15-381 Artificial Intelligence

15-381 Artificial IntelligenceInformation Retrieval

(How to Power a Search Engine)

Jaime Carbonell20 September 2001

Topics Covered:• “Bag of Words” Hypothesis• Vector Space Model & Cosine Similarity• Query Expansion Methods

Information Retrieval: The Challenge (1)

Text DB includes:(1) Rainfall measurements in the Sahara continue to show a steadydecline starting from the first measurements in 1961. In 1996 only12mm of rain were recorded in upper Sudan, and 1mm in SouthernAlgiers...

(2) Dan Marino states that professional football risks loosing the numberone position in heart of fans across this land. Declines in TV audienceratings are cited...

(3) Alarming reductions in precipitation in desert regions are blamed fordesert encroachment of previously fertile farmland in Northern Africa.Scientists measured both yearly precipitation and groundwater levels...

Information Retrieval: The Challenge (2)

User query states:"Decline in rainfall and impact on farms near Sahara"

Challenges•How to retrieve (1) and (3) and not (2)?•How to rank (3) as best?•How to cope with no shared words?

Information Retrieval Assumptions (1)

Basic IR task•There exists a document collection {Dj }

•Users enters at hoc query Q

•Q correctly states user’s interest

•User wants {Di } < {Dj } most relevant to Q

"Shared Bag of Words" assumptionEvery query = {wi }Every document = {wk }...where wi & wk in same Σ

All syntax is irrelevant (e.g. word order)All document structure is irrelevantAll meta-information is irrelevant(e.g. author, source, genre)=> Words suffice for relevance assessment

Information Retrieval Assumption (2)

Information Retrieval Assumption (3)

Retrieval by shared words

If Q and Dj share some wi , then Relevant(Q, Dj )

If Q and Dj share all wi , then Relevant(Q, Dj )

If Q and Dj share over K% of wi , then Relevant(Q, Dj)

Boolean Queries (1)Industrial use of SilverQ: silverR: "The Count’s silver anniversary..."

"Even the crash of ’87 had a silver lining...""The Lone Ranger lived on in syndication...""Sliver dropped to a new low in London..."...

Q: silver AND photographyR: "Posters of Tonto and the Lone Ranger..."

"The Queen’s Silver Anniversary photos..."...

Boolean Queries (2)

Q: (silver AND (NOT anniversary)AND (NOT lining)AND emulsion)

OR (AgI AND crystalAND photography))

R: "Silver Iodide Crystals in Photography...""The emulsion was worth its weight in

silver..."...

Boolean Queries (3)

Boolean queries are:

a) easy to implement

b) confusing to compose

c) seldom used (except by librarians)

d) prone to low recall

e) all of the above

Beyond the Boolean Boondoggle (1)

Desiderata (1)

•Query must be natural for all users

•Sentence, phrase, or word(s)

•No AND’s, OR’s, NOT’s, ...

•No parentheses (no structure)

•System focus on important words

•Q: I want laser printers now

Beyond the Boolean Boondoggle (2)Desiderata (2)

• Find what I mean, not just what I say Q: cheap car insurance(pAND (pOR

"cheap" [1.0]"inexpensive" [0.9]"discount" [0.5)]

(pOR "car" [1.0]"auto" [0.8]"automobile" [0.9]"vehicle" [0.5])

(pOR "insurance" [1.0]"policy" [0.3]))

The Vector Space Model (1)

Let Σ = [w1, w2, ... wn ]

Let Dj = [c(w1, Dj), c(w2, Dj), ... c(wn, Dj)]

Let Q = [c(w1, Q), c(w2, Q), ... c(wn, Q)]

The Vector Space Model (2)

Initial Definition of Similarity:

SI(Q, Dj) = Q . Dj

Normalized Definition of Similarity:

SN(Q, Dj) = (Q . Dj)/(|Q| x |Dj|)

= cos(Q, Dj)

The Vector Space Model (3)

Relevance Ranking

If SN(Q, Di) > SN(Q, Dj)

Then Di is more relevant than Di to Q

Retrieve(k,Q,{Dj}) =

Argmaxk[cos(Q, Dj)]

Dj in {Dj}

Refinements to VSM (2)

Stop-Word Elimination

• Discard articles, auxiliaries, prepositions, ... typically 100-300 most frequent small words

• Reduce document length by 30-40%

• Retrieval accuracy improves slightly (5-10%)

Refinements to VSM (3)

Proximity Phrases• E.g.: "air force" => airforce• Found by high-mutual information

p(w1 w2) >> p(w1)p(w2)

p(w1 & w2 in k-window) >>

p(w1 in k-window) p(w2 in same k-window)

• Retrieval accuracy improves slightly (5-10%)• Too many phrases => inefficiency

Refinements to VSM (4)

Words => Terms

• term = word | stemmed word | phrase

• Use exactly the same VSM method on terms (vs words)

Evaluating Information Retrieval (1)

Contingency table:

docsretrievedall

relevantretrieveedprecision

docsrelevantall

relevantretrieveedrecall

&

&

relevant not-relevant

retrieved a b

not retrieved c d

Evaluating Information Retrieval (2)

P = a/(a+b) R = a/(a+c)

Accuracy = (a+d)/(a+b+c+d)

F1 = 2PR/(P+R)

Miss = c/(a+c) = 1 - R

(false negative)

F/A = b/(a+b+c+d)

(false positive)

Query Expansion (1)Observations:• Longer queries often yield better results• User’s vocabulary may differ from document

vocabularyQ: how to avoid heart diseaseD: "Factors in minimizing stroke and cardiac arrest: Recommended dietary and exercise regimens"

• Maybe longer queries have more chances to help recall.

