SCSI E-86 Building the Brain

A survey of artificial intelligence and its implications.

 

Course Description (from catalog):

What will our artificial intelligence (AI) overlords look like? The world is abuzz with deep learning, machine learning, data science, and artificial intelligence. There are countless videos with artificial neural networks depicted as simple circles interconnected with lines or tutorials that are long on the mathematical concepts; however, the truth is a bit more complicated. If you want to know enough to write a neural network, you find precious little out there. If you want to model real neurons, neural networks fall short. In this course we explore the true physical neuron found in nature and in our heads, how it works, how it can be modelled in a very limited way, and how the nodes in artificial neural networks are very different from these. We explore why AI is suddenly hot again, after thirty years in hibernation. We learn how to create an artificial neural network and see inside the black box of libraries that do this automagically for us. We explore some of the ethical, political, and socio-economic ramifications and questions highlighted by advances in these fields. There is much work that can still be done to help enhance these networks to better approximate natural networks of neurons, and how it may be possible to meld different versions of networks and more traditional programmatic control (as is being used in autonomous vehicles) to create far more complex emergent systems behavior. This course is not intended to replace other artificial intelligence courses but is intended to be a survey course of sufficient depth to allow you to better understand the technology, its potential benefits and pitfalls, and to be able to speak intelligently about the subject matter.

 

Introduction

Artificial Intelligence, one of a number of Machine Learning paradigms, has seen accelerated advances in the past decade that only stand to get more interesting as time passes. It is therefore important for those who are interested to understand what it is, what is it no (at least not yet), how it works, and what are the opportunities and risks associated with it.

 

This course has the lofty aim of helping you learn the difference between artificial neural networks, natural neural networks (e.g., the human brain), and what those differences may mean. You should also have an intuitive grasp of Neurological and AI concepts, had played around with some AI (Deep Learning) code, understand a bit of the history and trajectory, and be well informed enough to discuss the potential implications of the progress of the field.

 

Notes

Below is a DRAFT Syllabus which is subject to change. There may be guest lecturers in the fields of study pending ability to arrange schedules.

 

Readings

Most readings will be optional and most of the required information will be available within the quizzes. If you would like to view a list of interesting readings from which I may pick things for dicussion see: Interesting Readings.

Other pages:

• ANN: Artificial Neurons
• ANN: Deep Learning Networks
• ANN: Forward and Backpropagation
• ANN: XOR Implementation
• CS: Historical Timeline
• Intelligence
• ML: Machine Learning
• ML: The Perceptron
• NNN: Complexity
• NNN: Encoding & Signal Processing
• Week 2 Part 1: Neuroanatomy: History
• Week 2 Part 1: Neuron Biology Page 1
• Week 2, Part 1: Neurophysiology: The Cells

 

Grades

	Undergrad	Graduate
Participation (surveys, sections, discussions, classroom, etc.)	10%	10%
Quizzes (5 to 30 minutes each)	45%	40%
Midterm	20%	15%
Final	25%	20%
Paper / Project (Short)	-	15%

 

Individuals taking this course for graduate credit will have some additional readings and assignments.

Taking surveys, asking questions, participating in class, pointing out typos, suggesting readings, all count as participation.

DISTANCE LEARNING:

All required items that will result in a grade may be done asynchronously for those who wish to take all or part of this course on-line.

Syllabus

Part1

1.     Intro to the class:

1.1. Learning objectives

1.2. Prerequisites

1.3. Topics

1.4. Expectations

1.4.1.    Homeworks

1.4.2.    Tests

1.4.3.    Grades

1.4.4.    Participation

1.5. Instructors and student Intros

2.     In the News

Part 2

3.     Disambiguation (aka “uh, wut?”}:

3.1. Natural Intelligence

3.1.1.    (Contested) Definitions

3.1.2.    Nature vs. Nurture

3.1.3.    Genetics

3.1.4.    Environment

3.2. Neurons, Nerves, and more

3.3. Brains

3.3.1.    General

3.3.2.    Human

3.4. Machine Learning

3.5. Artificial Intelligence

3.6. Neuroinformatics

Week 2 Part 1: Neuroanatomy: History

Week 2, Part 1: Neurophysiology: The Cells

Week 2 Part 1: Neuron Biology Page 1

(Note: I removed the link to the page with Cell Membrane images because of issues with the files. They are all included in the slides)

