This is a paper by E. T. Jaynes and it is called: Information Theory and Statistical Mechanics. It is one of the most influential pioneering papers of this field and I have to read this paper! I suspect much of the content of this paper might already be familiar to me, having transmitted to me through the folklore, but still this would be fun!
And this is a paper by I. Csiszar and it is called: I-divergence Geometry of Probability Distributions and Minimization Problems, which is another classic. This is one of those papers that I have to read very carefully.
p.s.: I recently read this paper: Learning Markov Structure by Maximum Entropy Relaxation. It is very exciting work. I am presenting the paper on Monday to my guide.