Applications in finite state automata Presentation Topics and First - - PowerPoint PPT Presentation

People Possible topics First Steps

Applications in finite state automata

Presentation Topics and First Steps Kurt Eberle

kurt.eberle@uni-tuebingen.de

(includes material from Karttunen, Beesley, Butt and others)

November 3, 2016

1 / 14

People Possible topics First Steps

Outline

People Possible topics First Steps

2 / 14

People Possible topics First Steps

Goals of this session

Xerox people Possible topics First steps: Karttunen’s presentation FS methods in NLP Linguistic society of America, Summer institute 2005

3 / 14

People Possible topics First Steps

Outline

People Possible topics First Steps

4 / 14

People Possible topics First Steps

People

Lauri Karttunen

◮ Semantics and Morphology ◮ Xerox Research Centre Europe (Meylan, Grenoble) ◮ Palo Alto Research Center (PARC) ◮ Xerox Finite State Toolkit ◮ Many awards

5 / 14

People Possible topics First Steps

People

Kimmo Koskenniemi

◮ Two-level morphology :

a general computational model for word-form recognition and production

◮ originally for Finnish ◮ cooperation with the Xerox people

6 / 14

People Possible topics First Steps

Martin Kay

◮ Cambridge Language Research Unit ◮ 1961 Rand Corporation in Santa Monica, California, head of

research in linguistics and machine translation.

◮ 1972 Chair of the Department of Computer Science at the

University of California, Irvine.

◮ 1974 Xerox Palo Alto Research Center, Research Fellow. ◮ 1985,in addition, Professor of Linguistics at Stanford University and

Honorary Professor of Computational Linguistics at Saarland University.

◮ very influential in the field (advisor in many projects,

e.g. Verbmobil)

◮ finite-state morphology ◮ MT : (The Proper Place of Men and Machines in Language

Translation)

7 / 14

People Possible topics First Steps

Ron Kaplan

◮ Research Fellow at the Palo Alto Research Center ◮ Chief Scientist and a Principal Researcher at the Powerset division

• f Microsoft Bing.

◮ Distinguished Scientist at Nuance Communications. ◮ Vice President at Amazon.com ◮ Chief Scientist for Amazon Search ◮ Adjunct Professor in the Linguistics Department at Stanford

University

◮ Lexical Functional Grammar (LFG) (with Joan Bresnan) ◮ Development (with Martin Kay) of the mathematical, linguistic, and

computational concepts that underlie the use of finite-state phonological and morphological descriptions.

◮ helped to embed finite-state methods in a wide range of commercial

products

• ffered by Xerox and Xerox spin-off companies: Microlytics, Inxight,

and Scansoft.

8 / 14

People Possible topics First Steps

Outline

People Possible topics First Steps

9 / 14

People Possible topics First Steps

Possible topics

Papers

◮ Chomsky & Halle 1968 The Sound Pattern of English

(excerpt, basic features of the theory - transformational setting)

◮ Koskenniemi 1983: Two-level morphology ◮ Karttunen 1993: Finite-State Constraints, The Last

Phonological Rule.

• J. Goldsmith (ed.), pages 173-194,

◮ Gregory T. Stump 2001: Inflectional Morphology. A Theory of

Paradigm Structure. Cambridge U. Press. 2001. (An excerpt)

◮ Johnson, C. D. 1972: Formal Aspects of Phonological

Description

◮ Kaplan & Kay 1981: Phonological rules and finite-state

transducers

◮ others

10 / 14

People Possible topics First Steps

Possible topics

Chapters of FSM

◮ Testing and Debugging ◮ Flag Diacritics ◮ Non-concatenative morphotactics ◮ Tokenize and lookup utility

11 / 14

People Possible topics First Steps

Possible topics

Presentations of Problems and solutions

◮ Brazilian Portuguese Pronunciation ◮ Esperanto Noun, Adjectives and Verbs ◮ others . . .

12 / 14

People Possible topics First Steps

Outline

People Possible topics First Steps

13 / 14

People Possible topics First Steps

First Steps

◮ Xerox software ◮ Presentation Karttunen: FS Methods in NLP

Lecture 1

◮ Morphotactics ◮ Phonological alternations ◮ Morphology is regular ◮ Compilation into FSA ◮ Generation, Analysis, Transducers ◮ models ◮ xfst examples ◮ regular expressions 14 / 14