07.01.2013 Views

Lecture Notes in Computer Science 3472

Lecture Notes in Computer Science 3472

Lecture Notes in Computer Science 3472

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

614 Bengt Jonsson<br />

of <strong>in</strong>put symbols and observ<strong>in</strong>g the result<strong>in</strong>g output. An experiment can be<br />

either preset or adaptive. Apreset experiment (or preset sequence) is a fixed<br />

<strong>in</strong>put sequence x ∈ I ∗ , and we are <strong>in</strong>terested <strong>in</strong> observ<strong>in</strong>g the output sequence<br />

produced by the mach<strong>in</strong>e <strong>in</strong> response to x .Inanadaptive experiment (or<br />

adaptive sequence), each symbol <strong>in</strong> the <strong>in</strong>put sequence depends on the output<br />

produced <strong>in</strong> response to the previous <strong>in</strong>put symbols. An adaptive experiment<br />

can be formalized as a decision tree, <strong>in</strong> which the <strong>in</strong>ternal nodes are labeled<br />

with <strong>in</strong>put symbols, and the edges are labeled with output symbols, such that<br />

edges emanat<strong>in</strong>g from a common node have dist<strong>in</strong>ct output symbols. Each leaf<br />

of an adaptive experiment can be labeled with a suitable def<strong>in</strong>ed outcome of the<br />

experiment for the particular case that the experiment ends up <strong>in</strong> this leaf.

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!