Mapper
Training
data T
Similarity Functions (F)
Mapped
labeled
instances
Pool of
mapped
unlabeled
instances
Select
instances
Train classifier
Active Learner
Unlabeled
Input
records
D
Initial
training
records
Dp
Predicate for uncertain region
Lp
Mapper
Similarity
Indices
S
Infer  pairs 
using
transitivity
Deduplication function
Large record lists
A
Evaluation engine
Groups of duplicates in A
Architecture of ALIAS