4 ms·
A question - does anyone utilize the fact observed in cognitive neuroscience in robotics, that the best learning happens in organically growing systems, not in
by bitL 10y ago
A question - does anyone utilize the fact observed in cognitive neuroscience in robotics, that the best learning happens in organically growing systems, not in universal systems where you run a huge batch on a clean, pre-configured state? So far I see everyone designing the structure of an ML system in advance and hoping for the best (e.g. define the whole structure in TensorFlow in the beginning), instead of allowing the system to self-organize as the learning happens.
- tedmiston 10y agoAre you referring to reinforcement learning [1] as opposed to "standard" supervised learning? [1]: https://en.wikipedia.org/wiki/Reinforcement_learning https://en.wikipedia.org/wiki/Reinforcement_learning