3 papers
cs.LG2026
Trees from Marginals: Autoregressive drafting with factorized priors
Yuma Oda, Ryan Mathieu, Roman Knyazhitskiy +1
Speculative decoding greatly increases the interactivity of autoregressive language models by trading off computation for extra tokens generated in a single forward pass. Factorize…
cs.LG2026
Early-Exit Graph Neural Networks for Link Prediction
Roman Knyazhitskiy, Andrea Giuseppe Di Francesco
Graph Neural Networks are great for link prediction in various network-like structures; however, the question of their speed/quality tradeoff has been barely studied. While in prac…
cs.CV2025
ARC: Anchored Representation Clouds for High-Resolution INR Classification
Joost Luijmes, Alexander Gielisse, Roman Knyazhitskiy +1
Implicit neural representations (INRs) encode signals in neural network weights as a memory-efficient representation, decoupling sampling resolution from the associated resource co…