cmsc 473/673 natural language processing fall 2018 · december 2016. august 2018. potential...

117
CMSC 473/673 Natural Language Processing Fall 2018

Upload: others

Post on 09-Jul-2020

8 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

CMSC 473/673Natural Language Processing

Fall 2018

Page 2: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Frank Ferraro

ITE 358

[email protected]

Monday: 2:15-3

Tuesday: 11:00-11:30

by appointment

Natural language processing

Semantics

Vision & language processing

Learning with low-to-no supervision

Page 3: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Caroline Kery

Location TBD

[email protected]

Tuesday: 2-3:30pm

Thursday: 1-2:30pm

by appointment

Semantic parsing

Active learning

Data visualization

Analysis of educational data

Page 4: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

December 2016

Page 5: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

August 2018

Page 6: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Potential Applications

ASR (automatic speech recognition)

Machine translation

Natural language generation

Document labeling/classification

Document summarization

Corpus exploration

Relation/information extraction

Entity identification

Page 7: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Automatic speech recognition

Page 8: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

SPORTSDocument classification

Page 9: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Machine translation

Page 10: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

https://cdn.arstechnica.net/wp-content/uploads/2015/11/Screen-Shot-2015-11-02-at-9.11.40-PM-640x543.png

Natural language generation

Page 11: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Document summarization

Page 12: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Corpus exploration

Page 13: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Pat and Chandler agreed on a plan.

He said Pat would try the same tactic again.

Relation extraction

Page 14: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Pat and Chandler agreed on a plan.

He said Pat would try the same tactic again.

Entity identification

Page 15: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Pat and Chandler agreed on a plan.

He said Pat would try the same tactic again.

is “he” the same person as “Chandler?”

?

Entity identification

Page 16: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Course Goals

Be introduced to some of the core problems and solutions of NLP (big picture)

Page 17: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes
Page 18: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Course Goals

Be introduced to some of the core problems and solutions of NLP (big picture)

Learn different ways that success and progress can be measured in NLP

Page 19: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Natural Language Processing

tensorflow

Page 20: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Course Goals

Be introduced to some of the core problems and solutions of NLP (big picture)

Learn different ways that success and progress can be measured in NLP

Relate to statistics, machine learning, and linguistics

Implement NLP programs

Page 21: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Course Goals

Be introduced to some of the core problems and solutions of NLP (big picture)

Learn different ways that success and progress can be measured in NLP

Relate to statistics, machine learning, and linguistics

Implement NLP programs

Read and analyze research papers

Practice your (written) communication skills

Page 22: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

http://www.qwantz.com/index.php?comic=170

Page 23: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

http://www.qwantz.com/index.php?comic=170

Page 24: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Natural Language Processing≈

Computational Linguistics

Page 25: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Natural Language Processing≈

Computational Linguisticsscience focus

computational biocomputational chemistry

computational X

Page 26: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Natural Language Processing≈

Computational Linguisticsscience focus

computational biocomputational chemistry

computational X

build a system to translatecreate a QA system

engineering focus

Page 27: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Natural Language Processing ≈ Computational Linguistics

Machine learning

Page 28: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Natural Language Processing ≈ Computational Linguistics

Machine learning

Information Theory

Page 29: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Natural Language Processing ≈ Computational Linguistics

Machine learning

Information Theory

Data Science

Page 30: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Natural Language Processing ≈ Computational Linguistics

Machine learning

Information Theory

Data Science

Systems Engineering

Page 31: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Natural Language Processing ≈ Computational Linguistics

Machine learning

Information Theory

Data Science

Systems Engineering

Logic

Theory of Computation

Page 32: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Natural Language Processing ≈ Computational Linguistics

Machine learning

Information Theory

Data Science

Systems Engineering

Logic

Theory of Computation

Linguistics

Page 33: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Natural Language Processing ≈ Computational Linguistics

Machine learning

Information Theory

Data Science

Systems Engineering

Logic

Theory of Computation

Linguistics

Cognitive Science

Psychology

Page 34: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Natural Language Processing ≈ Computational Linguistics

Machine learning

Information Theory

Data Science

Systems Engineering

Logic

Theory of Computation

Linguistics

Cognitive Science

Psychology

Political Science

Digital Humanities

Education

Page 35: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Natural Language Processing≈

Computational Linguisticsscience focus

computational biocomputational chemistry

computational X

build a system to translatecreate a QA system

engineering focus

these views can co-exist peacefully

Page 36: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

What Are Words?

Linguists don’t agree

(Human) Language-dependent

White-space separation is a sometimes okay (for written English longform)

Social media? Spoken vs. written? Other languages?

Page 37: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

What Are Words? Tokens vs. Types

The film got a great opening and the film went on to become a hit .

