INFO/CS 4300: Language and Information, Spring 2016

Time and place TuTh 2:55pm-4:10pm, Hollister Hall B14
Instructor Prof. Cristian Danescu-Niculescu-Mizil --- Office hours: TuTh 4:30-5:30pm, you need to make an appointment
Head PhD TA Jack Hessel "jhessel@"+".".join(["cs", "cornell", "edu"]) --- Office hours: M 4:30-6:30PM Gates G17
PhD TA Xanda Schofield "xanda@"+".".join(["cs", "cornell", "edu"]) --- Office hours: W 2:30-4:30pm Gates G21
Course Piazza page https://piazza.com/cornell/spring2016/cs4300info4300/home
Course homepage http://www.cs.cornell.edu/Courses/cs4300/2016sp/
Summary How can we make sense of the vast amounts of information available online, and how do we relate it to the social context in which it appears? This course introduces basic tools for retrieving and analyzing unstructured textual information from the web and social media. Applications include information retrieval (with human feedback), question answering, sentiment analysis and social analysis of text. The coursework will include small projects that play on the interaction between knowledge and social factors.
Prerequisites:
INFO 2950 (or CS 2800 and a linear algebra course)
INFO 3300 (or a machine learning course)
Good Python programming skills and familiarity with IPython Notebooks, of which we’ll make extensive use.
Related courses offered this semester at Cornell:
CS 5740 Natural Language Processing
CS 4786/5786 Machine Learning for Data Science
CS 4850: Mathematical Foundations for the Information Age
Textbooks:
- [MRS] Manning, Raghavan, and Schutze. 2008. Introduction to Information Retrieval. Cambridge University Press.
- [J&M] Jurafsky and Martin. 2009. Speech and Language Processing (2nd Edition). Pearson.
Syllabus

Lecture and assignment information will be posted often on this website. Check back soon.

Thu, Jan 28, 2016

Lecture topics:

Intro: Dimensions of Information Systems

Conversational Behavior and Social information

Related material:

Telephone

NPR story: Before The Internet, Librarians Would 'Answer Everything' — And Still Do

Trailer for the movie "Her"

References:

Cristian Danescu-Niculescu-Mizil, Lillian Lee, Bo Pang and Jon Kleinberg. Echoes of power: Language effects and power differences in social interaction.
Proceedings of WWW 2012.

Cristian Danescu-Niculescu-Mizil, Michael Gamon and Susan Dumais. Mark my words! Linguistic style accommodation in social media. Proceedings of WWW, 2011.

Kate G. Niederhoffer and James W. Pennebaker. Linguistic Style Matching in Social Interaction. Journal of Language and Social Psychology 2002 21: 337.

Assignment 1 out [Description, ZIP]

First part due on Thursday Feb 4 at noon.

Second part due on Thursday Feb 11 at noon. You might want this pickle

Tue, Feb 2, 2016

Lecture topics:

Text similarity measures: Minimum Edit Distance

Edit Distance worksheet (includes sketch of the Wagner Fisher algorithm we used in class)

Related material:

Readings:

J&M Chapters 3.11

Th, Feb 4, 2016

No lecture

Second part of Assignment 1 due on Thursday Feb 11 at noon. You might want this pickle

Tu, Feb 9, 2016

Basic text processing concepts: Sentence Splitting, Word Tokenization, Types, Tokens

Text similarity measures: Overlap, Jaccard similarity

Classic (ad hoc) information retrieval systems

Vector space model: binary representation

In-class demo: Proto Information Retrieval System: IPython notebook and html

Vector space model cheatsheet (useful to keep track of notation)

Related material:

Online Porter Stemmer Demo

Readings:

J&M Chapters 3.8 and 23.1.1

Th, Feb 11, 2016

Lecture topics:

Vector Space Model: geometric intuition

Cosine similarity

Inverse document frequency (IDF)

