CS 7180: Special Topics in Artificial Intelligence

Time and Location: Wednesdays and Fridays from 11:45 am - 1:25 pm in Richards Hall 233

Instructor: Lu Wang, Office 448 WVH

Discussion Forum: Piazza, sign up at piazza.com/northeastern/fall2015/cs7180

Course Description

With the development of Internet, it has never been so easy to obtain and share information from different sources, such as online news outlets or fast-growing social networks. The explosive growth of textual data far outpaces human beings' speed of digesting its content. Natural language processing (NLP) techniques are developed to help people understand, absorb, and learn knowledge from massive amount of textual data in an easy manner.

In this graduate-level research-oriented seminar course, we will discuss and examine the recent advances in NLP. Potential topics include text summarization, sentiment analysis, information extraction, text analysis in online social interactions or political domain, word embedding or representation learning, and deep learning for NLP. We will read papers in those topics, and students will work on course projects in small groups.

Graduate students who are interested in research in natural language processing, machine learning, computational social science, and related areas are encouraged to enroll.

Prerequisites

• Being able to write code in some programming languages proficiently.
• Finishing at least one course in natural language processing, machine learning, information retrieval, or any relevant subfield of artificial intelligence.

Schedule

Sep 9

• Topic
  • Course Setup
• TODO: email instructor your information by next class (Sep 11, 11:45am)
• Handout [download]
• References
  • Speech and Language Processing, by Daniel Jurafsky, James H. Martin
  • Foundations of Statistical Natural Language Processing, by Christopher D. Manning, Hinrich Schütze

Sep 11

• Topic
  • Sentiment Analysis
• TODO: make appointment with instructor on course project (email your availibility on Sep 17, 18, & 21)
• Paper to discuss (Leaders: Jesse & Sri):
  • Sarcasm as Contrast between a Positive Sentiment and Negative Situation, by Riloff et al., EMNLP 2013
  • Open Domain Targeted Sentiment, by Mitchell et al., EMNLP 2013
• References
  • Thumbs Up?: Sentiment Classification Using Machine Learning Techniques, by Pang et al., EMNLP 2002
  • Recognizing Contextual Polarity in Phrase-Level Sentiment Analysis, by Wilson et al., EMNLP 2005
  • Identifying Expressions of Opinion in Context, by Breck et al., IJCAI 2007
  • Predicting Elections with Twitter: What 140 Characters Reveal about Political Sentiment, by Tumasjan et al., ICWSM 2010
  • Conditional Random Fields: Probabilistic Models for Segmenting and Labeling Sequence Data, by Lafferty et al., ICML 2001

Sep 16

• Topic
  • Sentiment analysis & Argumentation mining
• TODO: make appointment with instructor on course project (email your availibility on Sep 17, 18, & 21)
• Paper to discuss (Leaders: Jesse):
  • Exploiting Debate Portals for Semi-Supervised Argumentation Mining in User-Generated Web Discourse, by Habernal and Gurevych, EMNLP 2015, [link]
• References:
  • Argumentation Mining: The Detection, Classification and Structure of Arguments in Text, by Palau and Moens, ICAIL 2009
  • Identifying Argumentative Discourse Structures in Persuasive Essays, by Stab and Gurevych, EMNLP 2014

Sep 18

• Topic
  • NLP for social media
• Paper to discuss (Leaders: Sri & Dina)
  • Finding Deceptive Opinion Spam by Any Stretch of the Imagination, by Ott et al., ACL 2011
  • From Tweets to Polls: Linking Text Sentiment to Public Opinion Time Series, by O'Connor et al., ICWSM 2010
• References:
  • Computational Text Analysis for Social Science: Model Assumptions and Complexity, by O'Connor et al., NIPS workshop 2011
  • Computational Social Science, by Lazer et al., [link]

