machine_learning

Superintelligence - A critique of Nick Bostroms arguments

Why is everyone afraid of artificial intelligence? I'm more afraid of natural human stupidity (which is already infinite according to Einstein!).  Superintelligence would just balance that out ;-)But seriously I think Nick Bostrom has some flawed arguments in his talk:1. He assumes that the work which is necessary to create a higher and higher intelligent being is linear (or even sublinear) with the IQ. But maybe this is not the case.

Can a computer ask questions

Computers are useless. They can only give you answers. (Pablo Picasso) Is Picasso correct? In the scientific literature the question if a computer can be made to ask questions is hardly discussed. However literature containing the following terms implicitly discusses this topic: Unsupervised learning Semi-supervised learning Active learning Bibliography [http://evanjones.ca/picasso.html] Olsson (2009) : A literature survey of active machine learning in the context of natural language processing Castro et al.

Data Mining Algorithms

Bibliography Wu (2008):Top 10 algorithms in data mining Segaran (2007): Programming Collective Intelligence: Building Smart Web 2.0 Applications

Knowledge Paths

The process of knowledge generation involves finding links or connections between already existing pieces of knowledge. There are some applications out there, that can help with that: PaperCube PathWay, see also CNET: New Mac Review: Keep Wikipedia research focused with Pathway Also helpful would be a Google Scholar API, but it does not exist yet: google-ajax-apis: Issue 109: Please add Google Scholar to the API, including # of citations and BibTex reference.

Knowledge generation machines

At the moment we can use internet search machines to find knowledge, that was explicitly written on some webpage. But in my opinion the future of internet search will not be in finding better algorithms to search for this explicit knowledge. Instead I expect a paradigm shift in that future search machines will generate new knowledge, by discovering the implicit associations between webpages. In that sense future search machines will also be knowledge generation machines!

Netflix prize

The Netflix Prize will be over soon. As possibly best algorithms emerged: Support Vector Machine (SVN) Restricted Boltzmann Machine (RBM) k-nearest neighbors algorithm (KNN) Bibliography Was macht den perfekten Film aus? The million dollar programming prize [http://www.netflixprize.com//community/index.php Netflix Prize Forum] [http://www.netflixprize.com//community/viewtopic.php?id=1056 Netflix Prize Forum: Best Single Algorithm] [http://www.timelydevelopment.com/demos/NetflixPrize.aspx Timely Development: Netflix Prize] [http://www.netflixprize.com//leaderboard Netflix price leaderboard] [http://hunch.net/?p=827 Netflix nearly done] Netflix 2.0