next up previous
Next: Scenes With Multiple Objects Up: LEARNING TO GENERATE ARTIFICIAL Previous: Parallel Learning of and

ONGOING AND FUTURE RESEARCH

The approach described above certainly does not solve all problems of adaptive attentive vision. So far our system has been tested only with fairly simple visual scenes involving objects with a rather simple geometrical shape. It is not clear how well the system will do with more complicated scenes. However, there are some promising directions for future research.



Subsections

Juergen Schmidhuber 2003-02-21

Back to Reinforcement Learning page
Back to main page on Learning attentive vision