← Search

Samuel Victor Norman-Haignere

2 accepted papers

2023

Large language models transition from integrating across position-yoked, exponential windows to structure-yoked, power-law windows

NeurIPS 2023poster

Modern language models excel at integrating across long temporal scales needed to encode linguistic meaning and show non-trivial similarities to biological neural systems. Prior work suggests that human brain responses to language exhibit hierarchically organized "integration windows" that substanti…

Cited by 6SourcePDFScholar
2021

Understanding Adaptive, Multiscale Temporal Integration In Deep Speech Recognition Systems

NeurIPS 2021poster

Natural signals such as speech are hierarchically structured across many different timescales, spanning tens (e.g., phonemes) to hundreds (e.g., words) of milliseconds, each of which is highly variable and context-dependent. While deep neural networks (DNNs) excel at recognizing complex patterns fro…