|
. |
|
|
|
|
We scored the papers of Rina Dechter, Zoubin Ghahramani, Geoffrey Hinton, Daphne Koller, Michael I. Jordan, Tom Mitchell, Padhraic Smyth and Suart Russell. For each document for each of these authors we calculate a perplexity score. Perplexity is widely used in language modeling to assess the predictive power of a model. It is a measure of how surprising the words are from the model's perspective, loosely equivalent to the effective branching factor. Our goal here is not to evaluate the out-of-sample predictive power of the model, but to explore the range of perplexity scores that the model assigns to papers from specific authors. Lower scores imply that the words are less surprising to the model. Note that some of the highly scored papers for Jordan_M as well Koller_D
and Russell_S, in fact are not written by Michael I. Jordan, Daphne Koller
and Sruart Russell but by researchers that have the same standardized name,
Jordan_M, Koller_D and Russell_S,
respectively (these are Mick Jordan and Daniel Koller, Dieter Koller, David
Koller as well as Stephen Russell). Since most papers are indeed writings of
the above researchers from Berkeley and Stanford who write about wide range
of topics, papers not written by them are scored as very untypical. |