Community
DeepFMI 2026 project posters
Final project posters from the DeepFMI 2026 edition are now online! View the gallery.
News
All team updates, publications, blog posts and more in one place!
Links: arXiv
Abstract Learning from a sequence of interactions, as soon as observations are perceived and acted upon, without explicitly storing them, …
Academic
Generalization & Interpretability
Fine-Tuning Regimes Define Distinct Continual Learning Problems
Links: arXiv
Abstract Continual learning (CL) studies how models acquire tasks sequentially while retaining previously learned knowledge. Despite …
Academic
Generalization & Interpretability
Temporal Taskification in Streaming Continual Learning: A Source of Evaluation Instability
Links: arXiv
Abstract Streaming Continual Learning (CL) typically converts a continuous stream into a sequence of discrete tasks through temporal …
Academic
Natural Language Processing
JumpLoRA: Sparse Adapters for Continual Learning in Large Language Models
Links: arXiv
Abstract Adapter-based methods have become a cost-effective approach to continual learning (CL) for Large Language Models (LLMs), by …
Academic
Generalization & Interpretability
Bridging Explainability and Embeddings: BEE Aware of Spuriousness
Links: OpenReview GitHub
Abstract Current methods for detecting spurious correlations rely on data splits or error patterns, leaving many harmful …
Introduction Fine tuning is usually framed as “adaptation”. In practice, it can also manufacture shortcuts.
A model can recognize the “right” object …
Community
Deep Learning @ FMI 2026 edition
The Deep Learning course at the Faculty of Mathematics and Computer Science, University of Bucharest is running again in 2026. It is aimed at …
Academic
Deepfake Detection
Investigating self-supervised representations for audio-visual deepfake detection
Links: arXiv GitHub
Abstract Self-supervised representations excel at many vision and speech tasks, but their potential for audio-visual deepfake …
Academic
Natural Language Processing
CLewR: Curriculum Learning with Restarts for Machine Translation Preference Learning
Links: arXiv GitHub
Abstract Large language models (LLMs) have demonstrated competitive performance in zero-shot multilingual machine translation …