[Research] Lab meeting this Wednesday 10/7 12noon in GHC 8102

[Artur Dubrawski]

BTW, Sajid submits the following title and abstract of his presentation:

AUTOMATIC STATE DISCOVERY FOR UNSTRUCTURED AUDIO SCENE ANALYSIS

We present a novel classification scheme for unstructured audio scene
analysis. Our scheme is based on our recently introduced machine learning
algorithm called Simultaneous Temporal And Contextual Splitting (STACS)
that discovers the appropriate number of states and efficiently learns
accurate Hidden Markov Model (HMM) parameters for the given data.
STACS-based algorithms train HMMs upto an order of magnitude faster than
Baum-Welch, avoid the overfitting problem commonly encountered in learning
large state complex HMMs using Expectation Maximization (EM) methods
such as Baum-Welch, and achieve superior classification results on a very
diverse dataset with minimal pre-processing. Furthermore, our scheme has
proven to be highly effective for building real-world applications and has
been integrated into a commercial surveillance system as an event
detection component.

This is joint work with Julian Ramos, Artur Dubrawski and Geoff Gordon of
CMU and Abhishek Sharma of MobileFusion, Inc.


Artur Dubrawski wrote:
> Dear Autonians,
>
> The first 2009/10 Lab meeting will take place this Wednesday, Oct 7, 
> at 12noon in GHC 8102 (it is in that ugly new building next door :)
>
> We will first go around with the updates on the good things that have happened
> since our last meeting in the spring. Then we will be entertained by 
> a veteran member of the Auton Lab, Mr. Sajid Siddiqi, who will discuss the
> use of his STACS algorithm for learning classifiers of sound that are useful 
> for analysis and interpretation of unstructured audio scenes in cool real
> world applications.
>
> The intellectual feast will be complemented with food.
>
> See you all there
> Artur
>  
>
> _______________________________________________
> Research mailing list
> Research at autonlab.org
> https://www.autonlab.org/mailman/listinfo/research
>
>