recap_al.txt

   1 Introduction Model Learning
   2 ===============================================================================
   3 Model learning black box
   4 - Passive: Process mining
   5
   6 - Active: Automata learning
   7         - Teacher and learner
   8         - Learner can ask
   9                 - membership queries: yes/no from teacher
  10                 - equivalence queries: yes/no+counterexample from teacher
  11         - Can be used for fingerprinting
  12
  13 Active Automata Learning
  14 ===============================================================================
  15 - Mealy machine
  16 - Nerode:
  17         two words are equivalent if for all continuations they map the same output
  18 Observation table
  19 Columns: suffix closed
  20 Rows: prefix closed
  21
  22
  23 HYPOTHESIS
  24 - Two parts:
  25         - top part: states
  26         - bottom part: extensions
  27 - Closedness:
  28         Bottom part structure should also be present in upper part otherwise move
  29         up.
  30 - Completeness:
  31         Drawing should be input complete, otherwise add rows with extensions
  32         This means that a top part prefix plus suffix should be available in the
  33         bottom part
  34
  35 EQUIVALENCE QUERY
  36 Yes: successfully found model
  37 No: Counterexample:
  38         word w should give b
  39
  40 HANDLE CE
  41 - extract suffix that disproves nerode
  42         - there is a suffix that ends you up in the same state, the later part is
  43           the culprit.
  44         - add prefix above line, add suffix in column
  45 - add to column(note that columns should be suffix closed)
  46 - fill in table
  47 - refine hypothesis