Olivier Bousquet and Manfred K. Warmuth. Tracking a small set of experts by mixing past posteriors. In David Helmbold and B. Williamson, editors, Proceedings of the 14th Annual Conference on Computational Learning Theory and 5th European Conference on Computational Learning Theory, pages 31--47. Springer, July 2001. Home/Search Document Details and Download
Summary Related Articles Check |
This paper is cited in the following contexts:
On Combining Disjoint Caches - Gramacy (2003) (Correct)
....(Adaptive Replacement Cache) we study how to best combine disjoint lists of pages in order to construct an adaptive paging algorithm that has a miss rate lower than LRU. This paper contains our preliminary results. The approach we use takes advantage of the Expert Framework from online learning[7, 5, 8, 4, 3]. It is also an interesting application of Path Kernels [11] which are used to keep track exponentially many combinations of the disjoint lists implicitly, by only maintaining polynomially many weights. We introduce the concept of rollover [6] in order to better track a dynamically changing ....
....on the observed sequence of requests. Unfortunately, adaptive is seldom clearly de ned, and little theoretical or experimental evidence is ever given that policies are good at adapting. Even more recently, there has been a push by the on line learning community to apply the Expert Framework [7, 5, 8, 4, 3] to systems problems. Specifically, there has bee a push to use the framework to help develop more adaptive page replacement policies, and to better quantify what it means for a policy to be theoretically or experimentally [2, 1, 6] adaptive. Such approaches involve combining the criteria of ....
[Article contains additional citation context not shown here]
Olivier Bousquet and Manfred K. Warmuth. Tracking a small set of experts by mixing past posteriors. In COLT/EuroCOLT, pages 31-47, 2001.
Adaptive Caching by Experts - Gramacy (2003) Self-citation (Warmuth) (Correct)
....heuristics for predicting responses, reacting to events, and or altering systems. An expert s success is measured by loss. Loss is a quantification of the discrepancy between the actions, or predictions of an expert, and what is currently judged to be an optimal response. In the Expert Framework [16, 11, 20, 10, 6] a master algorithm enlists the advice of experts, and observes their losses, in order to to make its own predictions, decisions, or actions. The master algorithm is evaluated by the same loss functions. Ideally, its loss is smaller, or close to, the loss of its best experts. To our knowledge ....
....well. Multiplicative updates like those in (2.1) are a double edged sword: driving the weights of experts which are currently poor to zero so quickly that it becomes very difficult for them to recover if they start doing well. Herbster and Warmuth [16] and more recently Bousquet and Warmuth [6], developed a second set of updates in order to help prevent poor experts weights from becoming too small. After an intermediate LOSS UPDATE, identical to that of (2.1) these updates, dubbed SHARE (or MIXING)UPDATES, force experts with a large amount of weight to share (or mix) a small amount of ....
[Article contains additional citation context not shown here]
Olivier Bousquet and Manfred K. Warmuth. Tracking a small set of experts by mixing past posteriors. In COLT/EuroCOLT, pages 31--47, 2001.
Using Multiple Experts for File Prediction - Brandt (2004) (Correct)
No context found.
Olivier Bousquet and Manfred K. Warmuth. Tracking a small set of experts by mixing past posteriors. In David Helmbold and B. Williamson, editors, Proceedings of the 14th Annual Conference on Computational Learning Theory and 5th European Conference on Computational Learning Theory, pages 31--47. Springer, July 2001.
Online articles have much greater impact More about CiteSeer.IST Add search form to your site Submit documents Feedback
CiteSeer.IST - Copyright Penn State and NEC