Information Theory and Statistics

by Solomon Kullback

Highly useful text studies the logarithmic measures of information and their application to testing statistical hypotheses. Topics include introduction and definition of measures of information, their relationship to Fisher's information measure and sufficiency, fundamental inequalities of information theory, much more. Numerous worked examples and problems. References. Glossary. Appendix.

Genres: Mathematics

432 pages, Paperback

First published July 7, 1997

Reviews

[Liquidlasagna · Rating 5 out of 5]

good man


Solomon Kullback

Born April 3, 1907 Brooklyn, New York, United States
Died August 5, 1994 (aged 87) Boynton Beach, Florida

Alma mater
City College of New York (B.A., 1927; M.A., 1929)
George Washington University (Ph.D., Mathematics, 1934)

Known for
Work in Information theory, Kullback–Leibler divergence

Scientific career
Fields cryptanalysis, mathematics, information theory

Institutions
George Washington University - National Security Agency

Kullback attended Boys High School in Brooklyn. He then went to City College of New York, graduating with a BA in 1927 and an MA in math in 1929.

He completed a doctorate in math from George Washington University in 1934. His intention had been to teach, and he returned to Boy's High School to do so, but found it not to his taste; he discovered his real interest was using mathematics, not teaching it.

At the suggestion of Abraham Sinkov, who showed him a Civil Service flyer for 'junior mathematicians' at $2,000 per year, he took the examination. Both passed, and were assigned to Washington, D.C. as junior cryptanalysts.

Upon arrival in Washington, Kullback was assigned to William F. Friedman. Friedman had begun an intensive program of training in cryptology for his new civilian employees. For several summers running, the SIS cryptanalysts attended training camps at Fort Meade until they received commissions as reserve officers in the Army.

Kullback and Sinkov took Friedman's admonitions on education seriously and spent the next several years attending night classes; both received their doctorates in mathematics.

Afterward, Kullback rediscovered a love of teaching; he began offering evening classes in mathematics at George Washington University from 1939.

Once they had completed the training, the three were put to the work for which they had actually been hired, compilations of cipher or code material for the U.S. Army. Another task was to test commercial cipher devices which vendors wished to sell to the U.S. government.

Kullback worked in partnership with Frank Rowlett against RED cipher machine messages. Almost overnight, they unravelled the keying system and then the machine pattern – with nothing but the intercepted messages in hand. Using the talents of linguist John Hurt to translate text, SIS started issuing current intelligence to military decision-makers.

In May 1942, five months after attack on Pearl Harbor, Kullback, by then a Major, was sent to Britain.

He learned at Bletchley Park that the British were producing intelligence of high quality by exploiting the Enigma machine. He also cooperated with the British in the solution of more conventional German codebook-based systems. Shortly after his return to the States, Kullback moved into the Japanese section as its chief.

When the National Security Agency (NSA) was formed in 1952, Rowlett became chief of cryptanalysis. The primary problem facing research and development in the post-war period was development of high-speed processing equipment.

Kullback supervised a team of about 60 people, including such innovative thinkers in automated data processing development as Leo Rosen and Sam Snyder.

His staff pioneered new forms of input and memory, such as magnetic tape and drum memory, and compilers to make machines truly "multi-purpose".

Kullback gave priority to using computers to generate communications security (COMSEC) materials.

Kullback's book Information Theory and Statistics was published by John Wiley & Sons in 1959. The book was republished, with additions and corrections, by Dover Publications in 1968.