Artificial Intelligence Meets Blockchain

Artificial Intelligence Meets Blockchain

A Gentle Introduction To A Synergy

TUNDE ADEGBOLA EXECUTIVE DIRECTOR AFRICAN LANGUAGES TECHNOLOGY INITIATIVE

IBADAN, NIGERIA taintransit@

What is Intelligence?

A rather bewildering question

Howard Gardner Vs Jordan Peterson

But we can easily identify Intelligent behaviour

Information is required for making decisions and taking actions ... but, how do we know? Empiricism vs Rationalism

Encounter and recall (what is the answer to 2x2?)

Achieved with Databases and Look-up tables (no intelligence required)

Calculation (what is the answer to 573x758?)

Achieved in Number Crunching, Functions and Algorithms (no intelligence required)

Intelligence addresses the problem of obtaining information (intelligently) Requires more than the traditional strengths of computers (speed and accuracy)

Databases, Look-up Tables and Number Crunching through Functions and Algorithms

Beyond Database Management and Calculation: Reasoning

A database manages already available data and information

Ade is Abu's father, Chike is Ade's brother (database)

... but, who is Abu's uncle?

Certainly not a question for a database management system Knowledge (relationships between data and information units) required

Reasoning is demanded

A need for rules (knowledgebase) and logic (inference engine) An uncle is a parent's brother, a father is a parent (knowledgebase) Manipulating symbols rather than numbers in an inference engine Still relying on Recall and Calculation though

Data, Information and now Knowledge (Understanding)

Human Cognition

AI is an attempt to replicate human intelligence in machines AI mimics humans problem solving ("in a more human-like fashion")

As against brute force "memorization" and number crunching

Humans do not see the world as a collection of objects (OOP?)

Rather, we engage the world as a set of concepts

A concept is a mental representation of a class of like objects Concepts are linked by propositions

Rules by which generalisations can be realised An uncle is a parent's brother (uncle, parent and brother are concepts) Propositions (statements that) evaluate to True or False Predicates are parameterised propositions (i.e with variables)

Humans learn to categorise objects from childhood

AI seeks to make computers do so too

Classification, Clustering And Inferencing

Prediction (a core objective of science and technology)

Functions facilitate interpolation and extrapolation (unknown from known) But many phenomena offer no known functions

How do you distinguish between a man and a woman, good and bad loan application, authentic or scam mails

AI seeks to widen the range of useful things machines can do Making machines learn from history (historical data and information) Machines that can glean information directly from data Assessing values based on mixed-scale features All in a quest for informed decisions and actions

 ................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download