This talk shows how real learning is actually a kind of unlearning. The most powerful kind of adaptive learning we can do is to selectively ignore larger and larger sections of our past and the "experts" informed by it. It seems kind of paradoxical that to learn from our history we must unlearn it, but anyone who goes to a university can tell you that the vast majority of their time is spent unlearning what they previously learned.
I'm looking at this video and all I can think of is Google AdWords and their Conversion Optimizer. For quite some time I've been wondering how it operates and in essence this might be it.
Can you post some results where you have used the online learning framework for a portfolio management scenario? Any results that show performance compared to other approaches such as a Genetic Algorithm (where a gene represents a share in your portfolio chromosome), PGA, or others... Convergence results?
this is not on-line optimization, this is on-line optimization for decision problem
henkauser 3 months ago
Well we need to chose a big problem to put theory in practice...
jlopezj3 10 months ago
very good ,i have something good too...check it out
New4fashion 1 year ago
This talk shows how real learning is actually a kind of unlearning. The most powerful kind of adaptive learning we can do is to selectively ignore larger and larger sections of our past and the "experts" informed by it. It seems kind of paradoxical that to learn from our history we must unlearn it, but anyone who goes to a university can tell you that the vast majority of their time is spent unlearning what they previously learned.
wresing 2 years ago
I'm looking at this video and all I can think of is Google AdWords and their Conversion Optimizer. For quite some time I've been wondering how it operates and in essence this might be it.
svetoslavstefanov 3 years ago
sympathetic resonance > synaptic resonance - T
ion 3 years ago
Can you post some results where you have used the online learning framework for a portfolio management scenario? Any results that show performance compared to other approaches such as a Genetic Algorithm (where a gene represents a share in your portfolio chromosome), PGA, or others... Convergence results?
pav930t 3 years ago 4
I would be really interested in seeing some in-practice results as well.
Macuyiko 3 years ago