Français Anglais
Accueil Annuaire Plan du site
Home > Research results > Dissertations & habilitations
Research results
Ph.D de

Group : Large-scale Heterogeneous DAta and Knowledge

Question Answering with Hybrid Data and Models

Starts on 01/01/1970
Advisor : MA, Yue

Funding :
Affiliation : Université Paris-Sud
Laboratory : salle de conférences du LIMSI

Defended on 06/02/2020, committee :
Patrice Bellot (Rapporteur) - Université Aix-Marseille
Mohand Boughanem (Rapporteur) - Université Paul Sabatier
Catherine Berrut (Examinatrice) - Université Grenoble Alpes
Patrick Gallinari (Examinateur) - Sorbonne Université
Nicolas Sabouret (Examinateur) - Université Paris-Sud
Anne Vilnat (Examinatrice) - Université Paris-Sud
Brigitte Grau (Directeur de thèse) - ENSIIE/LIMSI
Yue Ma (Co-encadrant de thèse) - Université Paris-Sud

Research activities :

Abstract :
Question Answering is a discipline which lies in between natural language processing and information retrieval domains. Emergence of deep learning approaches in several fields of research such as computer vision, natural language processing, speech recognition etc. has led to the rise of end-to-end models.

In the context of GoASQ project, we investigate, compare and combine different approaches for answering questions formulated in natural language over textual data on open domain and biomedical domain data. The thesis work mainly focuses on 1) Building models for small scale and large scale datasets, and 2) Leveraging structured and semantic information into question answering models. Hybrid data in our research context is fusion of knowledge from free text, ontologies, entity information etc. applied towards free text question answering.

The current state-of-the-art models for question answering use deep learning based models. In order to facilitate using them on small scale datasets on closed domain data, we propose to use domain adaptation. We model the BIOASQ biomedical question answering task dataset into two different QA task models and show how the Open Domain Question Answering task suits better than the Reading Comprehension task by comparing experimental results. We pre-train the Reading Comprehension model with different datasets to show the variability in performance when these models are adapted to biomedical domain. We find that using one particular dataset (SQUAD v2.0 dataset) for pre-training performs the best on single dataset pre-training and a combination of four Reading Comprehension datasets performed the best towards the biomedical domain adaptation. We perform some of the above experiments using large scale pre-trained language models like BERT which are fine-tuned to the question answering task. The performance varies based on the type of data used to pre-train BERT. For BERT pre-training on the language modelling task, we find the biomedical data trained BIOBERT to be the best choice for biomedical QA.

Since deep learning models tend to function in an end-to-end fashion, semantic and structured information coming from expert annotated information sources are not explicitly used. We highlight the necessity for using Lexical and Expected Answer Types in open domain and biomedical domain question answering by performing several verification experiments. These types are used to highlight entities in two QA tasks which shows improvements while using entity embeddings based on the answer type annotations. We manually annotated an answer variant dataset for BIOASQ and show the importance of learning a QA model with answer variants present in the paragraphs.

Our hypothesis is that the results obtained from deep learning models can further be improved using semantic features and collective features from different paragraphs for a question. We propose to use ranking models based on binary classification methods to better rank Top-1 prediction among Top-K predictions using these features, leading to an hybrid model that outperforms state-of-art-results on several datasets. We experiment with several overall Open Domain Question Answering models on QA sub-task datasets built for Reading Comprehension and Answer Sentence Selection tasks. We show the difference in performance when these are modelled as overall QA task and highlight the wide gap in building end-to-end models for overall question answering task.

Ph.D. dissertations & Faculty habilitations


Creative work has been at the core of research in Human-Computer Interaction (HCI). I describe the results of a series of studies that look at how creators work, where creators include artists with years of professional practice, as well as learners, or novices and casual makers. My research focuses on three creation activities: drawing, physical modeling, and music composition. For these activities, I examine how artists switch between representations and how these representations evolve throughout their creative process, from early sketches to fine-grained forms or structured vocabularies. I present interactive systems that enrich their workflow (i) by extending their computer tools with physical user interfaces, or (ii) by making physical materials interactive. I also argue that sketch-based representations can allow for user interfaces that are more personal and less rigid. My presentation will reflect on lessons and limitations of this work and discuss challenges for future design-support tools.