TF-IDF weighting

Pivot length normalization

In-class demo: (continued and updated) IPython notebook and html

Readings:

MRS Chapters 6.2, 6.3, 6.4.1 and 6.4.4

Th, Feb 18, 2016

Lecture topics:

Assigment 1 discussion

Inverted Index

Posting merge algorithm

Boolean search

In-class demo: (continued and updated) IPython notebook and html

Related Material:

Readings:

MRS Chapter 1

Assigment 2 out

[Description, ZIP]

Due: Wednesday, March 2, 11:59pm

Tu, Feb 23, 2016

Lecture topics:

Efficient cosine similarity scoring using the inverted index

Fast cosine retrieval worksheet (includes sketch of the algorithm using the inverted index)

In-class demo: (continued and updated) IPython notebook and html

Before optimizing retrieval with inverted indexes (one query on a collection of 40,000 reality TV utterances): class start

After optimizing retrieval with inverted indexes (one query on a collection of 40,000 reality TV utterances): class end

Readings:

MRS Chapter 6.3.3

Th, Feb 25, 2016

Lecture topics:

Evaluation of ranked retrieval systems: Precision@k, Precision-recall curve

Search at Facebook (Guest speaker Ves Stoyanov)

Thinking about evaluation metrics worksheet

In-class demo: IPython notebook and html

Readings:

MSR Chapter 8

Tu, Mar 1, 2016

Lecture topics:

Evaluation of ranked retrieval systems: Mean Average Precision, Pooling, K-statistic

Relevance feedback

In-class demo: IPython notebook and html

Related material

Data with relevance judgements: TREC

Readings:

MSR Chapters 8, 9.1

Th, Mar 3, 2016

#10

Lecture topics:

Rocchio's method for query rewriting, Pseudo Relevance feedback

Query expansion, Coocurence matrix

Query update using relevance feedback worksheet (includes the Rocchio query update rule)

Readings:

MSR Chapter 9

Assigment 3 out

[Description, ZIP]

Due: Wednesday, March 9, 11:59pm

Midterm date: March 15, durring class time

Tu, Mar 8, 2016

#11

Lecture topics:

Term-document matrix recap, Pointwise Mutual Information

Readings:

MSR Chapter 9

Th, Mar 10, 2016

#12

Lecture topics:

Wrapping up ad hoc IR

Tu, Mar 15, 2016

#13

Lecture topics:

In-class midterm

Th, Mar 17, 2016

#14

Lecture topics:

Midterm discussion

Project discussion

Project milestone 1

Due date 1 (piazza): Monday, March 21 at midnight

Due date 2 (CMS): Thursday, March 23 at midnight

Tu, Mar 22, 2016

#15

Lecture topics:

Text Mining

Th, Mar 24, 2016

#16

Lecture topics:

Practical text mining (by Xanda Schofield)

Slides with text mining tips and libraries

Tu, Apr 5, 2016

#17

Lecture topics:

Text mining, naive bayes, generative models

In-class demo: IPython notebook and html

Th, Apr 7, 2016

#18

Lecture topics:

One on one project meetings

Tu, Apr 12, 2016

#19

Lecture topics:

Feature selection, Conversational Features, Ordinal ranking

In-class demo: IPython notebook and html

Th, Apr 14, 2016

#20

Lecture topics:

Practical unsupervised learning on textual data: SVD (by Jack Hessel)

Slides

In-class demo: IPython notebooks Data exploration, LSI

Tu, Apr 19, 2016

#21

Lecture topics:

Practical unsupervised learning on textual data: Latent semandic indexing and topic modeling (by Jack Hessel)

Slides

In-class demo: IPython notebooks Kickstarter success prediction

Related material:

Indexing by latent semantic analysis. Deerwester, Dumais and Harshman 1990

Th, Apr 21, 2016

#22

Lecture topics:

Opinions and Trust: Link Analisys, Hubs and Authorities, Spectral Analysis

Related material:

NetworkX python package for link analysis

Reading:

Chapters 14.2 & 14.6 A from Networks, Crowds, and Markets

Tu, Apr 26, 2016

#23

Project Prototype Madness (project links posted on Piazza): fun, fun, fun!