Type: an element of the vocabulary.

Token: an instance of that type in running text.

How many of each?

Page 38: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Terminology: Tokens vs. TypesThe film got a great opening and the film went on to become a hit .

Tokens• The• film• got• a• great• opening• and• the• film• went• on• to• become• a• hit• .

Types• The• film• got• a• great• opening• and• the• went• on• to• become• hit• .

Page 39: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Terminology: Tokens vs. TypesThe film got a great opening and the film went on to become a hit .

Tokens• The• film• got• a• great• opening• and• the• film• went• on• to• become• a• hit• .

Types• The• film• got• a• great• opening• and• the• went• on• to• become• hit• .

Page 40: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

http://www.qwantz.com/index.php?comic=170

Page 41: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Adapted from Jason Eisner, Noah Smith

Page 42: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

orthography

Adapted from Jason Eisner, Noah Smith

Page 43: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

orthography

morphology

Adapted from Jason Eisner, Noah Smith

Page 44: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

orthography

morphology

Adapted from Jason Eisner, Noah Smith

lexemes

Page 45: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

orthography

morphology

Adapted from Jason Eisner, Noah Smith

lexemes

syntax

Page 46: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

orthography

morphology

Adapted from Jason Eisner, Noah Smith

lexemes

syntax

semantics

Page 47: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

orthography

morphology

Adapted from Jason Eisner, Noah Smith

lexemes

syntax

semantics

pragmatics

Page 48: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

orthography

morphology

Adapted from Jason Eisner, Noah Smith

lexemes

syntax

semantics

pragmatics

discourse

Page 49: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Adapted from Jason Eisner, Noah Smith

NLP + Latent Modeling

explain what you see/annotate

with things “of importance” you don’t

orthography

morphology

lexemes

syntax

semantics

pragmatics

discourse

observed text

Page 50: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

orthography

morphology

lexemes

syntax

semantics

pragmatics

discourse

Page 51: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

orthography

morphology

lexemes

syntax

semantics

pragmatics

discourse

VISION

AUDIO

prosody

intonation

color

Page 52: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Language is Productive

Page 53: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes
Page 54: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Watergate

Page 55: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Troopergate

Watergate Bridgegate

Deflategate

Page 56: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Language is Ambiguous

Page 57: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Ambiguity

Kids Make Nutritious Snacks

Page 58: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Ambiguity

Kids Make Nutritious Snacks

Kids Prepare Nutritious Snacks

Kids Are Nutritious Snacks

sense ambiguity

Page 59: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Ambiguity

British Left Waffles on Falkland Islands

Page 60: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Ambiguity

British Left Waffles on Falkland Islands

British Left Waffles on Falkland Islands

British Left Waffles on Falkland Islands

lexical ambiguity

Page 61: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Part of SpeechTagging

British Left Waffles on Falkland Islands

British Left Waffles on Falkland Islands

British Left Waffles on Falkland Islands

Adjective Noun Verb

Noun Verb Noun

lexical ambiguity

Page 62: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Parts of Speech

Classes of words that behave like one another in “similar” contexts

Pronunciation (stress) can differ: object (noun: OB-ject) vs. object (verb: ob-JECT)

It can help improve the inputs to other systems

(text-to-speech, syntactic parsing)

Page 63: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Ambiguity

Pat saw Chris with the telescope on the hill.

I ate the meal with friends.

Page 64: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Ambiguity

Pat saw Chris with the telescope on the hill.

I ate the meal with friends.

syntactic ambiguity

Page 65: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Language Can Be Surprising

Page 66: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Garden PathSentences

Page 67: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Garden PathSentences

The

Page 68: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Garden PathSentences

The old

Page 69: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Garden PathSentences

The old man

Page 70: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Garden PathSentences

The old man the

Page 71: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Garden PathSentences

The old man the boat

Page 72: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Garden PathSentences

The old man the boat .

Page 73: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Garden PathSentences

The old man the boat .

Page 74: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Garden PathSentences

The complex houses married and single soldiers and their families.

Page 75: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Garden PathSentences

The complex houses married and single soldiers and their families.

Page 76: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Garden PathSentences

The rat the cat the dog chased killed ate the malt.

Page 77: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Garden PathSentences

The rat that the cat the dog chased killed ate the malt.

Page 78: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Garden PathSentences

The rat that the cat that the dog chased killed ate the malt.

Page 79: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Garden PathSentences

The rat that the cat that the dog chased killed ate the malt.

Page 80: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Garden PathSentences

The rat that the cat that the dog chased killed ate the malt.

Page 81: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Garden PathSentences

The rat that the cat that the dog chased killed ate the malt.

