General research interests
- Hybrid symbolic / statistical approaches to NLP ;
- Discourse and dialogue modeling ;
- Spoken dialogue systems ;
- Planning and learning under uncertainty;
- Dependency grammar ;
- Embodied and situated cognition ;
- Social and developmental robotics ;
- Foundations of artificial intelligence (philosophy of mind, philosophy of language, phenomenology, semiotics, anthropology)
PhD research synopsis
My general research interest lies in the development of efficient algorithms for spoken dialogue systems. The long-term aim is to build artificial agents which are able to interact with humans via natural language (e.g. spontaneous speech) to perform various tasks. Spoken dialogue systems are expected to play an ever-increasing role in our daily interactions with technology. Wouldn't it be great if we could simply talk to our technological devices instead of having to tediously configure or program them?
Given the inherent complexity of natural language (and the uncertainty associated with speech recognition errors), building such kind of systems is a non-trivial engineering task. Figure 1 illustrates a a simplified architecture schema. My Ph.D research concentrates on decision-making algorithms for dialogue management in open-ended domains. The goal is to provide a new hybrid approach to dialogue management which combines rich linguistic knowledge (pragmatic rules, models of dialogue structure) with probabilistic models for planning and learning under uncertainty into a single, unified framework.
Figure 1: Generic architecture of a spoken dialogue system
On the practical side, I am interested in the application of these ideas for intelligent user interfaces and human-robot interaction. For instance, tutoring systems to learn foreign languages, "social" robots capable of taking care of routine tasks in homes, offices, schools or hospitals, or cognitive assistants for disabled, mentally impaired or elderly persons.
In many of these applications, the dialogue system has to operate in a rich, dynamic environment which also needs to be properly captured. The agent must therefore relate - or ground - the interaction to an active understanding of its environment and what is intended to be done in it through goals and plans of actions. And since the real world is also highly dynamic in nature (things are constantly changing), it must also be able to quickly adapt its behaviour relative to the surrounding context and the intentional, attentional and affective state of its conversational partners. In practice, it is almost impossible for the system developer to encode such complex interaction domains entirely by hand, and we are therefore using machine learning techniques to automatically learn and improve the system's internal models based on prior experience.
The proposed approach is currently in development as part of my Ph.D. research. I am working on two fronts. On the theoretical side, I am investigating the conceptual and formal foundations of this new framework for dialogue management, notably the core representations and algorithms and the use of machine learning techniques to estimate the parameters of the probabilistic models. On the practical side, I also aim to instantiate our approach in a robotic platform and demonstrate its validity in a real-world scenario.
For more information, I invite you to have a look at my PhD Research Proposal, which outlines the motivation, approach, development and evaluation procedure of my PhD work. My publications pages is also worth visiting if you want more technical details on my work.
