Object level Knowledge
Lecture by Sohaib Saqib
• Object-level knowledge, also known as
entity-level knowledge, refers to information
and understanding about specific individual
objects.
• This type of knowledge is essential for various
tasks in artificial intelligence, including natural
language understanding, computer vision,
robotics, and knowledge representation.
Identification
• Object-level knowledge includes the ability to
identify and recognize individual objects
within a given context.
• This recognition can involve distinguishing
between different objects based on their
attributes, such as shape, colour, size, or other
distinguishing features.
Attributes
• It involves knowledge about the attributes or
properties of objects, which can be both
intrinsic and extrinsic.
• Intrinsic attributes may include an object's
colour, shape, texture, and weight,
• while extrinsic attributes could involve an
object's location, orientation, or state.
Classification
• Object-level knowledge enables the
classification of objects into categories or
classes.
• This can be based on shared characteristics or
relationships with other objects.
• For example, recognizing that a specific animal
is a cat and not a dog.
Relationships
• Understanding the relationships between
objects is a crucial aspect of object-level
knowledge.
• This includes knowledge about spatial
relationships (e.g., "the book is on the table"),
part-whole relationships (e.g., "the car has
four wheels"), and functional relationships
(e.g., "a key is used to unlock a door").
Contextual Knowledge
• Object-level knowledge is often
context-dependent.
• It involves understanding how an object
behaves or is used within a specific context.
• For instance, knowing that an umbrella is used
to protect against rain or that a knife is used
for cutting.