Query Expansion (2)

Bridging the Gap• Human query expansion (user or expert)• Thesaurus-based expansion

Seldom works in practice (unfocused)• Relevance feedback

– Widen a thin bridge over vocabulary gap– Adds words from document space to query

• Pseudo-Relevance feedback• Local Context analysis

Relevance Feedback

Rocchio Formula

Q’ = F[Q, Dret ]

F = weighted vector sum, such as:

W(t,Q’) =

αW(t,Q) + βW(t,Drel ) - γW(t,Dirr )

Term Weighting Methods (1)

Salton’s Tf*IDfTf = term frequency in a document

Df = document frequency of term= # documents in collection

with this term

IDf = Df-1

Term Weighting Methods (2)

Salton’s Tf*IDfTfIDf = f1(Tf)*f2(IDf)

E.g. f1(Tf) =Tf*ave(|Dj|)/|D|

E.g. f2(IDf) = log2(IDF)

f1 and f2 can differ for Q and D

Efficient Implementations of VSM (1)

• Build an Inverted Index (next slide)

• Filter all 0-product terms

• Precompute IDF, per-document TF

• …but remove stopwords first.

Efficient Implementations of VSM (3)[term IDFtermi

,

<doci, freq(term, doci )

docj, freq(term, docj )...>]

or:

[term IDFtermi,

<doci, freq(term, doci), [pos1,i, pos2,i, ...]

docj, freq(term, docj), [pos1,j, pos2,j, ...]...>]

posl,1 indicates the first position of term in documentj and so on.

Generalized Vector Space Model (1)

Principles• Define terms by their occurrence patterns in

documents• Define query terms in the same way• Compute similarity by document-pattern

overlap for terms in D and Q• Use standard Cos similarity and either

binary or TfIDf weights

Generalized Vector Space Model (2)

Advantages• Automatically calculates partial similarity

If "heart disease" and "stroke" and "ventricular" co-occur in many documents, then if the query contains only one of these terms, documents containing the other will receive partial credit proportional to their document co-occurrence ratio.

• No need to do query expansion or relevance feedback

GVSM, How it Works (1)

Represent the collection as vector of documents:

Let C = [D1, D2, ..., Dm ]Represent each term by its distributional frequency:

Let ti = [Tf(ti, D1), Tf(ti, D2 ), ..., Tf(ti, Dm )]Term-to-term similarity is computed as:

Sim(ti, tj) = cos(vec(ti), vec(tj))Hence, highly co-occurring terms like "Arafat" and

"PLO"will be treated as near-synonyms for retrieval

GVSM, How it Works (2)And query-document similarity is computed as

before: Sim(Q,D) = cos(vec(Q)), vec(D)), except that instead of the dot product calculation, we use a function of the term-to-term similarity computation above, For instance:

Sim(Q,D) = Σi[Maxj(sim(qi, dj)]or normalizing for document & query length:

Simnorm(Q, D) = ||||

)],(([

DQ

dqsimMax ji

A Critique of Pure Relevance (1)

IR Maximizes Relevance

• Precision and recall are relevance measures

• Quality of documents retrieved is ignored

A Critique of Pure Relevance (2)

Other Important Factors• What about information novelty, timeliness,

appropriateness, validity, comprehensibility, density, medium,...??

• In IR, we really want to maximize:P(U(f i , ..., f n ) | Q & {C} & U & H)where Q = query, {C} = collection set,U = user profile, H = interaction history

• ...but we don’t yet know how. Darn.

Maximal Marginal Relevance (1)

• A crude first approximation:

novelty => minimal-redundancy

• Weighted linear combination:

(redundancy = cost, relevance = benefit)

• Free parameters: k and λ

Maximal Marginal Relevance (2)

MMR(Q, C, R) =

Argmaxkdi in C[λS(Q, di) - (1-λ)maxdj

in R (S(di, dj))]

Maximal Marginal Relevance (MMR) (3)COMPUTATION OF MMR RERANKING1. Standard IR Retrieval of top-N docs

Let Dr = IR(D, Q, N)

2. Rank max sim(di ε Dr, Q) as top doc, i.e. Let

Ranked = {di}

3. Let Dr = Dr\{di}

4. While Dr is not empty, do:

a. Find di with max MMR(Dr, Q. Ranked)

b. Let Ranked = Ranked.di

c. Let Dr = Dr\{di}

Maximal Marginal Relevance (MMR) (4)

Applications:

• Ranking retrieved documents from IR Engine

• Ranking passages for inclusion in Summaries

Document Summarization in a Nutshell (1)Types of Summaries

Task Query-relevant

(focused)

Query-free

(generic)

INDICATIVE, for Filtering (Do I

read further?)

To filter search engine results

Short abstracts

CONTENTFUL, for reading in lieu

of full doc.

To solve problems for busy

professionals

Executive summaries

Summarization as Passage Retrieval (1)

For Query-Driven Summaries1. Divide document into passages

e.g, sentences, paragraphs, FAQ-pairs, ....2. Use query to retrieve most relevant passages, or better, use MMR to avoid redundancy.3. Assemble retrieved passages into a summary.