next up previous
Next: ACKNOWLEDGEMENTS Up: DISCOVERING PREDICTABLE CLASSIFICATIONS (Neural Previous: FINDING PREDICTABLE DISTRIBUTED REPRESENTATIONS

CONCLUSION

In contrast to the principled approach embodied by IMAX, our methods (1) tend to be simpler (e.g., do not require sequential layer by layer `bootstrapping' or learning rate adjustments - the stereo task can be solved more readily by our system), (2) do not require Gaussian assumptions about the input or output signals, (3) do not require something like $DET$MAX, (4) partly have (unlike $DET$MAX) a potential for creating classifications with statistically independent components (this holds for $D_l$ defined according to section 2.4). In addition, our approach makes it easier to decide whether the outputs of both networks essentially represent the same thing.

The experiments above show that the alternative methods of section 2 can be useful for implementing the $D_l$ terms in (4) to obtain predictable informative input transformations. More experiments are needed, however, to become clear about their mutual advantages and disadvantages. It also remains to be seen how well the methods of this paper scale to larger problems.



Juergen Schmidhuber 2003-02-13