George Cazenavette
@gcazenavette.bsky.social
Computer Vision PhD at MIT
georgecazenavette.github.io/
georgecazenavette.github.io/
This work was a lot of fun and was done in collaboration with my advisors Antonio Torralba and @vincentsitzmann.bsky.social. Please see our paper for more details and results.
We hope this work inspires future research in this area!
And please share your favorite images from our gallery!!
6/6
We hope this work inspires future research in this area!
And please share your favorite images from our gallery!!
6/6
November 21, 2025 at 4:01 AM
This work was a lot of fun and was done in collaboration with my advisors Antonio Torralba and @vincentsitzmann.bsky.social. Please see our paper for more details and results.
We hope this work inspires future research in this area!
And please share your favorite images from our gallery!!
6/6
We hope this work inspires future research in this area!
And please share your favorite images from our gallery!!
6/6
We distill images with different models, each yielding distinct styles that hint as to how these models "see."
Please see our gallery to browse all our images from many datasets (incl. ImageNet, Stanford Dogs, CUB 200, Flowers-102, Food-101): linear-gradient-matching.github.io/gallery/
5/6
Please see our gallery to browse all our images from many datasets (incl. ImageNet, Stanford Dogs, CUB 200, Flowers-102, Food-101): linear-gradient-matching.github.io/gallery/
5/6
November 21, 2025 at 4:01 AM
We distill images with different models, each yielding distinct styles that hint as to how these models "see."
Please see our gallery to browse all our images from many datasets (incl. ImageNet, Stanford Dogs, CUB 200, Flowers-102, Food-101): linear-gradient-matching.github.io/gallery/
5/6
Please see our gallery to browse all our images from many datasets (incl. ImageNet, Stanford Dogs, CUB 200, Flowers-102, Food-101): linear-gradient-matching.github.io/gallery/
5/6
Our method out-performs all real-image baselines on the standard ImageNet benchmarks and shines even brighter on datasets with fine grained classes!
The learned images seem to contain more discriminative features than any single real image, leading to a better classifier.
4/6
The learned images seem to contain more discriminative features than any single real image, leading to a better classifier.
4/6
November 21, 2025 at 4:01 AM
Our method out-performs all real-image baselines on the standard ImageNet benchmarks and shines even brighter on datasets with fine grained classes!
The learned images seem to contain more discriminative features than any single real image, leading to a better classifier.
4/6
The learned images seem to contain more discriminative features than any single real image, leading to a better classifier.
4/6
We directly learn our synthetic images such that they induce similar gradients as the real data when training a linear classifier.
Our meta loss is simply the distance between these gradients!
Critically, we also parameterize our images as pyramids as a form of implicit regularization.
3/6
Our meta loss is simply the distance between these gradients!
Critically, we also parameterize our images as pyramids as a form of implicit regularization.
3/6
November 21, 2025 at 4:01 AM
We directly learn our synthetic images such that they induce similar gradients as the real data when training a linear classifier.
Our meta loss is simply the distance between these gradients!
Critically, we also parameterize our images as pyramids as a form of implicit regularization.
3/6
Our meta loss is simply the distance between these gradients!
Critically, we also parameterize our images as pyramids as a form of implicit regularization.
3/6
While prior work focuses on distilling images to train models from scratch, this task becomes unreasonable with extremely small support sizes.
Instead, we focus on learning images to train *linear classifiers* on top of pre-trained models, a more relevant task in the era of foundation models.
2/6
Instead, we focus on learning images to train *linear classifiers* on top of pre-trained models, a more relevant task in the era of foundation models.
2/6
November 21, 2025 at 4:01 AM
While prior work focuses on distilling images to train models from scratch, this task becomes unreasonable with extremely small support sizes.
Instead, we focus on learning images to train *linear classifiers* on top of pre-trained models, a more relevant task in the era of foundation models.
2/6
Instead, we focus on learning images to train *linear classifiers* on top of pre-trained models, a more relevant task in the era of foundation models.
2/6
(tagging @csprofkgd.bsky.social @iccv.bsky.social for visibility)
May 22, 2025 at 8:08 PM
(tagging @csprofkgd.bsky.social @iccv.bsky.social for visibility)
We also have a fantastic lineup of invited speakers, including @sarameghanbeery.bsky.social (MIT), Alyosha Efros (Berkley), Olga Russakovsky (Princeton), and Zhuang Liu (Princeton).
We're looking forward to a great workshop and hope to see all of you there! 🌴
3/
We're looking forward to a great workshop and hope to see all of you there! 🌴
3/
bsky.app
May 22, 2025 at 8:08 PM
We also have a fantastic lineup of invited speakers, including @sarameghanbeery.bsky.social (MIT), Alyosha Efros (Berkley), Olga Russakovsky (Princeton), and Zhuang Liu (Princeton).
We're looking forward to a great workshop and hope to see all of you there! 🌴
3/
We're looking forward to a great workshop and hope to see all of you there! 🌴
3/
We're accepting both short (4 page) and long-form (8 page) papers on any topic related to data curation.
The deadline for long papers wishing to be published in the ICCV workshop proceedings is July 7, and all other submissions have until August 29.
Please reply or DM me with any question!
2/
The deadline for long papers wishing to be published in the ICCV workshop proceedings is July 7, and all other submissions have until August 29.
Please reply or DM me with any question!
2/
May 22, 2025 at 8:08 PM
We're accepting both short (4 page) and long-form (8 page) papers on any topic related to data curation.
The deadline for long papers wishing to be published in the ICCV workshop proceedings is July 7, and all other submissions have until August 29.
Please reply or DM me with any question!
2/
The deadline for long papers wishing to be published in the ICCV workshop proceedings is July 7, and all other submissions have until August 29.
Please reply or DM me with any question!
2/