4 citations · 8 across the 5 of their papers we have counts for
7 papers
On Trajectory-Aware Training for Masked Diffusion Language Models
Manuel Madeira, Amitis Shidani, Alice Bizeul +6
Masked diffusion models (MDMs) generate text by unmasking several tokens per step, but they are trained and sampled under different conditions. The model is trained on randomly mas…
Limits of Confidence in Diffusion
Russ Webb, Amitis Shidani, Alice Bizeul +1
Discrete diffusion, including remasking and uniform-state samplers, generate a sequence by writing multiple token positions per step, drawing each from a per-position distribution…
Interpretable Diffusion Models with B-cos Networks
Nicola Bernold, Moritz Vandenhirtz, Alice Bizeul +1
Text-to-image diffusion models generate images by iteratively denoising random noise, conditioned on a prompt. While these models have enabled impressive progress in image generati…
From Pixels to Components: Eigenvector Masking for Visual Representation Learning
Alice Bizeul, Thomas Sutter, Alain Ryser +3
Predicting masked from visible parts of an image is a powerful self-supervised approach for visual representation learning. However, the common practice of masking random patches o…
Cross-Entropy Is All You Need To Invert the Data Generating Process
Patrik Reizinger, Alice Bizeul, Attila Juhos +4
Supervised learning has become a cornerstone of modern machine learning, yet a comprehensive theory explaining its effectiveness remains elusive. Empirical phenomena, such as neura…
A Probabilistic Model Behind Self-Supervised Learning
Alice Bizeul, Bernhard Schölkopf, Carl Allen
In self-supervised learning (SSL), representations are learned via an auxiliary task without annotated labels. A common task is to classify augmentations or different modalities of…
Identifiability Results for Multimodal Contrastive Learning
Imant Daunhawer, Alice Bizeul, Emanuele Palumbo +2
Contrastive learning is a cornerstone underlying recent progress in multi-view and multimodal learning, e.g., in representation learning with image/caption pairs. While its effecti…