David Charatan (@DavidCharatan) profile picture

CS PhD Student at MIT

Cambridge, MA
Joined September 2016
David Charatan retweeted
Can use just 10 images to teach a model what a sea anemone looks like? In our #neurips2026 paper, "Dataset Distillation via Drifting," we distill ImageNet's 1,300 images per class down to just 10, optimized from noise. drifting-distillation.github… More info in thread! ⬇️ 1/N
28
67
12
614
68,491
David Charatan retweeted
If video models scale like LLMs, they'll soon generate a perfect video plan for any embodied task. Do we need anything else? Introducing Mirror, which learns to follow these plans just by practicing, and even crafts an iron pickaxe in Minecraft w/o rewards or action labels! (1/n)
9
44
14
395
54,369
David Charatan retweeted
We discovered that our latest Dataset Distillation project can be used to create some beautiful synthetic images based on an artist's body of work! Come see us at the @eccvconf Art Gallery starting today! Explanation and some of my favorites in thread below: 1/ (Claude Monet)
5
17
4
92
26,427
David Charatan retweeted
Introducing MilliVid, our new method for long-context video generation! MilliVid creates videos that are consistent over long time spans, without using retrieval heuristics or 3D maps! (1/n) davidcharatan.com/millivid/#
11
72
7
438
75,905
Happy to finally share our latest work on Dataset Distillation! "Dataset Distillation for Pre-Trained Self-Supervised Vision Models," set to appear at #NeurIPS 2025! We learn 1 image per class to train linear heads for pre-trained models. linear-gradient-matching.git… 1/6
1
9
2
32
4,619
David Charatan retweeted
Now in Nature! 🚀 Our method learns a controllable 3D model of any robot from vision, enabling single-camera closed-loop control at test time! This includes robots previously uncontrollable, soft, and bio-inspired, potentially lowering the barrier of entry to automation! Paper: nature.com/articles/s41586-0… (1/n)
5
68
11
423
100,554
David Charatan retweeted
Introducing Diffusion Forcing, which unifies next-token prediction (eg LLMs) and full-seq. diffusion (eg SORA)! It offers improved performance & new sampling strategies in vision and robotics, such as stable, infinite video generation, better diffusion planning, and more! (1/8)
13
209
25
1,346
208,995
Announcing (fashionably late) our #CVPR2024 paper on GenAI detection! with @AvneeshSud, Thomas Leung, and @b_usmn Fake-Inversion: Learning to Detect Images from Unseen Models by Inverting Stable Diffusion fake-inversion.github.io/ More info in 🧵 1/7
1
8
1
17
2,587
Introducing “FlowMap”, the first self-supervised, differentiable structure-from-motion method that is competitive with conventional SfM like Colmap! cameronosmith.github.io/flow… IMO this solves a major missing piece for internet-scale training of 3D Deep Learning methods. 1/n
11
100
13
605
128,746
Announcing the 1st Workshop on Dataset Distillation @CVPR 2024! #CVPR2024 sites.google.com/view/dd-cvp… More info in 🧵 1/4
2
27
4
107
41,935
Introducing pixelSplat: feed-forward Gaussian splats from image pairs! Led by @DavidCharatan and @sizhe_lester_li, collaborating with @taiyasaki! We propose a memory-efficient, fast and editable alternative to pixelNeRF based on 3D Gaussian Splatting! davidcharatan.com/pixelsplat… 1/n
8
45
6
268
47,720