1. Introduction to Information Retrieval Introduction to Information Retrieval Information Retrieval and Web Search Lecture 1: Introduction and Boolean retrieval Fall 2012

2. Introduction to Information Retrieval Outline ❶ Course details ❷ Information retrieval ❸ Boolean retrieval 2

3. Introduction to Information Retrieval Course details 3  Course weblog: IR-qom.blogfa.comIR-qom.blogfa.com  Useful information from previous terms.  Please check the weblog periodically.  Useful URL: cs276.stanford.educs276.stanford.edu  [a.k.a., http://www.stanford.edu/class/cs276/ ]http://www.stanford.edu/class/cs276/  Slides:  http://nlp.stanford.edu/IR-book/newslides.html http://nlp.stanford.edu/IR-book/newslides.html  Edited versions are placed in weblog.  Please print and bring the slides.

4. Introduction to Information Retrieval Course details 4  Why teaching by slides?  Why English?

5. Introduction to Information Retrieval Course details  Textbook:  Introduction to Information Retrieval  Online (http://informationretrieval.org/)http://informationretrieval.org/  And others ( http://nlp.stanford.edu/IR-book/information-retrieval.html ) http://nlp.stanford.edu/IR-book/information-retrieval.html  Translated books?  Work/Grading (approximately):  Exercises 15%  Project15%  Midterm exam35%  Final exam40%  5% leniency!  -1% each absence!  -0.5% coming after your name is read by teacher! 5

6. Introduction to Information Retrieval Syllabus  Chapter 1: Introduction and boolean retrieval  Chapter 2: The term vocabulary and postings lists  Chapter 3: Dictionaries and tolerant retrieval  Chapter 4: Index Construction 6/48

7. Introduction to Information Retrieval Syllabus  Chapter 5: Index Compression  Chapter 6: Scoring, Term Weighting and the Vector Space Model  Chapter 7: Scoring and results assembly  Chapter 8: Evaluation 7/48

8. Introduction to Information Retrieval Outline ❶ Course details ❷ Information retrieval ❸ Boolean retrieval 8

9. Introduction to Information Retrieval Data 9  Structured data  Example: databases  Unstructured data  Example: free-form texts  Semi-structured data  Example: these slides  In fact almost no data is “unstructured”  Which are related to information retrieval?

10. Introduction to Information Retrieval Structured data  Structured data tends to refer to information in “tables” 10 EmployeeManagerSalary SmithJones50000 ChangSmith60000 50000IvySmith

11. Introduction to Information Retrieval Search  Structured  Typically allows numerical range and exact match (for text) queries, e.g., Salary < 60000 AND Manager = Smith.  Unstructured  Keyword queries including operators  More sophisticated “concept” queries, e.g.,  find all web pages dealing with drug abuse  Semi-structured  “semi-structured” search such as Title contains data AND Bullets contain search 11

12. Introduction to Information Retrieval More sophisticated semi-structured search  Title is about Object Oriented Programming AND Author something like stro*rup  where * is the wild-card operator  Issues:  how do you process “about”?  how do you rank results?  The focus of XML search (IIR chapter 10) 12

13. Introduction to Information Retrieval Unstructured (text) vs. structured (database) data in 1996 13

14. Introduction to Information Retrieval Unstructured (text) vs. structured (database) data in 2009 14

15. Introduction to Information Retrieval Information Retrieval  Information Retrieval (IR) is finding material (usually documents) of an unstructured nature (usually text) that satisfies an information need from within large collections (usually stored on computers).  Example: Find web pages that contains some information about the university of Qom. 15

16. Introduction to Information Retrieval Two aspects of IR systems  Indexing  Search  In which, time is more important?  In which, space is more important? 16/48

17. Introduction to Information Retrieval Clustering and classification  Clustering: Given a set of docs, group them into clusters based on their contents.  Classification: Given a set of topics, plus a new doc D, decide which topic(s) D belongs to. 17

18. Introduction to Information Retrieval Ranking  Ranking: Can we learn how to best order a set of documents, e.g., a set of search results 18

19. Introduction to Information Retrieval Basic assumptions of Information Retrieval  Collection: Fixed set of documents  Goal: Retrieve documents with information that is relevant to the user’s information need and helps the user complete a task 19 Sec. 1.1

20. Introduction to Information Retrieval The classic search model Corpus TASK Info Need Query Verbal form Results SEARCH ENGINE Query Refinement Get rid of mice in a politically correct way Info about removing mice without killing them How do I trap mice alive? mouse trap Misconception?Mistranslation?Misformulation?

