lord1234
Contributor
This isn't exactly diving related, but it is definitely marine bio related, and maybe some of you might be interested...so here is Marine Biology...with a Comp Sci Twist
this is a copy of an email I recieved from the CS department chair, sent out to all students about this colloquium, sorry for length)
The UML web site says that the university will be open tomorrow, so
the colloquium will be held as planned.
This will be the last colloquium of the spring term.
----------------------------------------------------------------------
Computer Science Colloquium
----------------------------------------------------------------------
Information Entropy of Humpback Whale Songs
Prof. John R. Buck
University of Massachusetts Dartmouth
Dept of Electrical and Computer Engineering &
School for Marine Science and Technology
Wednesday, 17 May 2006
Olsen 311
Refreshments at 2:30, Talk from 3:00-4:00
Humpback whales produce songs consisting of more than a dozen sounds
repeated in a complicated pattern. Many researchers have analyzed
this song structure using manual classification and visual inspection
of the sequence of sounds. Payne and McVay (1971) proposed a
hierarchical syntax, which is currently the dominant hypothesis among
marine mammal researchers. This hypothesis raises the question of
whether the whales organize their songs using a hierarchical syntax,
or if this structure is a bias of the human observers.
We applied information theory techniques to these songs to analyze the
structure in an objective, quantitative manner. We converted the
songs into discrete symbol sequences using both neural nets and human
observers. The entropy of these symbol sequences estimated using an
independent identically distributed model, a first-order Markov model
and the nonparametric sliding window match length (SWML) estimator
(Wyner & Ziv, 1989). A comparison of these entropy estimates
indicates that the production of the songs are governed by strong
syntactic constraints. Also, the entropy estimates from parametric
models were significantly higher than that from the non-parametric
estimator, indicating that both the memoryless and first order Markov
models are too simple to model the syntax of humpback whale songs.
Correlation functions of units in the songs demonstrate repetitive
structure with two different periods of approximately 8 and 200 units.
A simple hierarchical syntax will produce this phenomenon, while any
nonhierarchical model consistent with the correlation behavior
requires a huge number of parameters. This result is consistent with
the Payne and McVay (1971) syntax.
Bio: John R. Buck received his Ph.D. from the MIT/WHOI Joint Program
in Ocean and Electrical Engineering in 1996. He is currently an
associate professor at the University of Massachusetts Dartmouth,
jointly appointed in the Department of Electrical and Computer
Engineering and the School for Marine Science and Technology. His
research interests include signal processing, underwater acoustics,
information theory and animal bioacoustics. Dr. Buck is the co-author
of two textbooks, and the recipient an NSF CAREER award (1998), an ONR
Young Investigator Award (2000), and the IEEE Education Society's Mac
Van Valkenburg Early Career Teaching Award (2005). During 2003-2004,
he was a Fulbright Senior Scholar in Australia, hosted by DSTO Sydney
and the University of Sydney.
----------------------------------------------------------------------
For more information on the CS Colloquium Series, see
http://www.cs.uml.edu/~holly/colloquia or contact Prof. Holly Yanco at
holly@cs.uml.edu.
For directions to Olsen Hall at UMass Lowell, see
http://www.cs.uml.edu/~fredm/directions/olsen.shtml
this is a copy of an email I recieved from the CS department chair, sent out to all students about this colloquium, sorry for length)The UML web site says that the university will be open tomorrow, so
the colloquium will be held as planned.
This will be the last colloquium of the spring term.
----------------------------------------------------------------------
Computer Science Colloquium
----------------------------------------------------------------------
Information Entropy of Humpback Whale Songs
Prof. John R. Buck
University of Massachusetts Dartmouth
Dept of Electrical and Computer Engineering &
School for Marine Science and Technology
Wednesday, 17 May 2006
Olsen 311
Refreshments at 2:30, Talk from 3:00-4:00
Humpback whales produce songs consisting of more than a dozen sounds
repeated in a complicated pattern. Many researchers have analyzed
this song structure using manual classification and visual inspection
of the sequence of sounds. Payne and McVay (1971) proposed a
hierarchical syntax, which is currently the dominant hypothesis among
marine mammal researchers. This hypothesis raises the question of
whether the whales organize their songs using a hierarchical syntax,
or if this structure is a bias of the human observers.
We applied information theory techniques to these songs to analyze the
structure in an objective, quantitative manner. We converted the
songs into discrete symbol sequences using both neural nets and human
observers. The entropy of these symbol sequences estimated using an
independent identically distributed model, a first-order Markov model
and the nonparametric sliding window match length (SWML) estimator
(Wyner & Ziv, 1989). A comparison of these entropy estimates
indicates that the production of the songs are governed by strong
syntactic constraints. Also, the entropy estimates from parametric
models were significantly higher than that from the non-parametric
estimator, indicating that both the memoryless and first order Markov
models are too simple to model the syntax of humpback whale songs.
Correlation functions of units in the songs demonstrate repetitive
structure with two different periods of approximately 8 and 200 units.
A simple hierarchical syntax will produce this phenomenon, while any
nonhierarchical model consistent with the correlation behavior
requires a huge number of parameters. This result is consistent with
the Payne and McVay (1971) syntax.
Bio: John R. Buck received his Ph.D. from the MIT/WHOI Joint Program
in Ocean and Electrical Engineering in 1996. He is currently an
associate professor at the University of Massachusetts Dartmouth,
jointly appointed in the Department of Electrical and Computer
Engineering and the School for Marine Science and Technology. His
research interests include signal processing, underwater acoustics,
information theory and animal bioacoustics. Dr. Buck is the co-author
of two textbooks, and the recipient an NSF CAREER award (1998), an ONR
Young Investigator Award (2000), and the IEEE Education Society's Mac
Van Valkenburg Early Career Teaching Award (2005). During 2003-2004,
he was a Fulbright Senior Scholar in Australia, hosted by DSTO Sydney
and the University of Sydney.
----------------------------------------------------------------------
For more information on the CS Colloquium Series, see
http://www.cs.uml.edu/~holly/colloquia or contact Prof. Holly Yanco at
holly@cs.uml.edu.
For directions to Olsen Hall at UMass Lowell, see
http://www.cs.uml.edu/~fredm/directions/olsen.shtml