Sep 23

• Attention: we change our regular time to 1:10pm - 2:50pm, and the location is 164WVH
• TODO: hand in hard copy of project proposal to course instructor (due on Sep 23, 1:10pm)
• Topic
  • Project proposal presentation (1/4)

Sep 25

• Topic
  • Project proposal presentation (2/4)

Sep 30

• Topic
  • Project proposal presentation (3/4)

Oct 2

• Topic
  • Project proposal presentation (4/4)

Oct 7

• Topic
  • Word Embedding & DNN
• Paper to discuss (Leader: Yash)
  • A Neural Probabilistic Language Model, Bengio et al., Journal of Machine Learning Research 2003 [link]

Oct 9

• Topic
  • Word Embedding & DNN
• Paper to discuss (Leader: Kunal)
  • Distributed Representations of Words and Phrases and their Compositionality, Mikolov et al., NIPS 2013 [link]

Oct 14

• Topic
  • Word Embedding & DNN
• Paper to discuss (Leader: Kunal)
  • Sequence to Sequence Learning with Neural Networks, Sutskever et al., NIPS 2014 [link]
• References:
  • Long Short-Term Memory, by Hochreiter and Schmidhuber [link]
  • Understanding LSTM Networks [link]
  • LSTM python code [link]

Oct 16

• Topic
  • Word Embedding & DNN
• Paper to discuss (Leader: Liwen)
  • GloVe: Global Vectors for Word Representation, by Pennington et al., EMNLP 2014 [link]
• References:
  • Website [link]

Oct 21

• Topic
  • NLP for social media
• Paper to discuss (Leader: Liwen)
  • "You're Mr. Lebowski, I'm the Dude": Inducing Address Term Formality in Signed Social Networks, by Krishnan and Eisenstein, NAACL 2015

Oct 23

• Topic
  • NLP for social media
• Paper to discuss (Leader: Jesse)
  • QUOTUS: The Structure of Political Media Coverage as Revealed by Quoting Patterns, by Niculae et al, WWW 2015

Oct 28

• Topic
  • Discussing project progress with instructor (optional, one-on-one, by appointment)

Oct 30

• Topic
  • Project progress presentation
• TODO: hand in hard copy of project proposal to course instructor (due on Oct 30, 11:45am)

Nov 4

• Topic
  • EM revisit (Leader: Kunal)
• Paper to discuss
  • EM tutorial: see Piazza
  • "You're Mr. Lebowski, I'm the Dude": Inducing Address Term Formality in Signed Social Networks, by Krishnan and Eisenstein, NAACL 2015

Nov 6

• Topic
  • Representation Learning & Embeddings
• Paper to discuss (Leader: Dina)
  • Representation Learning for Text-level Discourse Parsing, by Ji and Eisenstein, ACL 2014

Nov 11 (Veterans' Day, no class)

Nov 13

• Topic
  • Computational Social Science
• Paper to discuss (Leader: Liwen)
  • A Frame of Mind: Using Statistical Models for Detection of Framing and Agenda Setting Campaigns, by Tsur et al., ACL 2015

Nov 18 (no class)

Nov 20

• Topic
  • Computational Social Science, Information Extraction
• Paper to discuss (Leader: Sri)
  • The Bayesian Echo Chamber: Modeling Social Influence via Linguistic Accommodation, by Guo et al. AISTATS 2015

Nov 25 & 27 (Thanksgiving, no classes)

Dec 2

• Topic
  • NLP for Social Media
• Paper to discuss (Leader: Dina and Yash)
  • Leveraging Linguistic Structure For Open Domain Information Extraction, by Angeli et al. ACL 2015
  • Predicting Speech Acts in MOOC Forum Posts, by Arguello and Shaffer, ICWSM 2015

Dec 4

• Topic
  • NLP for Social Media
• Paper to discuss (Leader: Yash)
  • Values in Words: Using Language to Evaluate and Understand Personal Values, Boyd et al, ICWSM 2015

Dec 9

• Topic
  • Project final presentation