La matematica delle reti neurali. VII

Read full text Summarise with VIRAT
Authors: Silvano Salvador
Cited by: 0

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

Dedicato alle architetture e alla teoria più recenti. Copre i grandi modelli linguistici (tokenizzazione BPE/WordPiece, scaling laws, RLHF), modelli di diffusione (DDPM, score matching), graph neural networks (WL test, message passing), meta-apprendimento (MAML), federated learning, interpretabilità meccanicistica, modelli efficienti (LoRA, quantizzazione) e teoria moderna dell'ottimizzazione (grokking, loss landscape, fenomeni di generalizzazione).

Fields of Study & Subjects

Neural Networks and Applications Stochastic Gradient Optimization Techniques Domain Adaptation and Few-Shot Learning

Publication details

Publication date
22 March 2026
Language
IT