1 CS 430: Information Discovery Lecture 1 Overview of Information Discovery

2 Course Administration Web site: Instructor: William Arms, Information Science, Room 104 Teaching assistants: Pavel Dmitriev, Ariful Gani, Heng-Scheng Chuang Assistant: Anat Nidar-Levi, Information Science, Room 101 Sign-up sheet: Include your NetID Contact the course team: to Notices: See the home page of the course Web site

3 Discussion Classes Format of Wednesday evening classes: Topic announced on Web site with chapter or article to read Allow several hours to prepare for class by reading the materials Class has discussion format One third of grade is class participation Class time is 7:30 to 8:30 in Upson B17 (note change of room)

4 Assignments Four individual programming assignments, in Java or C++.

5 Code of Conduct Computing is a collaborative activity. You are encouraged to work together, but... Some tasks may require individual work. Always give credit to your sources and collaborators. To make use of the expertise of others and to build on previous work, with proper attribution is good professional practice. To use the efforts of others without attribution is unethical and academic cheating. Read and follow the University's Code of Academic Integrity.

6 Course Description This course studies techniques and human factors in discovering information in online information systems. Methods that are covered include techniques for searching, browsing and filtering information, descriptive metadata, the use of classification systems and thesauruses, with examples from Web search systems and digital libraries.

7 Information Discovery People have many reasons to look for information: Known item Where will I find the wording of the US Copyright Act? Facts What is the capital of Barbados? Introduction or overview How do diesel engines work? Related information (annotation) Is there a review of this item? Comprehensive search What is known of the effects of global warming on hurricanes?

8 Information Discovery People have many ways to look for information: Where will I find the wording of the US Copyright Act? Is government information outside copyright? Browse: What is the capital of Barbados? Search: How do diesel engines work? Search: What is known of the effects of global warming on hurricanes? Explore:

9 Searching and Browsing: The Human in the Loop Search index Return hits Browse repository Return objects

10 Definitions Information retrieval: subfield of computer science that deals with automated retrieval of documents based on their content. Searching: seeking for specific information within a body of information. The result of a search is a set of hits. Browsing: unstructured exploration of a body of information. Linking: Moving from one item to another following links, such as citations, references, etc.

11 The Basics of Information Retrieval Query: A string of text, describing the information that the user is seeking. Each word of the query is called a search term. A query can be a single search term, a string of terms, a phrase in natural language, or a stylized expression using special symbols. Full text searching: Methods that compare the query with every word in the text, without distinguishing the function of the various words. Fielded searching: Methods that search on specific bibliographic or structural fields, such as author or heading.

12 Searching: Physical Collections Document collection Index User Create catalog

13

14 Descriptive metadata Some methods of information discovery search descriptive metadata about the objects. Metadata typically consists of a catalog or indexing record, or an abstract, one record for each object. The record acts as a surrogate for the object. Usually the metadata is stored separately from the objects that it describes, but sometimes is embedded in the objects. Usually the metadata is a set of text fields. Textual metadata can be used to describe non-textual objects, e.g., software, images, music

15 Documents and Surrogates The sea is calm to-night. The tide is full, the moon lies fair Upon the straits;--on the French coast the light Gleams and is gone; the cliffs of England stand, Glimmering and vast, out in the tranquil bay. Come to the window, sweet is the night-air! Only, from the long line of spray Where the sea meets the moon-blanch'd land, Listen! you hear the grating roar Of pebbles which the waves draw back, and fling, At their return, up the high strand, Begin, and cease, and then again begin, With tremulous cadence slow, and bring The eternal note of sadness in. Author: Matthew Arnold Title: Dover Beach Genre: Poem Date: 1851 Document Surrogate (catalog record) Notes: 1. The surrogate is also a document 2. Every word is different!

16 Descriptive metadata Catalog: metadata records that have a consistent structure, organized according to systematic rules. (Example: Library of Congress Catalog) Abstract: a free text record that summarizes a longer document. Indexing record: less formal than a catalog record, but more structure than a simple abstract. (Example: Inspec, Medline)

17 Surrogates for non-textual materials Textual catalog record about a non-textual item (photograph) Surrogate Text based methods of information retrieval can search a surrogate for a photograph

18 Library of Congress catalog record (part) CREATED/PUBLISHED: [between 1925 and 1930?] SUMMARY: U. S. President Calvin Coolidge sits at a desk and signs a photograph, probably in Denver, Colorado. A group of unidentified men look on. NOTES: Title supplied by cataloger. Source: Morey Engle. SUBJECTS: Coolidge, Calvin, Presidents--United States Autographing--Colorado--Denver Denver (Colo.) Photographic prints. MEDIUM: 1 photoprint ; 21 x 26 cm. (8 x 10 in.)

19 Searching: Electronic Collections Document collection Index User Automatic indexing

20

21 Automatic indexing Creating catalog records manually is labor intensive and hence expensive. The aim of automatic indexing is to build indexes and retrieve information without human intervention. History Much of the fundamental research in automatic indexing was carried out by Gerald Salton, Professor of Computer Science at Cornell, and his graduate students. The reading for Discussion Class 2 is a paper by Salton and others that describes the SMART system used for their research.

22 Searching: Other Methods Natural language processing (Not part of this class) Language support Lexicon, thesaurus Classification Divide documents into groups based on subject

23 Lexicon and thesaurus Lexicon contains information about words, their morphological variants, and their grammatical usage. Thesaurus relates words by meaning: ship, vessel, sail; craft, navy, marine, fleet, flotilla book, writing, work, volume, tome, tract, codex search, discovery, detection, find, revelation (From Roget's Thesaurus, 1911)

24 Searching: Distributed Collections (e.g., the Web) Document collections Index User Web crawling

25

26 Multi-modal Information Discovery Incorporate information from many sources into indexes Google (Web documents) full text indexing hyperlink structures file names, anchor text, etc. Informedia (video segments) speech recognition from sound track closed captions text over video

27 Measuring Effectiveness: Recall and Precision If information retrieval were perfect... Every hit would be relevant to the original query, and every relevant item in the body of information would be found. Precision: percentage of the hits that are relevant, the extent to which the set of hits retrieved by a query satisfies the requirement that generated the query. Recall: percentage of the relevant items that are found by the query, the extent to which the query found all the items that satisfy the requirement.

28 Recall and Precision: Example Collection of 10,000 documents, 50 on a specific topic Ideal search finds these 50 documents and reject others Actual search identifies 25 documents; 20 are relevant but 5 were on other topics Precision: 20/ 25 = 0.8 Recall: 20/50 = 0.4

29 Measuring Precision and Recall Precision is easy to measure: A knowledgeable person looks at each document that is identified and decides whether it is relevant. In the example, only the 25 documents that are found need to be examined. Recall is difficult to measure: To know all relevant items, a knowledgeable person must go through the entire collection, looking at every object to decide if it fits the criteria. In the example, all 10,000 documents must be examined.

30 History Measuring the effectiveness of information discovery is an unresolved research topic. Much of the work on evaluation of information retrieval derives from the ASLIB Cranfied projects led by Cyril Cleverdon, which began in Recent work on evaluation has centered around the Text Retrieval Conferences (TREC), which began in 1992, led by Donna Harman and Ellen Voorhees of NIST.