UNIT 5

Analytical Learning-1- Introduction, learning with perfect domain theories:

PROLOG-EBG, remarks on explanation-based learning, explanation-based
learning of search control knowledge.
Analytical Learning-2-Using prior knowledge to alter the search objective,
using prior knowledge to augment search operators.
Combining Inductive and Analytical Learning – Motivation, inductive-
analytical approaches to learning, using prior knowledge to initialize the
hypothesis.
Analytical Learning
Inductive Learning – Based on Patterns
 Any Missing data it assumes and does Classification
Example: Decision Tree Learning, Neural Networks etc.
- Only Training Example is given for learning.
Analytical Learning
Uses prior knowledge and deductive reasoning
prior knowledge – Past data/Past Experience
deductive reasoning – Reasoning Based on facts
Analytical Learning is also called Explanation based learning (because along with
training example, we also give explanation of that example)
Also used where there is missing data, Duplicate data and Error data.
More Efficient than Inductive Learning because in AL. We have Explanation.
Eg. Chess Game
Intuition about Explanation-Based Learning I
Intuition about Explanation-Based Learning II
Analytical Learning: A Definition
 Consider an instance space X in which each instance is a pair of physical
objects.
 Each of the two physical objects in the instance is described by the predicates

Color, Volume, Owner, Material, Type, and Density, and the relationship
between the two objects is described by the predicate On.
 Given this instance space, the task is to learn the target concept "pairs of

physical objects, such that one can be stacked safely on the other," denoted by
the predicate SafeToStack(x,y).
 Learning this target concept might be useful, for example, to a robot system that

has the task of storing various physical objects within a limited workspace.
 The domain theory shown in the table includes assertions such as "it is safe to
stack x on y if y is not Fragile," and "an object x is Fragile if the Material
from which x is made is Glass."
 Like the learned hypothesis, the domain theory is described by a collection of

Horn clauses, enabling the system in principle to incorporate any learned

hypotheses into subsequent domain theories.
 Notice that the domain theory refers to additional predicates such as Lighter

and Fragile, which are not present in the descriptions of the training
examples, but which can be inferred from more primitive instance attributes
such as Material, Density, and Volume, using other other rules in the domain
theory.
 Finally, notice that the domain theory shown in the table is sufficient to prove

that the positive example shown there satisfies the target concept
SafeToStack.
An analytical learning problem: SafeToStack(x,y)
Learning with Perfect Domain Theories: Prolog-EBG
 Analytical Learning is explanation based learning -> related to domain theory
(Knowledge on specific field). E.g. Mathematics, Science Domain.
 Domain Theory is always said to be correct and complete.
 Correct : If each assertion made by Domain Theory(DT) is always True.
 Complete : If it Covers each and every positive example.
 Need of Domain Theory

1. Improved Performance
2. Difficult to achieve a perfect Domain
Example: Prolog-EBG (Programming with Logic)
Mainly based on Sequential Learning and horn Classes
3 Steps : 1. Explaining 2.Analysis 3.Refining
Learning with Perfect Domain Theories: Prolog-EBG

(1)explaining the training example in terms of the domain theory,

(2)analyzing this explanation to determine the relevant features of the example, then
(3) constructing a new Horn clause that concludes the target concept when this set of
features is satisfied.
ANALYZE THE EXPLANATION
REMARKS ON EXPLANATION-BASED LEARNING
 Discovering New Features
 Deductive Learning
 Inductive Bias in Explanation-Based Learning
 Knowledge Level Learning