10.07.2015 Views

Information Theory, Inference, and Learning ... - Inference Group

Information Theory, Inference, and Learning ... - Inference Group

Information Theory, Inference, and Learning ... - Inference Group

SHOW MORE
SHOW LESS
  • No tags were found...

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

Copyright Cambridge University Press 2003. On-screen viewing permitted. Printing not permitted. http://www.cambridge.org/0521642981You can buy this book for 30 pounds or $50. See http://www.inference.phy.cam.ac.uk/mackay/itila/ for links.4.8: Solutions 89Solution to exercise 4.20 (p.86).NoteThe function f(x) has inverse functiong(y) = y 1/y . (4.54)log g(y) = 1/y log y. (4.55)I obtained a tentative graph of f(x) by plotting g(y) with y along the verticalaxis <strong>and</strong> g(y) along the horizontal axis. The resulting graph suggests thatf(x) is single valued for x ∈ (0, 1), <strong>and</strong> looks surprisingly well-behaved <strong>and</strong>ordinary; for x ∈ (1, e 1/e ), f(x) is two-valued. f( √ 2) is equal both to 2 <strong>and</strong>4. For x > e 1/e (which is about 1.44), f(x) is infinite. However, it might beargued that this approach to sketching f(x) is only partly valid, if we define fas the limit of the sequence of functions x, x x , x xx , . . .; this sequence does nothave a limit for 0 ≤ x ≤ (1/e) e ≃ 0.07 on account of a pitchfork bifurcationat x = (1/e) e ; <strong>and</strong> for x ∈ (1, e 1/e ), the sequence’s limit is single-valued – thelower of the two values sketched in the figure.504030201000 0.2 0.4 0.6 0.8 1 1.2 1.45432100 0.2 0.4 0.6 0.8 1 1.2 1.40.50.40.30.20.100 0.2xxxx···Figure 4.15. f(x) = x , shownat three different scales.

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

Saved successfully!

Ooh no, something went wrong!