Mikhail Yurochkin, Yuekai Sun, Subha Maity, Lilian Ngweta, Alex Gittens
We lifted 2 functions out of this paper's own repositories and ran 0 of them in a sandbox. "Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results.
| Repository | Role | Ran |
|---|---|---|
| lilianngweta/pisco | canonical | 0 of 2 |
| Function | Status | Where it lives |
|---|---|---|
| Img2Vec | Not yet run | lilianngweta/pisco/experiments/feature_extractors/ImageNet_ResNet50_feature_extractor_stylized.py pointer only (licence: NONE) · get_code("b2dedbe21e0b9076") |
| get_features | Not yet run | lilianngweta/pisco/experiments/feature_extractors/ImageNet_ResNet50_feature_extractor_stylized.py pointer only (licence: NONE) · get_code("14e9567a9e02e3a6") |
Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.
Learning visual representations with interpretable features, i.e., disentangled representations, remains a challenging problem. Existing methods demonstrate some success but are hard to apply to large-scale vision datasets like ImageNet. In this work, we propose a simple post-processing framework to disentangle content and style in learned representations from pre-trained vision models. We model the pre-trained features probabilistically as linearly entangled combinations of the latent content and style factors and develop a simple disentanglement algorithm based on the probabilistic model. We show that the method provably disentangles content and style features and verify its efficacy empirically. Our post-processed features yield significant domain generalization performance improvements when the distribution shift occurs due to style changes or style-related spurious correlations.
The same record, over MCP at https://syntology.ai/mcp:
get_harvested_code_for_paper("2302.09795")
get_code_for_paper("2302.09795")
have("2302.09795")
Connect an agent — have() is free.