Th, Apr 28, 2016

#24

Opinions and Trust: Sentiment Analyis , Lexicon Expansion

Related Material:

On Piazza, April 23

Tu, May 3, 2016

#25

Project presentations

Wed, May 5, 2016

#26

Project presentations

Th,May Apr 10, 2016

#27

Lecture topics:

Opinions and Trust: Review Helpfulness, Social Aspects of Helpfulness Evaluation, Deception Analysis

Course wrap-up

Telephone

Related Material:

Fin.

Date	Lecture	Agenda	Assignments
Thu, Jan 28, 2016	#1	Lecture topics: Intro: Dimensions of Information Systems Conversational Behavior and Social information Related material: NPR story: Before The Internet, Librarians Would 'Answer Everything' — And Still Do Trailer for the movie "Her" References: Cristian Danescu-Niculescu-Mizil, Lillian Lee, Bo Pang and Jon Kleinberg. Echoes of power: Language effects and power differences in social interaction. Proceedings of WWW 2012. Cristian Danescu-Niculescu-Mizil, Michael Gamon and Susan Dumais. Mark my words! Linguistic style accommodation in social media. Proceedings of WWW, 2011. Kate G. Niederhoffer and James W. Pennebaker. Linguistic Style Matching in Social Interaction. Journal of Language and Social Psychology 2002 21: 337.	Assignment 1 out [Description, ZIP] First part due on Thursday Feb 4 at noon. Second part due on Thursday Feb 11 at noon. You might want this pickle
Tue, Feb 2, 2016	#2	Lecture topics: Text similarity measures: Minimum Edit Distance Edit Distance worksheet (includes sketch of the Wagner Fisher algorithm we used in class) Related material: Online Edit Distance calculator Useful regex tool Readings: J&M Chapters 3.11
Th, Feb 4, 2016	#3	No lecture	Second part of Assignment 1 due on Thursday Feb 11 at noon. You might want this pickle
Tu, Feb 9, 2016	#4	Basic text processing concepts: Sentence Splitting, Word Tokenization, Types, Tokens Text similarity measures: Overlap, Jaccard similarity Classic (ad hoc) information retrieval systems Vector space model: binary representation In-class demo: Proto Information Retrieval System: IPython notebook and html Vector space model cheatsheet (useful to keep track of notation) Related material: Online Porter Stemmer Demo Readings: J&M Chapters 3.8 and 23.1.1
Th, Feb 11, 2016	#5	Lecture topics: Vector Space Model: geometric intuition Cosine similarity Inverse document frequency (IDF) TF-IDF weighting Pivot length normalization In-class demo: (continued and updated) IPython notebook and html Readings: MRS Chapters 6.2, 6.3, 6.4.1 and 6.4.4
Th, Feb 18, 2016	#6	Lecture topics: Assigment 1 discussion Inverted Index Posting merge algorithm Boolean search In-class demo: (continued and updated) IPython notebook and html Related Material: The scale at which real IR systems operate. NY Times article emphasizing importance of data pre-processing David Mimno's take on it: Data carpentry Readings: MRS Chapter 1	Assigment 2 out [Description, ZIP] Due: Wednesday, March 2, 11:59pm
Tu, Feb 23, 2016	#7	Lecture topics: Efficient cosine similarity scoring using the inverted index Fast cosine retrieval worksheet (includes sketch of the algorithm using the inverted index) In-class demo: (continued and updated) IPython notebook and html Before optimizing retrieval with inverted indexes (one query on a collection of 40,000 reality TV utterances): After optimizing retrieval with inverted indexes (one query on a collection of 40,000 reality TV utterances): Readings: MRS Chapter 6.3.3
