Abstract
Representation Autoencoders (RAEs) generate images from pre-trained visual fea- tures, but their dense token grids make generative modeling expensive. Motivated by local feature correlations, we introduce PoolDINO, a learned affine pooling operator that merges neighboring tokens. Training the pooling operator jointly with the RGB decoder preserves the standard two-stage RAE procedure without a separate feature auto-encoder. On ImageNet-256, 4x token compression retains comparable generation quality under internal guidance, while 16x compression trades some quality for greater efficiency. At a fixed budget of 100 sampling steps, latent-sampling throughput increases by 3.7x and 9.0x, respectively, relative to the unpooled baseline. Classification and dense prediction evaluations show that comparable guided generation quality can coexist with weaker performance on other tasks.
Keywords
Subject
Publication details
- Journal
- Not available
- Open access
- Green open access
Cite this article
APA 7
Calvo-González, R., Saied, Y., & Fleuret, F. (2026). Pooling Representation Autoencoders for Efficient Diffusion. https://omanscience.com/en/articles/pooling-representation-autoencoders-for-efficient-diffusion
MLA 9
Calvo-González, Ramón, et al. "Pooling Representation Autoencoders for Efficient Diffusion." https://omanscience.com/en/articles/pooling-representation-autoencoders-for-efficient-diffusion.
Chicago (author–date)
Calvo-González, Ramón, Youssef Saied, and François Fleuret. 2026. "Pooling Representation Autoencoders for Efficient Diffusion." https://omanscience.com/en/articles/pooling-representation-autoencoders-for-efficient-diffusion.
Harvard
Calvo-González, R., Saied, Y. and Fleuret, F. (2026) 'Pooling Representation Autoencoders for Efficient Diffusion', Available at: https://omanscience.com/en/articles/pooling-representation-autoencoders-for-efficient-diffusion.
Vancouver
Calvo-González R, Saied Y, Fleuret F. Pooling Representation Autoencoders for Efficient Diffusion. https://omanscience.com/en/articles/pooling-representation-autoencoders-for-efficient-diffusion
IEEE
R. Calvo-González, Y. Saied, and F. Fleuret, "Pooling Representation Autoencoders for Efficient Diffusion," https://omanscience.com/en/articles/pooling-representation-autoencoders-for-efficient-diffusion.