5 papers
When One Adapter Speaks for Many: Discovering Low-Rank Redundancy in Continual Fine-Tuning
Tanguy Dieudonné, Giulia Lanzillotta, Enis Simsar +2
Low-Rank Adaptation (LoRA) has become the standard tool for parameter-efficient fine-tuning of large pretrained models. When applied sequentially across tasks in Continual Learning…
SeqLoRA: Bilevel Orthogonal Adaptation for Continual Multi-Concept Generation
Javad Parsa, Enis Simsar, Amir Joudaki +2
Parameter-efficient fine-tuning enables fast personalization of text-to-image diffusion models, but composing multiple custom concepts remains challenging due to representation int…
FullFlow: Upgrading Text-to-Image Flow Matching Models for Bidirectional Vision--Language Generation
Eric Tillmann Bill, Enis Simsar, Alessio Tonioni +1
Modern text-to-image diffusion models encode rich visual priors, but expose them only through one-way text-conditioned generation. Existing unified vision--language models derived…
Flash PD-SSM: Memory-Optimized Structured Sparse State-Space Models
Aleksandar TerziÄ, Francesco Carzaniga, Nicolas Menet +4
State-space models (SSMs) face a fundamental trade-off between efficiency and expressivity that is mainly dictated by the structure of the model's transition matrix. Unstructured t…
Barriers for Learning in an Evolving World: Mathematical Understanding of Loss of Plasticity
Amir Joudaki, Giulia Lanzillotta, Mohammad Samragh Razlighi +5
Deep learning models excel in stationary data but struggle in non-stationary environments due to a phenomenon known as loss of plasticity (LoP), the degradation of their ability to…