1

Task Singular Vectors: Reducing Task Interference in Model Merging
Task Arithmetic has emerged as a simple yet effective method to merge models without additional training. However, by treating entire …
Cite arXiv GitHub
Task Singular Vectors: Reducing Task Interference in Model Merging
Naturalistic Music Decoding from EEG Data via Latent Diffusion Models
In this article, we explore the potential of using latent diffusion models, a family of powerful generative models, for the task of …
Cite arXiv
Naturalistic Music Decoding from EEG Data via Latent Diffusion Models
C2M3: Cycle-Consistent Multi-Model Merging
In this paper, we present a novel data-free method for merging neural networks in weight space. Differently from most existing works, …
Cite URL PDF arXiv GitHub
C2M3: Cycle-Consistent Multi-Model Merging
Latent Functional Maps: a spectral framework for representation alignment
Neural models learn data representations that lie on low-dimensional manifolds, yet modeling the relation between these …
Cite URL PDF arXiv
Latent Functional Maps: a spectral framework for representation alignment
ReMatching: Low-Resolution Representations for Scalable Shape Correspondence
We introduce ReMatching, a novel shape correspondence solution based on the functional maps framework. Our method, by exploiting a new …
Cite arXiv GitHub
ReMatching: Low-Resolution Representations for Scalable Shape Correspondence
Multi-Source Diffusion Models for Simultaneous Music Generation and Separation
In this work, we define a diffusion-based generative model capable of both music synthesis and source separation by learning the score …
Cite PDF URL arXiv GitHub ICLR 2024 notable top 1.2% (oral)
Multi-Source Diffusion Models for Simultaneous Music Generation and Separation
SyncFusion: Multimodal Onset-synchronized Video-to-Audio Foley Synthesis
Sound design involves creatively selecting, recording, and editing sound effects for various media like cinema, video games, and …
Cite PDF URL arXiv GitHub
SyncFusion: Multimodal Onset-synchronized Video-to-Audio Foley Synthesis
Generalized Multi-Source Inference for Text Conditioned Music Diffusion Models
Multi-Source Diffusion Models (MSDM) allow for compositional musical generation tasks: generating a set of coherent sources, creating …
Cite URL arXiv
Generalized Multi-Source Inference for Text Conditioned Music Diffusion Models
Camoscio: an Italian Instruction-tuned LLaMA
In recent years Large Language Models (LLMs) have increased the state of the art on several natural language processing tasks. However, …
Cite PDF GitHub Best Student Paper Award
Camoscio: an Italian Instruction-tuned LLaMA
Latent Space Translation via Semantic Alignment
Different neural models often exhibit similar latent spaces when exposed to semantically similar data; however, this inherent …
Cite PDF URL GitHub
Latent Space Translation via Semantic Alignment