Natural Language Processing 7 Part of Speech Tagging

1
Natural Language Processing(7)Part of Speech
Tagging

• Dr. Xuan Wang(? ?)
• Intelligence Computing Research Center
• Harbin Institute of Technology Shenzhen Graduate
  School
• Part of Slides from Dr. Mary P. Harper
• ECE, Purdue University
• Part of Slides from Dr. Liu Ting and Dr.Guan Yi
  HIT

2
Part of Speech Tagging
3
Penn Treebank Tag Set
4
Tagging Example
5
Baseline Tagging Accuracy

• In general, if you always select the most likely
  part of speech for a word in the training data,
  it is easy to obtain a 90 success rate largely
  because approximately 50 of the words in a
  typical corpus are not ambiguous. Hence, this is
  a baseline against which a good model must
  measure itself.
• It is possible to do better than 90 accuracy by
  using more information from the sentence a word
  appears in. For example, if a word follows a
  determiner, it is not likely that it is a verb.

6
Information Sources
7
Early Approaches
8
Statistical Approaches (Overview)
9
Tagging and The Noisy Channel
10
Notation
11
A Tagging HMM Model
12
The Equations
13
The Equations
14
The Equations
15
The Equations
16
Estimating Probabilities
17
Estimating Probabilities
18
Obtaining Probabilities
19
Estimation of SyntagmaticProbabilities
20
Estimation of Lexical Probabilities
21
Word/Tag Counts
22
Lexical Probability Estimates
23
The HMM Tagger
24
The Probability of a State Sequence for a Sentence
25
Efficient Tagging
26
(No Transcript)
27
Viterbi Algorithm
28
Viterbi Algorithm
29
(No Transcript)
30
(No Transcript)
31
(No Transcript)
32
(No Transcript)
33
(No Transcript)
34
(No Transcript)
35
(No Transcript)
36
(No Transcript)
37
(No Transcript)
38
(No Transcript)
39
(No Transcript)
40
(No Transcript)
41
(No Transcript)
42
(No Transcript)
43
(No Transcript)
44
(No Transcript)
45
(No Transcript)
46
(No Transcript)
47
(No Transcript)
48
(No Transcript)
49
(No Transcript)
50
(No Transcript)
51
(No Transcript)
52
(No Transcript)
53
(No Transcript)
54
(No Transcript)
55
(No Transcript)
56
(No Transcript)
57
(No Transcript)
58
(No Transcript)
59
(No Transcript)
60
(No Transcript)
61
(No Transcript)
62
(No Transcript)
63
(No Transcript)
64
Worlds Superstar1999 Turing Award Winners

• Frederick P. Brooks, Jr.
• For landmark contributions to computer
  architecture, operating systems, and software
  engineering.

65
Frederick P. Brooks, Jr.

• EDUCATION Ph.D., Harvard University, Applied
  Mathematics (Computer Science), 1956 Howard H.
  Aiken, advisor dissertation The Analytic Design
  of Automatic Data Processing Systems S.M.,
  Harvard University, Applied Mathematics (Computer
  Science), 1955 A.B. summa cum laude, Duke
  University, Physics, 1953. First in class of
  1953.
• TEACHING EXPERIENCE University of North Carolina
  at Chapel Hill, Department of Computer Science
  Kenan Professor of Computer Science, 1975-
  Professor of Computer Science, 1964-75 Chairman,
  1964-1984 founder Twente Technical University,
  Enschede, The Netherlands Visiting Professor,
  1970 Columbia University Adjunct Assistant
  Professor, 1960-61 Vassar College Visiting
  Instructor, 1958 IBM Systems Research Institute,
  Voluntary Education Program, and Summer Student
  Program Teacher, 1957-59

66

• DEVELOPMENT EXPERIENCE IBM Corporation
  Poughkeepsie, New York Corporate Processor
  Manager for Development of System/360 Computer
  Systems, 1961-1965 Manager of Operating
  System/360, 1964-65 Manager, System/360 Hardware
  Development, Data Systems Division, 1961-64
  Systems Planning Manager, Data Systems Division
  (8000 series et al.), 1960-61 Yorktown Heights,
  New York Advisory Engineer, Thomas J. Watson
  Research Center, 1959-60 Poughkeepsie, New York
  Project STRETCH Advisory Engineer, 1958-59
  Staff Engineer, 1956-58 Associate Engineer, 1956
  Professional summer jobs at IBM Endicott, Bell
  Labs, North American Aviation, Marathon Oil Co.,
  1952-56
• COMPUTER DESIGNS IBM System/360 Manager of whole
  project, 1961-64 IBM 8000 Series (never
  produced) Manager of Architecture, 1960-61 IBM
  7950 (Harvest) Instruction set-up, adjustments,
  1957-58 IBM 7030 (Stretch) Instruction
  sequencing, interruption, variable-field-length
  arithmetic, editing, programmable console,
  1956-59 A specialized computer for payroll, Ph.D.
  Dissertation, 1955-56

67
Project Walkthrough
68

• Thanks!