Information Retrieval CMSC 476/676

Lecture 8: Evaluation and Result Summaries

This lecture

How do we know if our results are any good?

Evaluating a search engine

Benchmarks Precision and recall

Result Summaries

Having ranked the documents matching a query, we wish to present a results list Most commonly, a list of the document titles plus a short summary, aka 10 blue links

Summaries

The title is typically automatically extracted from document metadata. What about the summaries?

This description is crucial. User can identify good/relevant hits based on description. Static Dynamic

Two basic kinds:

A static summary of a document is always the same, regardless of the query that hit the doc A dynamic summary is a query-dependent attempt to explain why the document was retrieved for the query at hand

Static summaries

In typical systems, the static summary is a subset of the document Simplest heuristic: the first 50 (or so this can be varied) words of the document

Summary cached at indexing time

More sophisticated: extract from each document a set of key sentences

Simple NLP heuristics to score each sentence Summary is made up of top-scoring sentences.

Most sophisticated: NLP used to synthesize a summary

Seldom used in IR; cf. text summarization work

Dynamic summaries

Present one or more windows within the document that contain several of the query terms

KWIC snippets: Keyword in Context presentation If query found as a phrase, all or some occurrences of the phrase in the doc If not, document windows that contain multiple query terms

Generated in conjunction with scoring

The summary itself gives the entire content of the window all terms, not only the query terms how?

Generating dynamic summaries

If we have only a positional index, we cannot (easily)

reconstruct context window surrounding hits If we cache the documents at index time, can find windows in it, cueing from hits found in the positional index

E.g., positional index says the query is a phrase in position 4378 so we go to this position in the cached document and stream out the content Note: Cached copy can be outdated

Most often, cache only a fixed-size prefix of the doc

Dynamic summaries

Producing good dynamic summaries is a tricky optimization problem

The real estate for the summary is normally small and fixed Want short item, so show as many KWIC matches as possible, and perhaps other things like title Want snippets to be long enough to be useful Want linguistically well-formed snippets: users prefer snippets that contain complete phrases Want snippets maximally informative about doc

But users really like snippets, even if they complicate IR system design
8

Alternative results presentations?

An active area of HCI research An alternative: http://www.searchme.com / copies the idea of Apples Cover Flow for search results

Evaluating search engines

Measures for a search engine

How fast does it index

Number of documents/hour (Average document size)

How fast does it search

Latency as a function of index size

Ability to express complex information needs Speed on complex queries

Expressiveness of query language

Uncluttered UI Is it free?
11

Measures for a search engine

All of the preceding criteria are measurable: we can quantify speed/size; we can make expressiveness precise The key measure: user happiness

What is this? Speed of response/size of index are factors But blindingly fast, useless answers wont make a user happy

Need a way of quantifying user happiness

12

Measuring user happiness

Issue: who is the user we are trying to make happy?

Depends on the setting

Web engine: user finds what they want and return to the engine

Can measure rate of return users

eCommerce site: user finds what they want and make a purchase

Is it the end-user, or the eCommerce site, whose happiness we measure? Measure time to purchase, or fraction of searchers who become buyers?

13

Measuring user happiness

Enterprise (company/govt/academic): Care about user productivity

How much time do my users save when looking for information? Many other criteria having to do with breadth of access, secure access, etc.

14

Happiness: elusive to measure

Most common proxy: relevance of search results But how do you measure relevance? We will detail a methodology here, then examine its issues Relevant measurement requires 3 elements:
1. A benchmark document collection 2. A benchmark suite of queries 3. A usually binary assessment of either Relevant or Nonrelevant for each query and each document
Some work on more-than-binary, but not the standard
15

Evaluating an IR system

Note: the information need is translated into a query Relevance is assessed relative to the information need not the query E.g., Information need: I'm looking for information on whether drinking red wine is more effective at reducing your risk of heart attacks than white wine. Query: wine red white heart attack effective You evaluate whether the doc addresses the information need, not whether it has these words 16

