This article argues that future generations of computer-based systems will need cognitive user interfaces to achieve sufficiently robust and intelligent human interaction. These cognitive user interfaces will be characterized by the ability to support inference and reasoning, planning under uncertainty, short-term adaptation, and long-term learning from experience. An appropriate engineering framework for such interfaces is provided by partially observable Markov decision processes (POMDPs) that integrate Bayesian belief tracking and reward-based reinforcement learning. The benefits of this approach are demonstrated by the example of a simple gesture-driven interface to an iPhone application. Furthermore, evidence is provided that humans appear to use similar mechanisms for planning under uncertainty.
%0 Journal Article
%1 Young2010
%A Young, Steve J.
%D 2010
%J IEEE Signal Processing Magazine
%K cognitive interfaces user
%N 3
%P 128-140
%R 10.1109/MSP.2010.935874
%T Cognitive User Interfaces
%U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.183.8405 http://ieeexplore.ieee.org/xpl/articleDetails.jsp?arnumber=5447049
%V 27
%X This article argues that future generations of computer-based systems will need cognitive user interfaces to achieve sufficiently robust and intelligent human interaction. These cognitive user interfaces will be characterized by the ability to support inference and reasoning, planning under uncertainty, short-term adaptation, and long-term learning from experience. An appropriate engineering framework for such interfaces is provided by partially observable Markov decision processes (POMDPs) that integrate Bayesian belief tracking and reward-based reinforcement learning. The benefits of this approach are demonstrated by the example of a simple gesture-driven interface to an iPhone application. Furthermore, evidence is provided that humans appear to use similar mechanisms for planning under uncertainty.
@article{Young2010,
abstract = {This article argues that future generations of computer-based systems will need cognitive user interfaces to achieve sufficiently robust and intelligent human interaction. These cognitive user interfaces will be characterized by the ability to support inference and reasoning, planning under uncertainty, short-term adaptation, and long-term learning from experience. An appropriate engineering framework for such interfaces is provided by partially observable Markov decision processes (POMDPs) that integrate Bayesian belief tracking and reward-based reinforcement learning. The benefits of this approach are demonstrated by the example of a simple gesture-driven interface to an iPhone application. Furthermore, evidence is provided that humans appear to use similar mechanisms for planning under uncertainty.},
added-at = {2021-02-01T10:51:23.000+0100},
author = {Young, Steve J.},
biburl = {https://www.bibsonomy.org/bibtex/26bce9769d28aff541d1809b5c4345abb/m-toman},
doi = {10.1109/MSP.2010.935874},
interhash = {f2d7f58fb973720329ac5b75ebd3134b},
intrahash = {6bce9769d28aff541d1809b5c4345abb},
issn = {1053-5888},
journal = {IEEE Signal Processing Magazine},
keywords = {cognitive interfaces user},
mendeley-tags = {cognitive user interfaces},
month = may,
number = 3,
owner = {mtoman},
pages = {128-140},
timestamp = {2021-02-01T10:51:23.000+0100},
title = {Cognitive User Interfaces},
url = {http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.183.8405 http://ieeexplore.ieee.org/xpl/articleDetails.jsp?arnumber=5447049},
volume = 27,
year = 2010
}