Rafael Elberg

Rafael Elberg, Denis Parra, Mircea Petrache

Image and multimodal machine learning tasks are very challenging to solve in the case of poorly distributed data. In particular, data availability and privacy restrictions exacerbate these hurdles in the medical domain. The state of the art in image generation quality is held by Latent Diffusion models, making them prime candidates for tackling this problem. However, a few key issues still need to be solved, such as the difficulty in generating data from under-represented classes and a slow inference process. To mitigate these issues, we propose a new method for image augmentation in long-tailed data based on leveraging the rich latent space of pre-trained Stable Diffusion Models. We create a modified separable latent space to mix head and tail class examples. We build this space via Iterated Learning of underlying sparsified embeddings, which we apply to task-specific saliency maps via a K-NN approach. Code is available at https://github.com/SugarFreeManatee/Feature-Space-Augmentation-and-Iterated-Learning

Rafael Elberg, Felipe del Rio, Mircea Petrache et al.

Compositional generalization--the ability to interpret novel combinations of familiar elements--is a hallmark of human cognition and language. Despite recent advances, deep neural networks still struggle to acquire this property reliably. In this work, we introduce CELEBI (Compressive-Expressive Language Emergence through a discrete Bottleneck and Iterated learning), a novel self-supervised framework for inducing compositionality in learned representations from pre-trained models, through a reconstruction-based communication game between a sender and a receiver. Building on theories of language emergence, we integrate three mechanisms that jointly promote compressibility, expressivity, and efficiency in the emergent language. First, interactive decoding incentivizes intermediate reasoning by requiring the receiver to produce partial reconstructions after each symbol. Second, a reconstruction-based imitation phase, inspired by iterated learning, trains successive generations of agents to imitate reconstructions rather than messages, enforcing a tighter communication bottleneck. Third, pairwise distance maximization regularizes message diversity by encouraging high distances between messages, with formal links to entropy maximization. Our method significantly improves both the efficiency and compositionality of the learned messages on the Shapes3D and MPI3D datasets, surpassing prior discrete communication frameworks in both reconstruction accuracy and topographic similarity. This work provides new theoretical and empirical evidence for the emergence of structured, generalizable communication protocols from simplicity-based inductive biases.