Published July 1991
| Published
Book Section - Chapter
Open
Objective functions for probability estimation
- Creators
- Miller, John W.
- Goodman, Rod
- Smyth, Padhraic
Abstract
Backpropagation was originally derived in the context of minimizing a mean-squared error (MSE) objective function. More recently there has been interest in objective functions that provide accurate class probability estimates. In this paper we derive necessary and sufficient conditions on the required form of an objective function to provide probability estimates. This leads to the definition of a general class of functions which includes MSE and cross cutropy (CE) as two of the simplest cases.
Additional Information
© 1991 IEEE. The research described in this paper was supported in part by DARPA under grant number AFOSR-90-0199 and by NSF grant ENG-8711673. In addition this work was carried out in part by the Jet Propulsion Laboratories, California Institute of Technology, under a contract with the National Aeronautics and Space Administration.Attached Files
Published - 00155295.pdf
Files
00155295.pdf
Files
(261.2 kB)
| Name | Size | Download all |
|---|---|---|
|
md5:27daaba12e33af7cff16fd9848ee3b10
|
261.2 kB | Preview Download |
Additional details
- Eprint ID
- 93835
- Resolver ID
- CaltechAUTHORS:20190314-142000675
- Defense Advanced Research Projects Agency (DARPA)
- AFOSR-90-0199
- Air Force Office of Scientific Research (AFOSR)
- ENG-8711673
- NSF
- NASA/JPL/Caltech
- Created
-
2019-03-14Created from EPrint's datestamp field
- Updated
-
2021-11-16Created from EPrint's last_modified field