deeplearning4j/deeplearning4j-examples

Increasing errors with MultiLayerNetworkExternalErrors.java Example

Open

#914 opened on Oct 18, 2019

View on GitHub
 (1 comment) (0 reactions) (0 assignees)Java (2,395 stars) (1,841 forks)batch import
help wantedquestion

Description

Hi,

I am trying to using train a simple 2 layer network on a simple synthetic dataset. With two features X1 and X2. and lables: y = X1^2 + X2^2

I am calculating the error for each epoch manually and using the errors to call backprop, similar to the MultiLayerNetworkExternalErrors.java example

The prediction error is increasing with each epoch, I would have expected decreasing errors with on this simple dataset. Could someone help me with if the issue is with how I am using dl4j methods ?

Code: https://gist.github.com/Saurabh7/9b5ea7def2a167903e7d206e272e2662 Data: https://gist.github.com/Saurabh7/982fb4ddfbd58bd866782ea31b95aef1

Contributor guide

Tech stack
java
Domain
machine learning
Issue type
bug
DifficultyEstimated implementation difficulty for a new contributor, from 1 for very small changes to 5 for expert-level work.
3
Estimated timeA rough time range for an experienced contributor to investigate, implement, test, and prepare a pull request.
1-3 hours
Activity statusHow available the issue appears right now: fresh, active, stale, blocked, or waiting on maintainer input.
stale
ClarityHow clearly the issue explains the expected change, acceptance criteria, and next step.
needs investigation
Prerequisites
Java basicsNeural network training concepts
Newbie friendlinessA 1-100 score estimating how approachable this issue is for first-time contributors.
30
Research direction
Examine the user's code at the gist and data link. Compare with the MultiLayerNetworkExternalErrors.java example in the repository. Determine if the issue stems from incorrect use of dl4j methods or a deeper bug. The increasing error suggests a possible learning rate issue, data normalization problem, or gradient calculation error. Review the external errors implementation and ensure the manual error calculation matches the network's output.