Language Modeling and Speech -- Sanjeev Khudanpur <khudanpur@jhu.edu>
language modeling
speech recognition
speech synthesis
multimodal communication
Morphology, Phonology, and Finite-State Methods -- Alex Clark <alexc@cs.rhul.ac.uk>
morphology
phonology
word boundaries
part-of-speech tagging
finite-state methods
transliteration
Parsing -- Anoop Sarkar <anoop@cs.sfu.ca>
parsing
syntactic chunking
NOTES:
"shallow parsing" goes here too, but a separate keyword seems unnecessary.
Possibly we should have reserved this track for parsing algorithms, formal grammars,
syntax, and sent new models and features to the grammar learning track?
Word Senses -- Carlo Strapparava <strappa@itc.it>
word senses
ontologies
Computational Semantics -- Suzanne Stevenson <suzanne@cs.toronto.edu>
computational semantics
lexical semantics
textual entailment
semantic roles
lexical acquisition
paraphrase
NOTES:
There's some overlap with "word senses" track.
Discourse, Dialogue, and Pragmatics -- Lyn Walker <walker@dcs.shef.ac.uk>
discourse
dialogue
pragmatics
coreference resolution
NOTES:
Spoken dialogue systems could go here or to speech, depending on paper focus.
Information Extraction -- Dekang Lin <lindek@google.com>
term extraction
named-entity recognition
information extraction
relation extraction
Generation and Summarization -- Manabu Okumura <oku@pi.titech.ac.jp>
generation
summarization
Text Categorization and Sentiment -- Jan Wiebe <wiebe@cs.pitt.edu>
sentiment analysis
opinion extraction
text categorization
text segmentation
Document Collections (includes IR) -- Hal Daume <me@hal3.name>
large document collections
information retrieval
data mining
knowledge discovery
topic discovery
dimensionality reduction
clustering
scalable algorithms
NOTES:
"word sense discovery" is not a keyword (even thought it appears in the CFP),
since that might go here or to the word senses track, depending on methodology.
Should have included "term identification"?
User Applications (includes QA) -- Tomek Strzalkowski <tomek@albany.edu>, Hans Uszkoreit <uszkoreit@dfki.de>
question answering
user applications
information navigation
educational applications
NOTES:
Other specific applications (e.g., helpdesk, medical) are probably not worth listing.
Email routing, etc., might go either here or to Text Cat.
"user interfaces" is not a keyword (even though it appears in the CFP), since that
could go to User Applications, Discourse/Dialogue, or Speech/Multimodality, depending
on topic.
Machine Translation -- David Chiang <chiang@isi.edu>
machine translation
bilingual alignment
parallel text
NOTES:
Possibly some overlap with Multilinguality, especially for cross-lingual projection.
"synchronous grammars" is not a keyword, since that could go
to MT, Parsing, or Formalisms, depending on focus of paper.
Multilinguality -- Katrin Kirchhoff <katrin@ee.washington.edu>
multilinguality
low-density languages
linguistic resources
Machine Learning Models and Methods -- Michael Collins <mcollins@ai.mit.edu>
machine learning
modeling techniques
NOTES:
Applications of existing ML techniques go to the relevant tracks.
Some ML papers will go to Document Collections, Grammar Learning, or Learning Theory.
Specific subtopics of machine learning are not listed as keywords
(although several are listed in the CFP).
Grammar Learning (empirical) -- Menno van Zaanen <menno@ics.mq.edu.au>
grammar learning (empirical)
syntax induction
automaton induction
NOTES:
Empirical inference of grammars and automata.
Covers both structural induction and parameter estimation.
Covers supervised, unsupervised, and semi-supervised learning techniques.
Does not include statistical parsing papers where training isn't central.
"adaptation" is not a keyword (track depends on paper topic).
Learning Theory and Formalisms -- David McAllester <mcallester@tti-c.org>
learning theory
grammar formalisms
grammar learning (theoretical)
learnability
PAC learning
NOTES:
Other formalisms might go here too, but some will go to
Finite-State or Semantics tracks.
MISCELLANEOUS
NOTES:
Resource development should go to the track whose reviewers would use the resources.
Pure linguistics should go to the track whose reviewers would use/appreciate the ideas.
Work on human language processing or acquisition should go to
the track as if a machine were doing the same task.
Work on language evolution, comp bio, etc. should go to the track
that uses appropriate techniques (e.g., ML).
OTHER KEYWORDS WE SHOULD HAVE USED:
synonymy
lexical taxonomy
error detection/correction [e.g., for errors by non-native speakers]
selectional preferences
Sorted list of all 60 keywords above:
automaton induction --> Grammar Learning (empirical)
bilingual alignment --> Machine Translation
clustering --> Document Collections (includes IR)
computational semantics --> Computational Semantics
coreference resolution --> Discourse, Dialogue, and Pragmatics
data mining --> Document Collections (includes IR)
dialogue --> Discourse, Dialogue, and Pragmatics
dimensionality reduction --> Document Collections (includes IR)
discourse --> Discourse, Dialogue, and Pragmatics
educational applications --> User Applications (includes QA)
finite-state methods --> Morphology, Phonology, and Finite-State Methods
generation --> Generation and Summarization
grammar formalisms --> Learning Theory and Formalisms
grammar learning (empirical) --> Grammar Learning (empirical)
grammar learning (theoretical) --> Learning Theory and Formalisms
information extraction --> Information Extraction
information navigation --> User Applications (includes QA)
information retrieval --> Document Collections (includes IR)
knowledge discovery --> Document Collections (includes IR)
language modeling --> Language Modeling and Speech
large document collections --> Document Collections (includes IR)
learnability --> Learning Theory and Formalisms
learning theory --> Learning Theory and Formalisms
lexical acquisition --> Computational Semantics
lexical semantics --> Computational Semantics
linguistic resources --> Multilinguality
low-density languages --> Multilinguality
machine learning --> Machine Learning Models and Methods
machine translation --> Machine Translation
modeling techniques --> Machine Learning Models and Methods
morphology --> Morphology, Phonology, and Finite-State Methods
multilinguality --> Multilinguality
multimodal communication --> Language Modeling and Speech
named-entity recognition --> Information Extraction
ontologies --> Word Senses
opinion extraction --> Text Categorization and Sentiment
PAC learning --> Learning Theory and Formalisms
parallel text --> Machine Translation
paraphrase --> Computational Semantics
parsing --> Parsing
part-of-speech tagging --> Morphology, Phonology, and Finite-State Methods
phonology --> Morphology, Phonology, and Finite-State Methods
pragmatics --> Discourse, Dialogue, and Pragmatics
question answering --> User Applications (includes QA)
relation extraction --> Information Extraction
scalable algorithms --> Document Collections (includes IR)
semantic roles --> Computational Semantics
sentiment analysis --> Text Categorization and Sentiment
speech recognition --> Language Modeling and Speech
speech synthesis --> Language Modeling and Speech
summarization --> Generation and Summarization
syntactic chunking --> Parsing
syntax induction --> Grammar Learning (empirical)
term extraction --> Information Extraction
text categorization --> Text Categorization and Sentiment
text segmentation --> Text Categorization and Sentiment
textual entailment --> Computational Semantics
topic discovery --> Document Collections (includes IR)
transliteration --> Morphology, Phonology, and Finite-State Methods
user applications --> User Applications (includes QA)
word boundaries --> Morphology, Phonology, and Finite-State Methods
word senses --> Word Senses