Standard relevance benchmarks

TREC - National Institute of Standards and Technology (NIST) has run a large IR test activity for many years Retrieval tasks specified

sometimes as queries

Human experts mark, for each query and for each doc, Relevant or Nonrelevant

But there are way too many documents So only judge that subset of docs that some system returned for that query
17

Precision and Recall

Precision: fraction of retrieved docs that are relevant = P(relevant|retrieved) Recall: fraction of relevant docs that are retrieved = P(retrieved|relevant)
Relevant Retrieved Not Retrieved

Nonrelevant False postive

True positive

False negative True negative

Precision P = tp/(tp + fp) Recall R = tp/(tp + fn)

18

Precision/Recall

You can get 100% recall (but low precision) by retrieving all docs for all queries! Recall is a non-decreasing function of the number of docs retrieved Precision tends to decrease, asymptotically, as either the number of docs retrieved or recall increases

This is not a theorem, but a result with strong empirical confirmation To say that a system is better than before, show that either P or R or both are better
19

Difficulties in using precision/recall

Should average over large document collection/query ensembles Need human relevance assessments

People arent reliable assessors Nuanced assessments?

Assessments have to be binary

Heavily skewed by collection/authorship

Results may not translate from one domain to another

20

A combined measure: F

The F measure (weighted harmonic mean) is a combined measure that assesses the precision/recall tradeoff:

( 1) PR F 2 1 1 PR (1 ) P R 1
2

People usually use balanced F1 measure

i.e., with = 1 or = See van Rijsbergen, Information Retrieval

Harmonic mean is a conservative average

21

Evaluating ranked results

Evaluation of ranked results:

The system can return any number of results By taking various numbers of the top returned documents (levels of recall), the evaluator can produce a precision-recall curve

22

A precision-recall curve
1.0 0.8

Precision

0.6 0.4 0.2 0.0 0.0 0.2 0.4 0.6 0.8 1.0
23

Recall

Averaging over queries

A precision-recall graph for one query isnt a very sensible thing to look at You need to average performance over several queries. But theres a technical issue:

Precision-recall calculations place some points on the graph How do you determine a value (interpolate) between the points?

24

Interpolated precision

Idea: If locally precision increases with increasing recall, then you should get to count that So you max of precisions to right of value

25

Evaluation

Graphs are good, but people want summary measures! Precision at fixed retrieval level

Precision-at-k: Precision of top k results Perhaps appropriate for most of web search: all people want are good matches on the first one or two results pages But: averages badly and has an arbitrary parameter of k The standard measure in the early TRECs: you take the precision at 11 levels of recall varying from 0 to 1 by tenths of the documents, using interpolation (the value for 0 is always interpolated!), and average them Evaluates performance at all recall levels
26

11-point interpolated average precision

Typical (good) 11 point precisions

SabIR/Cornell 8A1 11pt precision from TREC 8 (1999)

1

0.8

Precision

0.6

0.4

0.2

0 0 0.2 0.4 Recall 0.6 0.8 1

27

Yet more evaluation measures

Mean average precision (MAP)

Average of the precision value obtained for the top k documents, each time a relevant doc is retrieved Avoids interpolation, use of fixed recall levels MAP for query collection is arithmetic ave.

Macro-averaging: each query counts equally

R-precision

If have known (though perhaps incomplete) set of relevant documents of size Rel, then calculate precision of top Rel docs returned Perfect system could score 1.0.
28

Variance

For a given test collection, it is common for a system to do poorly on some information needs (e.g., MAP = 0.1) and well on others (e.g., MAP = 0.7) Indeed, it is usually the case that the variance in performance of the same system across queries is much greater than the variance of different systems on the same query. That is, there are easy information needs and hard ones!
29

Creating Test Collections for IR Evaluation

Test Collections

31

From document collections to test collections

Still need

Test queries Relevance assessments

Test queries

Must be germane to docs available Best designed by domain experts Random query terms generally not a good idea
Human judges, time-consuming Are human panels perfect?
32

Relevance assessments

Unit of Evaluation

We can compute precision, recall, F, and ROC curve for different units. Possible units

Documents (most common) Facts (used in some TREC evaluations) Entities (e.g., car companies)

May produce different results. Why?