Page 82: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Garden PathSentences

[The rat [the cat [the dog chased] killed] ate the malt].

Language can have recursive patterns

Syntactic parsing can help identify those

Page 83: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Syntactic Parsing

I ate the meal with friends

NP VP

VP NP PP

S

Syntactic parsing: perform a “meaningful” structural analysis according to grammatical rules

Page 84: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Syntactic Parsing Can Help Disambiguate

I ate the meal with friends

NP VP

VP NP PP

S

Page 85: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Syntactic Parsing Can Help Disambiguate

I ate the meal with friends

NP VP

VP NP PP

S

NP VP

S

VP NP

PPNP

Page 86: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Clearly Show Ambiguity…But Not Necessarily All Ambiguity

I ate the meal with friends

NP VP

VP NP PP

S

I ate the meal with gusto

I ate the meal with a fork

Page 87: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Discourse Processing

John stopped at the donut store.

Courtesy Jason Eisner

Page 88: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Discourse Processing

John stopped at the donut store.

Courtesy Jason Eisner

Page 89: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Discourse Processing

John stopped at the donut store before work.

Courtesy Jason Eisner

Page 90: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Discourse Processing

John stopped at the donut store on his way home.

Courtesy Jason Eisner

Page 91: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Discourse Processing

John stopped at the donut shop.

John stopped at the trucker shop.

John stopped at the mom & pop shop.

John stopped at the red shop.

Courtesy Jason Eisner

Page 92: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Discourse Processing through Coreference

I spread the cloth on the table to protect it.

I spread the cloth on the table to display it.

Courtesy Jason Eisner

Page 93: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

I spread the cloth on the table to protect it.

I spread the cloth on the table to display it.

Courtesy Jason Eisner

Discourse Processing through Coreference

Page 94: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

I spread the cloth on the table to protect it.

I spread the cloth on the table to display it.

Courtesy Jason Eisner

Discourse Processing through Coreference

Page 95: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

http://www.qwantz.com/index.php?comic=170

Page 96: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Three people have been fatally shot, and five people, including a mayor, were seriously wounded as a result of a Shining Path attack today.

Page 97: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Three people have been fatally shot, and five people, including a mayor, were seriously wounded as a result of a Shining Path attack today.

score( )

Page 98: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Three people have been fatally shot, and five people, including a mayor, were seriously wounded as a result of a Shining Path attack today.

pθ( )

Page 99: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

pθ(X)probabilistic model

objective F(θ)

Page 100: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Gradient Ascent

θ2

θ1

Page 101: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Gradient Ascent

θ2

θ1

Page 102: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Gradient Ascent

θ2

θ1

Page 103: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Gradient Ascent

“gradient of F with respect to θ”

θ2

θ1

Page 104: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Gradient Ascent

“gradient of F with respect to θ”

gradient: a vector of derivatives, each with respect to θk while

holding all other variables constant

θ2

θ1

Page 105: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

http://www.qwantz.com/index.php?comic=170

Page 106: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

http://universaldependencies.org/part-of-speech & syntax for > 120 languages

Page 107: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

From Syntax to Shallow Semantics

http://corenlp.run/ (constituency & dependency)

https://github.com/hltcoe/predpatt

http://openie.allenai.org/

http://www.cs.rochester.edu/research/knext/browse/ (constituency trees)

http://rtw.ml.cmu.edu/rtw/

Angeli et al. (2015)

“Open Information Extraction”

a sampling of efforts

Page 108: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Semantic Projection

Page 109: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Administrivia

Page 110: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Grading

Component 473 673

FiveAssignments

45% 30%

Midterm 10% 10%

GraduatePaper

--- 30%

Course Project 45% 30%

Page 111: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Final Grades≥ Letter

90 A

80 B

70 C

65 D

0 F

≥ Letter

90 A-

80 B-

70 C-

65 D

0 F

473

673

Page 112: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

https://www.csee.umbc.edu/courses/undergraduate/473/f18

Page 113: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Online Discussions

https://piazza.com/umbc/fall2018/cmsc473673

Page 114: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Important Dates

Page 115: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Late Policy

Everyone has a budget of 10 late days

If you have them left: assignments turned in after the deadline will be graded and recorded, no questions asked

If you don’t have any left: still turn assignments in. They could count in your favor in borderline cases

Page 116: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Late Policy

Everyone has a budget of 10 late days

Use them as needed throughout the course

They’re meant for personal reasons and emergencies

Do not procrastinate

Page 117: CMSC 473/673 Natural Language Processing Fall 2018 · December 2016. August 2018. Potential Applications ASR (automatic speech recognition) Machine translation ... Noah Smith lexemes

Late Policy

Everyone has a budget of 10 late days

Contact me privately if an extended absence will occur