Summary

This summary outlines a professional trajectory that bridges interdisciplinary research and foundational academic training. The speaker's journey begins with a PhD at Sorbonne University in Paris, establishing a strong background in machine learning and data science. A significant milestone in this path occurred during an internship at Facebook AI Research, which focused on Neural Machine Translation with Memory Networks. During this high-stakes professional environment, the speaker successfully completed a research project titled *Measuring Idiomaticity in Text Embedding Models with Compositionality*, a collaborative work with Sondre Wold. This dissertation provides critical insights into how language patterns are organized and measured within neural text embeddings. The speaker now plans to present this groundbreaking research work at the upcoming EACL 2026 conference, offering a fresh perspective on the relationship between idiomology and model architecture. The research highlights that idiomatic language is not merely a collection of word combinations but emerges from complex compositional patterns that define meaningful communication. This approach underscores how deep theoretical studies can illuminate practical applications in AI systems like translation models, creating a more robust foundation for future development. Ultimately, the speaker emphasizes that the intersection of creative linguistics and engineering demands rigorous experimental design to validate claims about how humans actually construct language in the digital realm.

Title

Étienne Simon's homepage: ~

Description

Étienne Simon's homepage: ~

Keywords

machine, learning, simon, navigation, research, teaching, working, extraction, paris, deep, team, internship, home, details, stuff, repositories, path

NS Lookup

A 51.15.8.18

Dates

Created 2026-03-07

Updated 2026-03-07

Summarized 2026-03-22