Hi,
On Sun, 2008-05-04 at 08:37 +0200, Georg Holzmann wrote:
in the description of the algorithm they talk about some pending patents... I guess they use more than simple ann comparison to get this working in realtime on a game console.
Yes, I also would not use a feedforward neural network (as implemented in ann) for this. You have to use a model which also incorporates time.
Hm. You can incorporate changes over time using a standard feedforward ANN by wrapping your time-ordered vectors over a given time period into a single input vector and increasing the number of inputs to the network accordingly. But of course this introduces latency and other problems (e.g. it could massively increase the number of training examples required).
Pd has the ann_td external, which provides a 'time delay' neural network which I believe incorporates time using a method similar to that described above.
For example a hidden markov model or echo state network (= special kind of recurrent neural network) should work.
I'm intrigued! Presumably these approaches avoid the latency problem by maintaining the network's state? Are there other advantages -- easier to train?
Jamie