Computational Information Geometry Wonderland

Edwin Thompson Jaynes is nowadays mostly known for the maximum entropy principle exposed in the 1957 papers:

@article{jaynes-1957,
  author = {Edwin Thompson Jaynes},
  journal = {The Physical Review},
  number = 4,
  pages = {620--630},
  publisher = {American Physical Society},
  title = {Information Theory and Statistical Mechanics},
  volume = 106,
  year = 1957,
  numpages = {10},
  doi = {10.1103/PhysRev.106.620},
  month = may
}

There it is shown that distributions with expectation constraints maximizing Shannon entropy are precisely exponential families. There a nice biography here.

Jaynes work was profound as he brought a new viewpoint of statistical mechanics as inference based on incomplete information.

Note that now-famous papers (his 1957 Phys. rev. papers) was objected by a reviewer then.