Center for Advanced Intelligence Project

Speaker : Dr. Christophe Gravier, PhD, is an Associate Professor at Université Jean Monnet, Université de Lyon, France

Title : Short-texts generation from structured knowledge
Presentation Slides: Download here (PDF/ 5.7MB)

Abstract : Many of Natural Language Processing modern pipelines share a two-step process approach : a first network make it possible to learn word representation, while a second is dedicated to (deep) learning models for downstream tasks. Such downstream tasks usually require a large amount of labeled data, and text generation is not an exception. Despite the recent findings on attention mechanisms and residual connections in such networks, it is still very difficult to train a model to generate text from concepts it has never seen at training time (zero-shot setup). My talk will emphasize how neural models in NLP can benefit from external structured knowledge, and I will especially look at some of our recent works in leveraging knowledge bases for Natural Language Generation models, especially for zero-shot question generation and learning to generate textual Wikipedia summaries.

This talk is based on the following publications :

ElSahar et al. (2018) – NAACL-HLT 2018 (pp. 218–228)
Zero-Shot Question Generation from Knowledge Graphs for Unseen Predicates and Entity Types

Vougiouklis et al. (2018) – J. Web Semantics (52–53, pp. 1–15)
Neural Wikipedian: Generating Textual Summaries from Knowledge Base Triples

Kaffee et al. (2018) – NAACL-HLT 2018 (pp. 640–645)
Learning to Generate Wikipedia Summaries for Underserved Languages from Wikidata

Kaffee et al. (2018) – ESWC 2018 (pp. 319–334)
Mind the (Language) Gap: Generation of Multilingual Wikipedia Summaries from Wikidata for ArticlePlaceholders

ElSahar et al. (2018) – LREC 2018 (pp. 3448 – 3452)
T-REx: A Large Scale Alignment of Natural Language with Knowledge Base Triples

Short bio : Dr. Christophe Gravier, PhD, is an Associate Professor at Université Jean Monnet, Université de Lyon, France. His research interest is Natural Language Processing, including text generation, chatbots, learning binary embeddings, and question answering. Christophe Gravier is a Computer Science teacher at Télécom Saint-Etienne, a French engineering school (Master diploma), where he also serves as director of development and innovation. Julien Tissier, currently visiting RIKEN AIP, is one of his current PhD student.