TY - GEN
T1 - The Segmented iHMM
T2 - 33rd International Conference on Machine Learning, ICML 2016
AU - Saeedi, Ardavan
AU - Hoffman, Matthew
AU - Johnson, Matthew
AU - Adams, Ryan
N1 - Funding Information:
Thanks to Asma Ghandeharioun for her help with the sensors data and Mira Dontcheva for her feedback on segmenting user behavior traces. RPA is supported by NSF IIS-1421780 and the Alfred P. Sloan Foundation.
Publisher Copyright:
© 2016 by the author(s).
PY - 2016
Y1 - 2016
N2 - We propose the segmented iHMM (siHMM), a hierarchical infinite hidden Markov model (iHMM) that supports a simple, efficient inference scheme. The siHMM is well suited to segmentation problems, where the goal is to identify points at which a time series transitions from one relatively stable regime to a new regime. Conventional iHMMs often struggle with such problems, since they have no mechanism for distinguishing between high- and low-level dynamics. Hierarchical HMMs (HHMMs) can do better, but they require much more complex and expensive inference algorithms. The siHMM retains the simplicity and efficiency of the iHMM, but out-performs it on a variety of segmentation problems, achieving performance that matches or ex-ceeds that of a more complicated HHMM.
AB - We propose the segmented iHMM (siHMM), a hierarchical infinite hidden Markov model (iHMM) that supports a simple, efficient inference scheme. The siHMM is well suited to segmentation problems, where the goal is to identify points at which a time series transitions from one relatively stable regime to a new regime. Conventional iHMMs often struggle with such problems, since they have no mechanism for distinguishing between high- and low-level dynamics. Hierarchical HMMs (HHMMs) can do better, but they require much more complex and expensive inference algorithms. The siHMM retains the simplicity and efficiency of the iHMM, but out-performs it on a variety of segmentation problems, achieving performance that matches or ex-ceeds that of a more complicated HHMM.
UR - http://www.scopus.com/inward/record.url?scp=84998785413&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84998785413&partnerID=8YFLogxK
M3 - Conference contribution
AN - SCOPUS:84998785413
T3 - 33rd International Conference on Machine Learning, ICML 2016
SP - 3949
EP - 3959
BT - 33rd International Conference on Machine Learning, ICML 2016
A2 - Weinberger, Kilian Q.
A2 - Balcan, Maria Florina
PB - International Machine Learning Society (IMLS)
Y2 - 19 June 2016 through 24 June 2016
ER -