Having worked with them and having programmed a rudimentary 2-layer neural net in college, I will personally never consider using them for investment purposes.
It’s been said in the AI community that neural nets are always the ‘second best’ solution. Meaning, that a well-groomed expert system will always beat them.
That said, they still have some value in tuning/optimizing existing systems. But I think the focus is more on genetic algorithms these days.
