András Kornai - Biography#

Professor Kornai earned his mathematics degree in 1983 from Eötvös Loránd University, Budapest where his advisor was Miklós Ajtai. His thesis work became the basis of his joint work with G Pullum "The X-bar theory of phrase structure", Language (1990). He earned his linguistics degree in 1991 from Stanford University, where his advisor was Paul Kiparsky. His thesis work "Formal Phonology" was published in the "Outstanding Dissertations in Linguistics" series by Garland (1995), and is republished in 2nd edition by Routledge (2018). He brought both finite state and probabilistic techniques to bear on fundamental issues of theoretical linguistics such as placing natural languages in the Chomsky hierarchy, poverty of stimulus, and the issue of infinitude.

Besides academia, Kornai worked extensively in industrial research, both at large companies (IBM, BBN) and at startups, progressing from individual researcher to senior scientist, principal scientist, and for his last three industrial appointments, chief scientist. He works the entire gamut of natural language processing from Optical Character Recognition (where he is considered a pioneer of extending Hidden Markov techniques, already broadly used in speech recognition, to OCR) to information extraction and information retrieval (search engines).

He is full professor at the Budapest Institute of Technology, and Senior Scientific Advisor at the Computer and Automation Research Institute of the Hungarian Academy of Sciences. For the past decade his research focuses on bringing the methods of algebra to semantics, see his upcoming monograph http://kornai.com/Drafts/sem.pdf. The technological impact of his work is seen in the HunSpell library built by L. Nemeth under his direction (over 100m downloads, standard in Firefox, Thunderbird, and Open/Libre Office), the HunPOS part of speech tagger (P Halacsy, Cs Oravecz), the HunAlign parallel corpus aligner (D Varga), and the related Hun* toolchain (see http://hlt.bme.hu/en/software).