Slides in PDF - of Marcus Hutter
Slides in PDF - of Marcus Hutter
Slides in PDF - of Marcus Hutter
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
<strong>Marcus</strong> <strong>Hutter</strong> - 40 - Universal Induction & Intelligence<br />
Universal Inductive Inference: Abstract<br />
Solomon<strong>of</strong>f completed the Bayesian framework by provid<strong>in</strong>g a rigorous,<br />
unique, formal, and universal choice for the model class and the prior. I<br />
will discuss <strong>in</strong> breadth how and <strong>in</strong> which sense universal (non-i.i.d.)<br />
sequence prediction solves various (philosophical) problems <strong>of</strong> traditional<br />
Bayesian sequence prediction. I show that Solomon<strong>of</strong>f’s model possesses<br />
many desirable properties: Strong total and weak <strong>in</strong>stantaneous bounds<br />
, and <strong>in</strong> contrast to most classical cont<strong>in</strong>uous prior densities has no zero<br />
p(oste)rior problem, i.e. can confirm universal hypotheses, is<br />
reparametrization and regroup<strong>in</strong>g <strong>in</strong>variant, and avoids the old-evidence<br />
and updat<strong>in</strong>g problem. It even performs well (actually better) <strong>in</strong><br />
non-computable environments.