4.     Neuroanatomy: History

5.     The Neuron: Physiology

5.1. Classical representation

5.2. Subtypes

5.2.1.    Histological examples

5.3. Neurophysiology

5.3.1.    Membrane

5.3.2.    Myelin Sheaths

5.3.3.    Sodium/Potassium Pumps

5.3.4.    Synapses

5.3.5.    Vesicles

5.3.6.    Neurotransmitters

5.4. "Firing", Action Potentials, Refractory periods

Week 1 Quiz Due before start of class.

Part 1

6.     The Neuron: Function

6.1. Encoding

6.1.1.    Temporal

6.1.2.    Spatial

6.1.3.    Capacity

6.2. Processing

6.2.1.    Input / Output complexity

Part 2

7.     Natural Neuronal Networks

7.1. Examples

7.2. Complexity

7.3. Functional Connectome

7.4. TEM Connectome

Part 1

8.     Artificial Intelligence: History

8.1. Mythology & Ancient

8.2. Principia Mathematica (1910) [Wikipedia, Vol I, Vol II, Vol III] 

8.3. 1950’s

8.3.1.    Turing

8.3.2.    Logic Theorist (Allen Newell et. al)

8.3.3.    Dartmouth & “The Birth”

8.3.4.    Reasoning as search

8.3.5.    General problem solver

8.4. Golden Age (60’s and 70’s)

8.4.1.    STUDENT

8.4.2.    ELIZA

8.4.3.    SHRDLU

8.4.4.    Animal Game

Part 2

8.5. The Dark Ages (1974-1980)

8.6. Expert Systems (80’s)

8.7. Pause v. 2.0

8.8. Restart (90’s-ish)

9.     Artificial Intelligence: History

9.1. New Millennium (Y2K)

9.2. The missing weight

9.3. The hardware

9.4. Deep Learning reborn

9.5. Watson & Jeopardy

Part 1

NOTE: Modified. We will go over Machine Learning and the Perceptron during this lecture.

10. Artificial Intelligence: Networks

10.1.               Flat (Infinite Theory)

10.2.               Deep

10.3.               Cyclic vs. Acyclic

Part 2

11. Artificial Intelligence: General Principles

11.1.               Nodes

11.2.               Edges

11.3.               Weights

11.4.               Transformational Function

11.5.               Outputs

Part 1

NOTE: Modified. Due to feedback from students, we will go over ANNs to review and expound, particularly on forward and back propagation. I am reducing the amount of statistics and probability review since it is not essential to the broad concepts.

 

12. Statistics and Probability: Review

12.1.               Classical

12.2.               Bayesian

Part 2

12.3.               Bayesian (Continued)

Part 1

13. Classification

14. Task completion

15. Aggregated complexity

Part 2

Midterm Exam

Mar 14 – Spring Break

Part 1

16. Turing Test and beyond

17. Consciousness

Part 2

18. Intent - Bumped to next week.

19. Emotions (Desire) - Bumped to next week.

Part 1

20. Rights of Sentient Beings

20.1.               Existing discourses

21. Rights of Artificial Intelligence

21.1.               In the news

Part 2

22. The general ethical questions of AI

22.1.               Historical

22.1.1.                   Asimov

22.2.               Current

Part 1

23. The Ethics: Autonomous Vehicles

23.1.               Trolley Problem

23.2.               Training control (Mercedes Benz)

Part 2

24. Economic Models

24.1.               Post-producer economies

Part 1

25. Policies and Laws

25.1.               Liability

25.2.               Funding

25.3.               Resources

25.4.               Rights

Part 2

26. Coding Examples

Various Topics

CANCELLED

Part 1

32. The End (aka “Are we all going to die?”)

33. Course Review for Exam

Part 2

34. Further Review & Discussion

Final Exam