Combining Multilayer Networks to Combine Learning
- Creators
- Bax, Eric
Abstract
This paper explores methods to combine networks, implementations of network combination on multicomputers, and applications of network combination. We divide the network combination process into two steps. First, a matching algorithm rearranges the networks so that corresponding weights in the networks being combined play corresponding roles in the networks' functions. Then a proportioning algorithm chooses a convex combination of the matched networks. The combination network is a weight-by-weight convex combination of the matched networks. This paper begins by examining combinations of pairs of networks. Matching and proportioning algorithms are developed, analyzed, implemented, and tested. Next, algorithms are developed to combine several networks. Then the combination process is poked and prodded to explore its nature, its utility, and its limits.
Additional Information
© 1997 California Institute of Technology. March 28, 1997. I thank Yaser Abu-Mostafa for his advice and guidance concerning the development of these results. I thank Zehra Cataltepe , Malik Magdon-Ismaiel, Joseph Sill, and Xubo Song for many helpful pointers and educational conversations.Attached Files
Submitted - CSTR1997.pdf
Submitted - postscript.ps
Files
Name | Size | Download all |
---|---|---|
md5:647f9d3fe613b9f7c51b55a597eb2aed
|
4.0 MB | Preview Download |
md5:72bc4bdccc547175525e25971f0dab9e
|
5.0 MB | Download |
Additional details
- Eprint ID
- 26831
- Resolver ID
- CaltechCSTR:1997.cs-tr-98-01
- Created
-
2001-04-30Created from EPrint's datestamp field
- Updated
-
2019-10-03Created from EPrint's last_modified field
- Caltech groups
- Computer Science Technical Reports