I can’t speak for his personality, but I’m certainly a big fan of Schmidhuber’s work. It’s refreshing to see the stream of small, concrete results in AI (e.g. machine translation, traffic sign identification, etc.) scattered with the occasional big, abstract idea (goedel machines, speed prior, etc.). I don’t necessarily agree with some of the more philosophical musings, but I can respect his willingness to take extrapolations seriously.
One thing lacking in this article is giving credit to the laundry list of students and collaborators that have worked with Schmidhuber, e.g. Sepp Hochreiter in the case of LSTM. In research it’s rarely clear from the outside how much each co-author contributed to a result, but when I see Schmidhuber’s name on all these impressive papers I’d sooner assume that he’s an effective facilitator, coordinator, “ideas man”, sounding board, etc. rather than a singular genius who’s swept up unwitting co-authors in his wake ;)
I can’t speak for his personality, but I’m certainly a big fan of Schmidhuber’s work. It’s refreshing to see the stream of small, concrete results in AI (e.g. machine translation, traffic sign identification, etc.) scattered with the occasional big, abstract idea (goedel machines, speed prior, etc.). I don’t necessarily agree with some of the more philosophical musings, but I can respect his willingness to take extrapolations seriously.
One thing lacking in this article is giving credit to the laundry list of students and collaborators that have worked with Schmidhuber, e.g. Sepp Hochreiter in the case of LSTM. In research it’s rarely clear from the outside how much each co-author contributed to a result, but when I see Schmidhuber’s name on all these impressive papers I’d sooner assume that he’s an effective facilitator, coordinator, “ideas man”, sounding board, etc. rather than a singular genius who’s swept up unwitting co-authors in his wake ;)