Zobrazeno 1 - 10
of 292
pro vyhledávání: '"Scardapane, Simone"'
Spatiotemporal graph neural networks (STGNNs) have shown promising results in many domains, from forecasting to epidemiology. However, understanding the dynamics learned by these models and explaining their behaviour is significantly more complex tha
Externí odkaz:
http://arxiv.org/abs/2410.13469
Autor:
Verdini, Francesco, Melucci, Pierfrancesco, Perna, Stefano, Cariaggi, Francesco, Gaido, Marco, Papi, Sara, Mazurek, Szymon, Kasztelnik, Marek, Bentivogli, Luisa, Bratières, Sébastien, Merialdo, Paolo, Scardapane, Simone
The remarkable performance achieved by Large Language Models (LLM) has driven research efforts to leverage them for a wide range of tasks and input modalities. In speech-to-text (S2T) tasks, the emerging solution consists of projecting the output of
Externí odkaz:
http://arxiv.org/abs/2409.17044
Graph Neural Networks based on the message-passing (MP) mechanism are a dominant approach for handling graph-structured data. However, they are inherently limited to modeling only pairwise interactions, making it difficult to explicitly capture the c
Externí odkaz:
http://arxiv.org/abs/2409.12033
Autor:
Devoto, Alessio, Alvetreti, Federico, Pomponi, Jary, Di Lorenzo, Paolo, Minervini, Pasquale, Scardapane, Simone
Recently, foundation models based on Vision Transformers (ViTs) have become widely available. However, their fine-tuning process is highly resource-intensive, and it hinders their adoption in several edge or low-energy applications. To this end, in t
Externí odkaz:
http://arxiv.org/abs/2408.08670
The deployment of large language models (LLMs) is often hindered by the extensive memory requirements of the Key-Value (KV) cache, especially as context lengths increase. Existing approaches to reduce the KV cache size involve either fine-tuning the
Externí odkaz:
http://arxiv.org/abs/2406.11430
Autor:
Telyatnikov, Lev, Bernardez, Guillermo, Montagna, Marco, Vasylenko, Pavlo, Zamzmi, Ghada, Hajij, Mustafa, Schaub, Michael T, Miolane, Nina, Scardapane, Simone, Papamarkou, Theodore
This work introduces TopoBenchmarkX, a modular open-source library designed to standardize benchmarking and accelerate research in Topological Deep Learning (TDL). TopoBenchmarkX maps the TDL pipeline into a sequence of independent and modular compon
Externí odkaz:
http://arxiv.org/abs/2406.06642
Autor:
Scardapane, Simone
Neural networks surround us, in the form of large language models, speech transcription systems, molecular discovery algorithms, robotics, and much more. Stripped of anything else, neural networks are compositions of differentiable primitives, and st
Externí odkaz:
http://arxiv.org/abs/2404.17625
In this paper, we propose a novel design for AI-native goal-oriented communications, exploiting transformer neural networks under dynamic inference constraints on bandwidth and computation. Transformers have become the standard architecture for pretr
Externí odkaz:
http://arxiv.org/abs/2405.02330
Influence based explainability of brain tumors segmentation in multimodal Magnetic Resonance Imaging
Autor:
Torda, Tommaso, Ciardiello, Andrea, Gargiulo, Simona, Grillo, Greta, Scardapane, Simone, Voena, Cecilia, Giagu, Stefano
In recent years Artificial Intelligence has emerged as a fundamental tool in medical applications. Despite this rapid development, deep neural networks remain black boxes that are difficult to explain, and this represents a major limitation for their
Externí odkaz:
http://arxiv.org/abs/2405.12222
Autor:
Scardapane, Simone, Baiocchi, Alessandro, Devoto, Alessio, Marsocci, Valerio, Minervini, Pasquale, Pomponi, Jary
Publikováno v:
Intelligenza Artificiale, vol. Pre-press, pp. 1-16, 2024
This article summarizes principles and ideas from the emerging area of applying \textit{conditional computation} methods to the design of neural networks. In particular, we focus on neural networks that can dynamically activate or de-activate parts o
Externí odkaz:
http://arxiv.org/abs/2403.07965