I revised the readings list and made Manning et al chapter as the main reading for the latent semantic indexing. Although my own lecture is not based on that specific chapter, I think the chapter provides the
best stand-alone short intro. For the mathematically brave, I also put in a link to a proof of the connection between eigen decomposition and optimal low-rank approximation.
Kepler's laws of planetary motion are here
http://en.wikipedia.org/wiki/Kepler%27s_laws_of_planetary_motion
I was trying to remember the third law (which states"The
squares of the
orbital periods of planets are directly
proportional to the
cubes of the
axes of the orbits.").
The farside explanation of how Einstein discovered special theory of
relativity is here:
http://rakaposhi.eas.asu.edu/cse494/notes/einstein-farside.jpgBefore you laugh at it, consider that historians of science think Kepler probably did discover his laws empirically by poring over (and looking for regularities in) Tycobrahe's planetary motion data.
In fact, one of the first empirical discovery systems in Machine Learning, called Bacon, written by our very own Pat Langley, for his PhD thesis, shows how this approach can be
"automated" (and one of the examples it uses is Kepler's laws--I recall this because I reimplemented bacon system as a class project for my intro to AI course in 1984). Here is a link to a short paper
on Bacon (from 1979) :
http://rakaposhi.eas.asu.edu/cse494/notes/bacon-langley.pdfcheers
Rao