21. Introduction to Information Retrieval How good are the retrieved docs?  Two types of error:  false positive  false negative.  Two evaluation measures:  Precision : Fraction of retrieved docs that are relevant to user’s information need  FP is the counterpoint of precision.  Recall : Fraction of relevant docs in collection that are retrieved  FN is the counterpoint of recall.  More precise definitions and measurements to follow in later lectures 21 Sec. 1.1

22. Introduction to Information Retrieval Outline ❶ Course details ❷ Information retrieval ❸ Boolean retrieval 22

23. Introduction to Information Retrieval 23 Boolean retrieval  The Boolean model is perhaps the simplest model to base an information retrieval system on.  Queries are Boolean expressions,  e.g., C AESAR AND B RUTUS  The search engine returns all documents that satisfy the Boolean expression.

24. Introduction to Information Retrieval Term-document incidence 1 if play contains word, 0 otherwise Sec. 1.1 indexing First we collect keywords from each document to avoid searching over the whole document: some kind of indexing

25. Introduction to Information Retrieval Incidence vectors  So we have a 0/1 vector for each term.  To answer query: take the vectors for Brutus, Caesar and Calpurnia (complemented)  bitwise AND.  110100 AND 110111 AND 101111 = 100100. 25 Sec. 1.1 Brutus AND Caesar BUT NOT Calpurnia

26. Introduction to Information Retrieval What is wrong?  Consider N = 1 million documents, each with about 1000 words.  Avg 6 bytes/word including spaces/punctuation  6GB of data in the documents.  Say there are M = 500K distinct terms among these. 26 Sec. 1.1

27. Introduction to Information Retrieval Can’t build the matrix  500K x 1M matrix has half-a-trillion 0’s and 1’s.  But it has no more than one billion 1’s.  matrix is extremely sparse.  What’s a better representation?  We only record the 1 positions. 27 Why? Sec. 1.1

28. Introduction to Information Retrieval Inverted index  For each term t, we must store a list of all documents that contain t.  Identify each by a docID, a document serial number 28 2 Brutus Calpurnia Caesar 124561657132 124113145173 31 What happens if the word Caesar is added to document 14? Sec. 1.2 174 54101 Dictionary Postings Sorted by docID (more later on why).

29. Introduction to Information Retrieval Can we use fixed-size arrays for postings?  We need variable-size postings lists  On disk, a continuous run of postings is normal and best  In memory, can use linked lists or variable length arrays  Linked lists generally preferred to arrays  Dynamic space allocation  Insertion of terms into documents easy  Space overhead of pointers  Some tradeoffs in size/ease of insertion Sec. 1.2

30. Introduction to Information Retrieval Tokenizer Token stream Friends RomansCountrymen Inverted index construction Linguistic modules Modified tokens friend romancountryman Indexer Inverted index friend roman countryman 24 2 13 16 1 More on these later. Documents to be indexed Friends, Romans, countrymen. Sec. 1.2

31. Introduction to Information Retrieval Indexer steps: Token sequence  Sequence of (Modified token, Document ID) pairs. I did enact Julius Caesar I was killed i' the Capitol; Brutus killed me. Doc 1 So let it be with Caesar. The noble Brutus hath told you Caesar was ambitious Doc 2 Sec. 1.2

32. Introduction to Information Retrieval Indexer steps: Sort  Sort by terms  And then docID Core indexing step Sec. 1.2

33. Introduction to Information Retrieval Indexer steps: Dictionary & Postings  Multiple term entries in a single document are merged.  Split into Dictionary and Postings  Doc. frequency information is added. Why frequency? Will discuss later. Sec. 1.2

34. Introduction to Information Retrieval Where do we pay in storage? 34 Pointers Terms and counts Chapter 4: How do we index efficiently? How much storage do we need? Sec. 1.2 Lists of docIDs

