Song Duong

About Me

I am a Member of Technical Staff at Liquid AI in the Post-Training team.

Previously I was a Ph.D. student at Criteo and CNRS-ISIR, Sorbonne University, under the supervision of Patrick Gallinari and Alberto Lumbreras. My research focused on building faithful and efficient text generation models from structured/unstructed data.

Research Interests

Natural Language Processing: LLMs, efficient architectures, Post-Training

News

[Jul. 2025] Joined Liquid AI as a Member of Technical Staff, Post-Training.
[Jul. 10th 2025] Succesfully defended my PhD 🎉 (Link to recording)
[Apr. 2025] Attended the ALPS 2025 winter school. 🏔️
[Jan. 2025] Our paper ‘SCOPE’ is accepted at ICLR 2025.
[Mar. 2024] Our paper ‘LOCOST’ is honored with the Best Paper Award.
[Feb. 2024] Our paper ‘LOCOST’ is accepted and selected for an oral presentation at EACL 2024.

Publications

ICLR

SCOPE: A Self-supervised Framework for Improving Faithfulness in Conditional Text Generation

Song Duong*, Florian Le Bronnec*, Alexandre Allauzen, Vincent Guigue, Alberto Lumbreras, Laure Soulier, Patrick Gallinari (*equal contribution)

13th International Conference on Learning Representations (ICLR), 2025.

PDF Code BibTex

EACL

LOCOST: State-Space Models for Long Document Abstractive Summarization

Florian Le Bronnec*, Song Duong*, Mathieu Ravaut, Alexandre Allauzen, Nancy F. Chen, Vincent Guigue, Alberto Lumbreras, Laure Soulier, Patrick Gallinari (*equal contribution)

18th Conference of the European Chapter of the Association for Computational Linguistics (EACL), 2024.

PDF Code BibTex Best Paper Award

AISTATS

Learning from Multiple Sources for Data-to-Text and Text-to-Data

Song Duong, Alberto Lumbreras, Mike Gartrell, Patrick Gallinari

26th International Conference on Artificial Intelligence and Statistics (AISTATS), 2023.

PDF Code BibTex

About Me

Research Interests

News

Publications

Services

Conference Reviewers