Th, Feb 25, 2016	#8	Lecture topics: Evaluation of ranked retrieval systems: Precision@k, Precision-recall curve Search at Facebook (Guest speaker Ves Stoyanov) Thinking about evaluation metrics worksheet *In-class demo:* IPython notebook and html Readings: MSR Chapter 8
Tu, Mar 1, 2016	#9	Lecture topics: Evaluation of ranked retrieval systems: Mean Average Precision, Pooling, K-statistic Relevance feedback *In-class demo:* IPython notebook and html Related material Data with relevance judgements: TREC Readings: MSR Chapters 8, 9.1
Th, Mar 3, 2016	#10	Lecture topics: Rocchio's method for query rewriting, Pseudo Relevance feedback Query expansion, Coocurence matrix Query update using relevance feedback worksheet (includes the Rocchio query update rule) Readings: MSR Chapter 9	Assigment 3 out [Description, ZIP] Due: Wednesday, March 9, 11:59pm Midterm date: March 15, durring class time
Tu, Mar 8, 2016	#11	Lecture topics: Term-document matrix recap, Pointwise Mutual Information Readings: MSR Chapter 9
Th, Mar 10, 2016	#12	Lecture topics: Wrapping up ad hoc IR
Tu, Mar 15, 2016	#13	Lecture topics: In-class midterm
Th, Mar 17, 2016	#14	Lecture topics: Midterm discussion Project discussion	Project milestone 1 Due date 1 (piazza): Monday, March 21 at midnight Due date 2 (CMS): Thursday, March 23 at midnight
Tu, Mar 22, 2016	#15	Lecture topics: Text Mining
Th, Mar 24, 2016	#16	Lecture topics: Practical text mining (by Xanda Schofield) Slides with text mining tips and libraries
Tu, Apr 5, 2016	#17	Lecture topics: Text mining, naive bayes, generative models *In-class demo:* IPython notebook and html
Th, Apr 7, 2016	#18	Lecture topics: One on one project meetings
Tu, Apr 12, 2016	#19	Lecture topics: Feature selection, Conversational Features, Ordinal ranking *In-class demo:* IPython notebook and html
Th, Apr 14, 2016	#20	Lecture topics: Practical unsupervised learning on textual data: SVD (by Jack Hessel) Slides *In-class demo:* IPython notebooks Data exploration, LSI
Tu, Apr 19, 2016	#21	Lecture topics: Practical unsupervised learning on textual data: Latent semandic indexing and topic modeling (by Jack Hessel) Slides *In-class demo:* IPython notebooks Kickstarter success prediction Related material: Indexing by latent semantic analysis. Deerwester, Dumais and Harshman 1990
Th, Apr 21, 2016	#22	Lecture topics: Opinions and Trust: Link Analisys, Hubs and Authorities, Spectral Analysis Related material: NetworkX python package for link analysis Reading: Chapters 14.2 & 14.6 A from Networks, Crowds, and Markets
Tu, Apr 26, 2016	#23	Project Prototype Madness (project links posted on Piazza): fun, fun, fun!
Th, Apr 28, 2016	#24	Opinions and Trust: Sentiment Analyis , Lexicon Expansion Related Material: NLTK part of speech tagging NLTK WordNet Interface Predicting the Semantic Orientation of Adjectives WordNet Polarity Label Propagation Method Get out the vote: Determining support or opposition from Congressional floor-debate transcripts	On Piazza, April 23
Tu, May 3, 2016	#25	Project presentations
Wed, May 5, 2016	#26	Project presentations
Th,May Apr 10, 2016	#27	Lecture topics: Opinions and Trust: Review Helpfulness, Social Aspects of Helpfulness Evaluation, Deception Analysis Course wrap-up Related Material: Automatically assesing review helpfulness How opinions are received by online communities New York Times article on fake reviews Finding Deceptive Opinion Spam by Any Stretch of the Imagination Fin.