The systems I've build observe a person or are told something by them, form a judgement about it and then have to decide what to do with it including whether to do anything at all. My research is about that decision and about what makes it hard when the judgement is uncertain.

Feedback given at the wrong moment could interrupt a person, add to what they are already keeping track of or cost them confidence in the system. An agent which could offer a hint has to weigh what the hint is worth against what it costs to interrupt. A tool which responds to what someone tells it about themselves has to decide what it is allowed to say before it says anything. In each case the question is not only what a system should say but whether and when it should say anything at all and what should happen when it has misunderstood the person.

My work so far falls into three groups.
Sensing and response in health. Systems which observe a person during an activity and decide whether to respond, where a badly timed response has a cost and the safety rules belong to a clinician rather than to the model.

Deciding when to speak. Agents and tools that hold something a person might need and have to weigh the value of offering it against the cost of interrupting.

Tools for design and learning. Interfaces which help people apply research to concrete decisions and make the reasoning behind a suggestion available for inspection.

In each of these I build the system, decide in advance what evidence would show whether it works and test it before it reaches a participant.

1. Current Research

2. Publications

3. Key Projects

4. Activity Projects