ABOUT THIS COURSE

General Overview

Organization	The bulk of the course is built on reading 1-3 papers every day and discussing it. - The first week involves 6 topics; the second week involves two of the topics and two programming assignments. The last week consists of invited talks and hands-on projects.
Implementation Assignments	Implementation Assignments are intended to give you insight into some of the core principles of intelligence.
Exam	The Exam is designed to measure how much of the material you actually have understood – how much of it you have “ingested” – through a written exam of 2 hours. The content of the exam is the readings and the invited talks.
How to get the most out of the course	Pay attention and spend a lot of time reading in the first week. Spend time on thinking about the content ! Ask questions in all sessions !

What this Course Is / Is not

Terminology	*Terms used in this course are usually taken in their most general* sense, unless otherwise noted. Terms in AI are regularly morphed and redefined to mean something different than their typical meaning when encountered in general discourse. One example is “attention” – a term that people use all the time but means something very** different in the context of humans than in the context of ANN-based systems. The reasons for such term re-definitions are numerous. Here it should be noted that, since we are talking about the natural phenomenon of intelligence, when we use terms like “attention”, “reasoning”, “knowledge” and the like, we are most often talking about the intuitive term that typically would be meant if the words were used in general conversations in general discourse (e.g. between family members, friends, etc.).
Intelligence	This course is about a phenomenon we often refer to as “intelligence”. A number of features of natural intelligence remain unexplained. Like the focus of any good scientist should reflect, it is the unexplained and ill-understood aspects of this phenomenon that is our key focus here.
GMI / AGI	A number of terms have been used to refer to the various aspects that people study WRT intelligence. We use the terms “general machine intelligence” (GMI) and “artificial general intelligence” (AGI), in their most general sense (no pun intended), to refer to the various aspects of intelligence that allow an agent to deal with variety, information incompleteness, and incremental information gathering and modeling.
Advanced topics	The main focus of course is not the latest and greatest methods to come out of the field called “AI”. Instead, it is about unanswered scientific questions. Nevertheless, we will make some references to contemporary AI methods along the way (much of which is ANN-based), and you may even learn something about them. What is meant by “next-generation” are the upcoming-yet-unadapted ideas, methods, and approaches that the field of AI is developing now and which will soon make contemporary AI methods mostly irrelevant.
Then what does “next-generation” refer to here?	It refers to advancement toward a deeper understanding of the phenomenon of intelligence, and how to create machines based on these.
History	The phenomenon of intelligence has been studied for ages. Some of the early notable contributions were the Greek philosophers' musings on reasoning and logic. This is not a history course, but we may make some references to the history of philosophy, AI, cybernetics, and computer science along the way.

Important Terms in This Course

Methodology	Contemporary methods in AI (specifically, artificial neural networks, reinforcement learning, and related methods) will not suffice for addressing the full scope of the phenomenon of intelligence, as observed in nature.
Attention	The ability to manage resources, including computational (thought), information from the external environment, energy, and time.
Meta-Cognition	The ability of a system to reason about itself, about its place in the world, and about its own reasoning.
Reasoning	The application of logical rules to knowledge.
Knowledge	Structured information in a way that it can be used for goal-achievement, explanation and prediction.
Information	A collection of data organized in a way to contain relations and some meta-data.
Data	A measurement.
Learning	Acquisition of knowledge - models of experience - that enables more successful (a) completion of tasks, (b) creation of goals, subgoals and plans, © prediction and explanation of the world.
Cumulative Learning	Incremental, continual acquisition of knowledge in such a way as generally making it more useful over time.
Transfer learning	The ability to transfer knowledge learned in one task to a different ask. Also called 'generalization'.
Autonomy	The ability to do tasks without help from a teacher (process designed to specifically help with a specific learning process).
Constructivist AI	Methodology that relies on systems acquiring their own knowledge.

Center for Analysis and Design of Intelligent Agents

Table of Contents

ABOUT THIS COURSE

General Overview

What this Course Is / Is not

Important Terms in This Course