Parsing the Stream of Time: The Value of Event-Based Segmentation in a Complex Real-World Control Problem

Temporal information-processing tasks generally require that the continuous stream of time be parsed or segmented, which involves determining boundaries that divide the stream into distinct intervals. In a clock-based segmentation, the boundaries are spaced equally in time, resulting in fixed-duration intervals. In an event-based segmentation, the boundaries depend on the state of the environment, resulting in variable-duration intervals. Models of temporal information processing in cognitive science and artificial intelligence generally rely on clock-based segmentation. However, event-based segmentation can greatly simplify temporal information-processing tasks. We illustrate by describing a complex control problem that appears intractable when cast in terms of a clock-based segmentation, but has a straightforward solution when cast in terms of an event-based segmentation.

[1]  P.J. Antsaklis,et al.  Learning to Coordinate Control Policies of Hybrid Systems , 1993, 1993 American Control Conference.

[2]  Geoffrey E. Hinton,et al.  Adaptive Mixtures of Local Experts , 1991, Neural Computation.

[3]  Jürgen Schmidhuber,et al.  LSTM can Solve Hard Long Time Lag Problems , 1996, NIPS.

[4]  K H Kowal,et al.  Apparent duration and numerosity as a function of melodic familiarity , 1987, Perception & psychophysics.

[5]  M. Reed,et al.  A watched pot sometimes boils: a study of duration experience. , 1980, Acta psychologica.

[6]  Ewart A. C. Thomas,et al.  Time perception and the filled-duration illusion , 1974 .

[7]  M. Mozer,et al.  The Neural Network House: An Overview , 1995 .

[8]  M.C. Mozer.  Neural-network speech processing for toys and consumer electronics , 1996, IEEE Expert.

[9]  Andrew G. Barto,et al.  Improving Elevator Performance Using Reinforcement Learning , 1995, NIPS.

[10]  Michael O. Duff,et al.  Reinforcement Learning Methods for Continuous-Time Markov Decision Problems , 1994, NIPS.

[11]  Jürgen Schmidhuber,et al.  Continuous history compression , 1993 .

[12]  R. E. Hicks,et al.  The repetition effect in judgments of temporal duration across minutes, days, and months. , 1979, The American journal of psychology.

[13]  Kimon P. Valavanis,et al.  A subject-indexed bibliography of discrete event dynamic systems , 1994, IEEE Robotics & Automation Magazine.

[14]  Chris Watkins,et al.  Learning from delayed rewards , 1989 .

[15]  Pavan Sikka,et al.  Monitoring contact using clustering and discriminant functions , 1996, Proceedings of IEEE International Conference on Robotics and Automation.

[16]  Tarek M. Sobh.  8.7 Hybrid Systems and Control , 1996 .

[17]  Thomas G. Dietterich.  What is machine learning? , 2020, Archives of Disease in Childhood.

[18]  Masayuki Inaba,et al.  Design and implementation of a system that generates assembly programs from visual recognition of human action sequences , 1990, EEE International Workshop on Intelligent Robots and Systems, Towards a New Frontier of Applications.

[19]  James J. Kellaris,et al.  Shaping time perceptions with background music: The effect of congruity and arousal on estimates of ad durations , 1996 .

[20]  Michael C. Mozer,et al.  Comparison of reinforcement algorithms on discrete functions: learnability, time complexity, and scaling , 1992, [Proceedings 1992] IJCNN International Joint Conference on Neural Networks.