Catching the Drift: Learning Broad Matches from Clickthrough Data Sonal Gupta, Mikhail Bilenko, Matthew Richardson University of Texas at Austin, Microsoft.

Catching the Drift: Learning Broad Matches from Clickthrough Data

Sonal Gupta, Mikhail Bilenko, Matthew Richardson

University of Texas at Austin, Microsoft Research

Introduction Keyword-based online advertising: bidded keywords are extracted from

context Context: query (search ads) or page (content ads)

Broad matching: expanding keywords via keyword-to-keywords mapping Example: electric cars tesla, hybrids, toyota prius, golf carts

Broad matching benefits advertisers (increased reach, less campaign tuning), users (more relevant ads), ad platform (higher monetization)

Expanded Keywords

kw1

kw11

kw12

kwn

kwn1

kwn2

BroadMatch

Expansion

AdSelection

andRanking

Ad1

Ad2

Adk

Extracted Keywords

Keyword Extraction

kw1

kw2

kwn

Queryor

Web Page

Selected Ads

Identifying Broad Matches

Good keyword mappings retrieve relevant ads that users click

How to measure what is relevant and likely to be clicked? Human judgments: expensive, hard to scale Past user clicks: provide click data for kw → kw’ when user was

shown ad(kw' ) in context of kw Highly available, less trustworthy

What similarity functions may indicate relevance of kw → kw' ? Syntactic (edit distance, TF-IDF cosine, string kernels, …) Co-occurrence (in documents, query sessions, bid campaigns,

…) Expanded representation (search result snippets, category bags,

…)

Approach

Task: train a learner to estimate p(click | kw → kw' ) for any kw → kw'

Data <kw, ad(kw' ), click> triples from clickthrough logs, where kw →

kw' was suggested by previous broad match mappings

Features Convert each pair to a feature vector capturing similarities etc.

(kw → kw') →

For each triple <kw, ad(kw'), click>, create an instance: (ϕ(kw, kw' ), click)

Learner: max-margin averaged perceptron (strong theory, very efficient)

ϕ1(kw, kw' )

ϕn(kw, kw' )

… where ϕi(kw, kw' ) can be any function of kw, kw' or both

5

Example: Creating an Instance Historical broad match clickthrough data: kw kw' ad(kw' ) click event

digital slr canon rebel Canon Rebel Kit for $499 clickseattle baseball mariners tickets Mariners season tickets no click

Feature functions

Instances[0.78 0.001 0.9], 1[0.05 0.02 0.2], 0

Original kw Broad match kw'

ϕ1 ϕ2 ϕ3

digital slr canon rebel 0.78 0.001 0.9

seattle baseball

mariners tickets 0.05 0.02 0.2

Experiments

Data 2 months of previous broad match ads from Microsoft Content Ads

logs 1 month for training, 1 month for testing

68 features (syntactic, co-occurrence based, etc.); greedy feature selection

Metrics LogLoss:

LogLoss Lift: difference between obtained LogLoss and an oracle that has access to empirical p(click | kw → kw' ) in test set.

CTR and revenue results in live test with users

Results

0.6033

0.6572

0.57

0.58

0.59

0.6

0.61

0.62

0.63

0.64

0.65

0.66

0.67

Prior FeatureSelection+Online

-LogLoss

0.0685

0.1224

0

0.02

0.04

0.06

0.08

0.1

0.12

0.14

Prior FeatureSelection+Online

LogL Lift

Live Test Results

Use CTR prediction to maximize expected revenue Re-rank mappings to incorporate revenue +18% revenue, -2% CTR

Revenue

0.0%

20.0%

40.0%

60.0%

80.0%

100.0%

120.0%

BM1 BM2 BM6 Our

CTR

0.0%

20.0%

40.0%

60.0%

80.0%

100.0%

120.0%

BM1 BM2 BM6 Our

Online Learning with Amnesia Advertisers, campaigns, bidded keywords and delivery

contexts change very rapidly: high concept drift

Recent data is more informative Goal: utilize older data while capturing changes in

distributions Averaged Perceptron doesn’t capture drift

Solution: Amnesiac Averaged Perceptron Exponential weight decay when averaging hypotheseswavg

t = α (wt + (1−α)wt−1 + (1−α)2wt−2 + ...)

Results

Model -LogLoss LogL Lift

Prior 0.6572 0.1224

Feature Selection + Online Learning + Amnesia

0.5709 0.0361

Online+Feature Selection, No Amnesia 0.6033 0.0685

Online+Amnesia, No Feature Selection 0.6563 0.1215

Feature Selection+Amnesia, Weekly Batch

0.5948 0.0600

Contributions and Conclusions learning broad matches from implicit feedback

Combining arbitrary similarity measures/features

Using clickthrough logs as implicit feedback

Amnesiac Averaged Perceptron Exponentially weighted averaging: distant examples “fade

out” Online learning adapts to market dynamics

Thank You!

13

Features and Feature Selection Co-occurrence feature examples:

User search sessions: keywords searched within 10 mins Advertiser campaigns: keywords co-bidded by the same advertiser

Past clickthrough rates of original and broad matched keywords Various syntactic similarities Various existing broad matching lists and so on…

Feature Selection: A total of 68 features Greedy feature selection

Additional Information Estimation of expected value of click over all the ads shown

for a broad match mapping E(p(click(ad(kw))|q))

Query Expansion vs. Broad Matching Our broad matching algorithm can be extended for query

expansion But, broad matching is for a fixed set of bidded keywords

Forgetron vs. Amesiac Averaged Perceptron Forgetron maintains a set of budget support vectors: stores

examples explicitly and does not take into account all the data

AAP: weighted average over all the examples, no need to store examples explicitly

Results

Model -LogLoss LogL Lift

Prior 0.6572 0.1224

Feature Selection + Online Learning + Amnesia

0.5709 0.0361

Online+Amnesia, No Feature Selection 0.6563 0.1215

Feature Selection+Amnesia, Weekly Batch

0.5948 0.0600

Online+Feature Selection, No Amnesia 0.6033 0.0685

16

Amnesiac Averaged Perceptron