun to work with. DM Joan if you are interested in the problem space!
Joan Rodriguez@joanrod_ai

Come build with us at @QuiverAI! We’re assembling a world-class team of researchers and engineers to push the frontier of AI for vector design and beyond. If you’re excited about building models and products for visual code generation and the future of design tools, this is for you.

English
2
4
29
5.8K
Hila Chefer
Hila Chefer@hila_chefer·
New research from @bfl_ml 🥳 Meet Self-Flow: our self-supervised framework for image, audio, video & world models 🤖 bfl.ai/research/self-… Do generative models really need DINO to learn strong representations? We propose teaching them directly via a joint framework instead 🧵
Hila Chefer tweet media
English
11
61
271
56.9K
Haochen Shi
Haochen Shi@HaochenShi74·
Excited to release Minimalist Compliance Control! We achieve robust, compliant robot interaction across robot arms, dexterous hands, and humanoids, with NO force sensors or learning. If you’re wondering what remains, please see the thread below😉 Website: …nimalist-compliance-control.github.io
English
9
49
268
28.6K
Yael Vinker🎗 รีทวีตแล้ว
Joel Simon
Joel Simon@_joelsimon·
New essay: Beyond Slop Reflections on 6 years making creative AI tools and why the field feels stuck on content production rather than personal expression. We're building molds with dials rather than new clay. On embracing craft, process, and open-ended mediums (with some new demos!) joelsimon.net/beyond-slop
English
10
25
145
10.5K
Reve
Reve@reve·
We redesigned Reve. Editing is easier than ever. Now all your tools are in one panel on the right. And you can make changes to anything just by clicking directly and prompting.
English
17
31
338
371.7K
Yael Vinker🎗 รีทวีตแล้ว
Judy Fan
Judy Fan@judyefan·
This collaboration with @parastooabtahi & colleagues has been such a treat! Super excited about this line of work on how people use both words and gestures to coordinate on shared understanding of multi-step assembly tasks—check it out at #CHI2026!
Parastoo Abtahi@parastooabtahi

If you saw @judyefan present our poster at #CogSci2025, the full paper will be presented at #CHI2026: “Gesturing Toward Abstraction: Multimodal Convention Formation in Collaborative Physical Tasks” Paper, dataset, data viewing app, analysis code, and simulation code: 🔗 multimodal-conventions.github.io
📄 arxiv.org/pdf/2602.08914 @PrincetonHCI @PrincetonCS @StanfordPsych 3/3

English
0
2
27
2.8K
Yael Vinker🎗 รีทวีตแล้ว
Vincent Sitzmann
Vincent Sitzmann@vincesitzmann·
In my recent blog post, I argue that "vision" is only well-defined as part of perception-action loops, and that the conventional view of computer vision - mapping imagery to intermediate representations (3D, flow, segmentation...) is about to go away. vincentsitzmann.com/blog/bitter_le…
English
43
157
1K
365.6K
Yael Vinker🎗 รีทวีตแล้ว
Runway
Runway@runwayml·
World models are the most transformative technology of our time. Our mission at Runway is to accelerate their development and ensure they have a positive impact on the world. Today we're announcing $315 million in Series E funding to help advance this work. Learn more at the link below.
Runway tweet media
English
21
55
238
77.2K
Yael Vinker🎗 รีทวีตแล้ว
Amir Bar
Amir Bar@_amirbar·
Planning is one of the most exciting uses of world models, but existing planners struggle on long horizons. Introducing GRASP: a fast gradient-based planner for world models that outperforms prior methods on long-horizon tasks. Two key ideas: 1.jointly optimize actions and intermediate subgoals 2.inject stochasticity to explore subgoal space Project page: michaelpsenka.io/grasp/ Paper: arxiv.org/abs/2602.00475 w/ @michaelpsenka @mikerabb @ask1729 and @ylecun
GIF
Michael Psenka@michaelpsenka

tl;dr New planner for world models! GRASP: gradient-based, stochastic, parallelized. Long range planning for world models has always been an issue. 0th order methods like CEM/MPPI dominate, but have degrading performance at longer contexts or higher-dimensional actions. We wanted to address this from the ground up. w/ Michael Rabbat, @ask1729 , @ylecun*, @_amirbar* (equally advised)

English
11
66
558
74.3K
Dan Haramati
Dan Haramati@DanHrmti·
Learning accurate World Models for long horizon planning is hard. So what minimal aspect of world dynamics must a model capture to achieve complex goals? We find a simple and effective solution in our #ICLR2026 paper, which we will present as an Oral at @worldmodel_26. (1/n)
English
6
46
283
23.6K
Yael Vinker🎗 รีทวีตแล้ว
idan shenfeld
idan shenfeld@IdanShenfeld·
People keep saying 2026 will be the year of continual learning. But there are still major technical challenges to making it a reality. Today we take the next step towards that goal — a new on-policy learning algorithm, suitable for continual learning! (1/n)
idan shenfeld tweet media
English
45
208
1.4K
191.4K

ค้นพบ

@Barium1022 @m_wulfmeier @patricktresset @GoogleDeepMind @arxiv @GuyTvt @stuffyokodraws @QuiverAI