CS7800 Information Retrieval

Course Objective

Course Prerequisite

Course Description

This course will cover models for information retrieval, techniques for indexing and searching, and algorithms for classification and clustering. It will also cover SVM, latent semantic indexing, link analysis and ranking, Map-Reduce architecture and Hadoop, and Word Embedding/Language models to different degree of detail, time permitting. 

Course Load

The course load includes programming projects and two tests. The format of the tests will be communicated to you before their administration and they will not be comprehensive. All the recorded lectures will be available in WebEx/Panopto from course webpage on Pilot and slide decks for the lectures from this publically available course webpage (https://cecs.wright.edu/~tkprasad/courses/cs7800/cs7800.html).  Note that  the lectures  and slide decks in Powerpoint  are organized  around  topics and may not clearly  map to  fixed lecture time periods. But they will  cover all the requisite content.

Assignment and Exams Policy

(1) Team work is allowed on the assignments and in person discussions are encouraged. While team work is not mandatory to allow flexibility, forming teams of 2 to 3 students to learn together and collaborate on  joint-assignments is strongly encouraged. This sharing is to improve your understanding of the material. Plagiarising from online resources or other teams or from from previous years  is prohibited.
(2) The tests will be conducted in-class and every registered student MUST take the tests in person. The tests will be closed notes and closed book. You will not be permitted to use any electronic devices, including calculators, laptops, and phones, during the test. Each test will provide helpful information and code snippets from the text book or course notes that are deemed necessary for answering the questions. The format of the test will be communicated to you before the test and they will not be comprehensive.
(3) The entire class is expected to take the test simulaneously though multiple versions of the questions (of comparable coverage and difficulty) will be administered. 
(4) Any academic impropriety during the test  (which includes copying from or communicating with other students, or accessing web resources) will have a minimum penalty of 'F' grade, plus additional disciplinary action for unethical behavior. See http://www.wright.edu/students/judicial/integrity.html for details. 

(5) While we expect constant exchanges on Course-Related Discord Channel and use of emails for personal matters, the course instructor and/or GTA will be available to the students on WebEx or in person for additional meetings as needed, through appointments.  The students will be notified in advance of any planned brief absences of the instructor due to conference travel, or research panel meetings, or for personal reasons, and every effort will be made to have the GTA fill-in to minimize disruption.

Tentatively,  the weightage for the  (earlier and shorter) Test 1will be  40%, and the Test 2 will be 60% for determining the final grade. If this policy is modified for any reason, you will be informed about it in the class.

Grading

Tentatively,  the programming assignments will not be scored. The weightage for  the first test (test1) will be 40% and the second test (Test2) will be 60% for determining the final grade. The exams are not comprehensive and the first test will cover less material compared to the second test. If this policy is modified for any reason, you will be informed. The letter grades will be assigned using the following scale: A[90-100], B[80-90), C[70-80), D[60-70), and F[0-60). However, we reserve the right to adjust the scale somewhat to utilize the gaps in the distribution. If  the class size is large, the grade cut-offs will be guided by the mean and the standard deviation of the cumulative scores. (A separate email will clarify the grading scheme.)

"Attendance" Policy

All registered students are strongly encouraged to attend the class in person. If a student is unable to attend a class, then the student is expected to catch-up by accessing lecture recording(s) from Pilot (available both in Panopto and in WebEx) and using slide decks from the publically available course webpage (https://cecs.wright.edu/people/faculty/tkprasad/courses/cs7800/cs7800.html) . 

Applicable University Policy on Generative AI

The use of generative AI tools (such as ChatGPT and GitHub Co-Pilot) for solving assignments is discouraged to improve understanding by doing. Note that, off and on, the answers provided by these tools or output generated by them can be invalid (due to hallucination) and you will not be able to detect it without first knowing what the "correct" answer is, which is what this course is trying to provide.

Required Texts

Recommended Texts

Reference URLs

 Tentative Class Schedule and Syllabus

  Topics Addl. Reading
Class 1 Information Retrieval; The Boolean Model MIR-1
Class 2 The Vector Space Model : Term Weighting and Scoring IIR-6, MIR-2 
Class 3 Inverted Index Construction IIR-1, MIR-8.2
Class 4 Dictionary and Postings; Query Processing IIR-2, MIR-7.2
Class 5 (continue)
Class 6 Tolerant Retrieval (B-Trees) IIR-3
Class 7 Index Construction IIR-4, MG-5
Class 8
 Map Reduce Architecture Hadoop
Class 9

(continue)
Class 10 Index Compression IIR-5, MG 3.3-4
Class 11
 Vector Space Model : TF-IDF IIR-6.2-4
Class 12 Vector Space Model : Ranking Revisited IIR-6.1, IIR-7
Class 13

Test 1: February 20, Friday, 11:15am-12:10pm,

Loc: 145 RC

  
Class 14 Evaluation in Information Retrieval IIR-8, MIR-3 IIR-8
Class 15 Relevance Feedback and Query Expansion IIR-9, MIR-5.2-4
Class 16 Text Classification and Naive Bayes IIR-13
Class 17 Vector Space Classification IIR-14
Class 18
 Support Vector Machines IIR-15, SVM
Class 19
 Flat and Hierarchical Clustering

IIR-16, IIR-17
Class 20 Latent Semantic Indexing

IIR-18, Refs
Class 21 Linear Algebra: Matrix Decompositions SVD-URL
Class 22 Link Analysis  IIR-21
Class 23
 (continue)
Class 24  Word-Embedding

 
Class * Web Characteristics

IIR-19
Class * Web Search: Crawling and Indexes

IIR-20
 

Test 2 / Meeting: May 1, Friday, 10:15am-12:15pm,

Loc: 145 RC

  

Assignments (Spring 2021)

Assignments (Spring 2015)

Exams (Fall 2012)


Lecture slide decks have been linked to this page. Lecture videos recorded in WebEx will be posted on Pilot.
Furthermore, old recordings in Panopto and/or WebEx will also be made available as a fall back in case there is
some recording snafu.
T. K. Prasad