35. Introduction to Information Retrieval The index we just built  How do we process a query? 35 Sec. 1.3

36. Introduction to Information Retrieval Query processing: AND  Consider processing the query: Brutus AND Caesar  Locate Brutus in the Dictionary;  Retrieve its postings.  Locate Caesar in the Dictionary;  Retrieve its postings.  “Merge” the two postings: 36 128 34 248163264123581321 Brutus Caesar Sec. 1.3

37. Introduction to Information Retrieval The merge  Walk through the two postings simultaneously, in time linear in the total number of postings entries 37 34 12824816 3264 12 3 581321 128 34 248163264123581321 Brutus Caesar 2 8 If list lengths are x and y, merge takes O(x+y) operations. Crucial: postings sorted by docID. Sec. 1.3

38. Introduction to Information Retrieval Intersecting two postings lists (a “merge” algorithm) 38

39. Introduction to Information Retrieval Boolean queries: More general merges  Exercise: Adapt the merge for the queries: Brutus AND NOT Caesar Brutus OR NOT Caesar  Can we still run through the merge in time O(x+y)?  What can we achieve? 39 Sec. 1.3

40. Introduction to Information Retrieval Merging  Exercise: What about an arbitrary Boolean formula? (Brutus OR Caesar) AND NOT (Antony OR Cleopatra)  Exercise : Extend the merge to an arbitrary Boolean query.  Can we always merge in “linear” time?  Linear in what?  Hint: Begin with the case of a Boolean formula query where each term appears only once in the query. 40 Sec. 1.3

41. Introduction to Information Retrieval Query optimization  Consider a query that is an AND of n terms.  What is the best order for query processing? Brutus Caesar Calpurnia 12358162134 248163264128 1316 Query: Brutus AND Calpurnia AND Caesar 41 Sec. 1.3

42. Introduction to Information Retrieval Query optimization example  Process in order of increasing freq:  start with smallest set, then keep cutting further. 42 This is why we kept document freq. in dictionary Execute the query as (Calpurnia AND Brutus) AND Caesar. Sec. 1.3 Brutus Caesar Calpurnia 12358162134 248163264128 1316

43. Introduction to Information Retrieval More general optimization  e.g., (madding OR crowd) AND (ignoble OR strife)  Get doc. freq.’s for all terms.  Estimate the size of each OR by the sum of its doc. freq.’s (conservative).  Process in increasing order of OR sizes. 43 Sec. 1.3

44. Introduction to Information Retrieval Example  Recommend a query processing order 44 (tangerine OR trees) AND (marmalade OR skies) AND (kaleidoscope OR eyes)

45. Introduction to Information Retrieval Query processing exercises  Exercise: If the query is friends AND romans AND (NOT countrymen), how could we use the freq of countrymen? 45

46. Introduction to Information Retrieval Boolean queries: Exact match  The Boolean retrieval model is being able to ask a query that is a Boolean expression:  Boolean Queries use AND, OR and NOT to join query terms  Views each document as a set of words  Is precise: document matches condition or not.  Perhaps the simplest model to build an IR system on  Primary commercial retrieval tool for 3 decades. 46 Sec. 1.3

47. Introduction to Information Retrieval Boolean queries: Exact match  Many search systems you still use are Boolean:  Email, library catalog, Mac OS X Spotlight  Many professional searchers still like Boolean search  You know exactly what you are getting  But that doesn’t mean it actually works better…. 47

48. Introduction to Information Retrieval What’s ahead in IR? Beyond term search  What about phrases? “Qom University”  Proximity: Find Gates NEAR Microsoft.  Need index to capture position information in docs.  Zones in documents: Find documents with (author = Ullman) AND (text contains automata). 48

49. Introduction to Information Retrieval Ranking search results  Boolean queries give inclusion or exclusion of docs.  Often we want to rank/group results  Need to measure proximity from query to each doc.  Need to decide whether docs presented to user are singletons, or a group of docs covering various aspects of the query. 49

50. Introduction to Information Retrieval The web and its challenges  Unusual and diverse documents  Unusual and diverse users, queries, information needs  Beyond terms, exploit ideas from social networks  link analysis, clickstreams...  How do search engines work? And how can we make them better? 50

51. Introduction to Information Retrieval More sophisticated information retrieval  Cross-language information retrieval  Question answering  Summarization  Text mining  … 51