[GitHub] sxjscience opened a new issue #9172: Wrong gradient of gather_nd when the indices have duplicates

[GitBox <gi...@apache.org>]

sxjscience opened a new issue #9172: Wrong gradient of gather_nd when the indices have duplicates
URL: https://github.com/apache/incubator-mxnet/issues/9172
 
 
   This issue is borrowed from https://discuss.gluon.ai/t/topic/3389.
   
   I find the cause is a bug in the gradient computation of `gather_nd `. Currently, the gradient of `gather_nd ` is `scatter_nd`, which has not considered the case that the indices may be the same
   https://github.com/apache/incubator-mxnet/blob/master/src/operator/tensor/indexing_op.h#L1156 . The correct way to implement it is to use the same backward logic as `take`.
   
   ```python
   import mxnet as mx
   import mxnet.ndarray as nd
   import numpy as np
   
   data = mx.nd.array([[0, 1], [2, 3]])
   indices = mx.nd.array([[1, 1, 0], [1, 1, 0]])
   data.attach_grad()
   with mx.autograd.record():
       ret = mx.nd.gather_nd(data, indices)
       loss = mx.nd.sum(ret)
   loss.backward()
   print(data.grad)
   ```
   
   ```
   [[ 1.  0.]
    [ 0.  1.]]
   <NDArray 2x2 @cpu(0)>
   ```
   
   The correct result should be
   ```
   [[ 1.  0.]
    [ 0.  2.]]
   <NDArray 2x2 @cpu(0)>
   ```
   
   @piiswrong @szha 

----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
users@infra.apache.org


With regards,
Apache Git Services