Corpora, Cognition and Composition - Exploring Semantics and Semantic Composition

Explore the intersection of language processing, cognitive science, and machine learning in this lecture by Alona Fyshe from the University of Victoria. Delve into the complexities of semantic composition and word meaning using large text corpora and brain recordings. Discover how latent representations can reveal overlapping and complementary information between linguistic data sources. Learn about innovative approaches to studying semantics and semantic composition grounded in brain activity and word usage patterns. Gain insights into cutting-edge research on adjective-noun phrases, sentiment analysis, and the mental processes involved in combining words to create higher-order meaning.

Introduction
Inspiration from Siri
Semantics
Magnetoencephalography
Machine Learning
First Data
Science Paper
Meg Data
Word Features
Matrix
Matrix Factorization
Nonnegative sparse embeddings
Latent representations
Data sources
Matrix factorisation
Online dictionary learning
Results
Mapping
Neural Representations
Train Test Matrix
Semantic Composition
Composition
Semantic violations