Prof. Verena Rieser

Potential Supervisor

My interests are interaction learning, speech and multimodal technology. Please see the following link http://www.macs.hw.ac.uk/~vtr1/

I specialize in data-driven machine learning approaches for sequential decision making as applied to interactive systems, including human-robot interaction, spoken dialogue systems (aka. conversational agents), multimodal output generation, decision support and multi-agent systems for computational sustainability.  I currently supervise 3 MSc, 3 PhD students and 2 postdoctoral researchers in these areas. 

Potential Research topics for MSc/EngD:
1.    Strategic Interaction: Using advanced machine learning methods, such as approximate DEC-POMDPs, for effective human-robot teamwork.
2.    Cognitive Interfaces: Develop human-robot interfaces which can adapt to the cognitive load of the user, as estimated from multimodal sensor data.
3.    Deep Learning for Open Domain Interaction: Investigate Deep Neural Networks to learn optimal interaction control from unlabelled data.
4.    Situated Output Generation: Develop a data-driven framework for multi-modal output generation (including language) for Human-Robot Interaction.

 

Research keywords: 
NLP, reinforcement learning, spoken and multimodal interaction
Phone: 
0131 451 4192