Chris Snijders - Irrelevant private stuff 2 Chris Snijders - - PDF document

1

Chris Snijders – Eindhoven University of Technology (NL)

Human-Technology Interaction Human Perception Human-Computer-Interaction Human-Games-Interaction Human-Data-Interaction Human-Model-Interaction

c.c.p.snijders@tue.nl

• Trust in technology / persuasion
• Recommender systems
• Online behavior
• …

“psychologists with a technical interest” “psychologists with a technical interest”

Chris Snijders – Eindhoven University of Technology (NL)

Human-Technology Interaction Human Perception Human-Computer-Interaction Human-Games-Interaction Human-Data-Interaction Human-Model-Interaction

c.c.p.snijders@tue.nl

• Trust in technology / persuasion
• Recommender systems
• Online behavior
• …

“psychologists with a technical interest” “psychologists with a technical interest”

2

Chris Snijders – Eindhoven University of Technology (NL)

Human-Technology Interaction Human Perception Human-Computer-Interaction Human-Games-Interaction Human-Data-Interaction Human-Model-Interaction

c.c.p.snijders@tue.nl

“psychologists with a technical interest” “psychologists with a technical interest”

Connection of abstract representation of data to user mental-states >> with Philips: air purifier data being used to understand users >> learning data: logs of the click-stream

• f students in a learning environment

>> online users in web stores: how can we understand what kind of user they are (how soon can we do this) Connection of abstract representation of data to user mental-states >> with Philips: air purifier data being used to understand users >> learning data: logs of the click-stream

• f students in a learning environment

>> online users in web stores: how can we understand what kind of user they are (how soon can we do this)

Chris Snijders - Irrelevant private stuff

3

Chris Snijders – @Dagstuhl

The models themselves are the smaller part of the problem: technology alone is not enough. You can have models but that need not imply that they are used. In my area, perhaps also in yours, there is more to gain by understanding the people who use these models better. Humans s&^ck at many tasks – but there might be consistency to their madness. Can’t these models be used for other uses than “just” improving the process? Personalization: what (types of models) work for which user? What determines whether people trust a model?