|
Activity Number:
|
490
|
|
Type:
|
Invited
|
|
Date/Time:
|
Thursday, August 2, 2007 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
IMS
|
| Abstract - #307941 |
|
Title:
|
Aggregating Algorithms for Individual Sequences
|
|
Author(s):
|
Nicolò Cesa-Bianchi*+
|
|
Companies:
|
Università degli Studi di Milano
|
|
Address:
|
Dipartimento di Scienze dell'Informazione, Milano, 20135, Italy
|
|
Keywords:
|
|
|
Abstract:
|
Prediction with expert advice is concerned with the design and analysis of aggregating algorithms without imposing any probabilistic assumption on the data-generating mechanism. The performance bounds that can be proven in such a general framework are surprisingly strong. Typically, these bounds depend on quantities that are empirical versions of analogous entities arising in statistical aggregation problems. We will review a number of results in prediction of individual sequences, including some recent results where the regret is bounded in terms of empirical variances. We will also describe some simple techniques for converting any individual sequence aggregation bound into a corresponding bound for i.i.d. sources. These techniques are effective, in the sense that little processing is required besides the computation performed by the original individual sequence algorithm.
|
- The address information is for the authors that have a + after their name.
- Authors who are presenting talks have a * after their name.
Back to the full JSM 2007 program |