33

Kappa measure for inter-judge (dis)agreement

Kappa measure

Agreement measure among judges Designed for categorical judgments Corrects for chance agreement

Kappa = [ P(A) P(E) ] / [ 1 P(E) ] P(A) proportion of time judges agree P(E) what agreement would be by chance Kappa = 0 for chance agreement, 1 for total agreement.

34

P(A)? P(E)?

Kappa Measure: Example

Number of docs Judge 1 Judge 2

300

Relevant

Relevant

70

Nonrelevant

Nonrelevant

20

Relevant

Nonrelevant

10

Nonrelevant

relevant
35

Kappa Example

P(A) = 370/400 = 0.925 P(nonrelevant) = (10+20+70+70)/800 = 0.2125 P(relevant) = (10+20+300+300)/800 = 0.7878 P(E) = 0.2125^2 + 0.7878^2 = 0.665 Kappa = (0.925 0.665)/(1-0.665) = 0.776 Kappa > 0.8 = good agreement 0.67 < Kappa < 0.8 -> tentative conclusions (Carletta 96) Depends on purpose of study 36 For >2 judges: average pairwise kappas

TREC

TREC Ad Hoc task from first eight TRECs is standard IR task

50 detailed information needs per year Human evaluation of pooled results returned More recently other related things: Web track, HARD

A TREC query (TREC 5)

<top> <num> Number: 225 <desc> Description: What is the main function of the Federal Emergency Management Agency (FEMA) and the funding level provided to meet emergencies? Also, what resources are available to FEMA such as people, equipment, facilities? 37 </top>

Standard relevance benchmarks: Others

GOV2

Another TREC/NIST collection 25 million web pages Largest collection that is easily available But still 3 orders of magnitude smaller than what Google/Yahoo/MSN index
East Asian language and cross-language information retrieval This evaluation series has concentrated on European languages and cross-language information retrieval.
38

NTCIR

Cross Language Evaluation Forum (CLEF)

Many others

Interjudge Agreement: TREC 3

39

Impact of Inter-judge Agreement

Impact on absolute performance measure can be significant (0.32 vs 0.39) Little impact on ranking of different systems or relative performance Suppose we want to know if algorithm A is better than algorithm B A standard information retrieval experiment will give us a reliable answer to this question.

40

Critique of pure relevance

Relevance vs Marginal Relevance

A document can be redundant even if it is highly relevant Duplicates The same information from different sources Marginal relevance is a better measure of utility for the user.

Using facts/entities as evaluation units more directly measures true relevance. But harder to create evaluation set See Carbonell reference

41

Can we avoid human judgment?

No Makes experimental work hard

Especially on a large scale

In some very specific settings, can use proxies

E.g.: for approximate vector space retrieval, we can compare the cosine distance closeness of the closest docs to those found by an approximate retrieval algorithm

But once we have test collections, we can reuse them (so long as we dont overtrain too badly)
42

Evaluation of large search engines

Search engines have test collections of queries and handranked results Recall is difficult to measure on the web Search engines often use precision at top k, e.g., k = 10 . . . or measures that reward you more for getting rank 1 right than for getting rank 10 right.

NDCG (Normalized Cumulative Discounted Gain) Clickthrough on first result Not very reliable if you look at a single clickthrough but pretty reliable in the aggregate. Studies of user behavior in the lab A/B testing

Search engines also use non-relevance-based measures.

43

A/B testing

Purpose: Test a single innovation Prerequisite: You have a large search engine up and running. Have most users use old system Divert a small proportion of traffic (e.g., 1%) to the new system that includes the innovation Evaluate with an automatic measure like clickthrough on first result Now we can directly see if the innovation does improve user happiness. Probably the evaluation methodology that large search engines trust most In principle less powerful than doing a multivariate regression analysis, but easier to understand 44

Resources for this lecture

IIR 8 MIR Chapter 3 MG 4.5 Carbonell and Goldstein 1998. The use of MMR, diversity-based reranking for reordering documents and producing summaries. SIGIR 21.

45