The Human Intelligence Enterprise

Information, Mind, and the Human Interface

A Proposal for Enterprise-Centered Education

January 1997

Robert C. Berwick, Jay L. Jaroslav, Thomas F. Knight, Jr.,
Gerald Jay Sussman, Shimon Ullman,
Patrick Henry Winston, and Kenneth Yip

Education and research are mutually enabling. A healthy educational program stimulates research, and an exciting research program enables new educational tracks.

Of course, the development and marketing of educational programs can take time, so we do not necessarily expect to rush to, say, a new major, and we will not despair if practical considerations were to rule out a substantial education program. Nevertheless, we are excited by what we think can be done in a substantial educational program, with a label such as Information, Mind, and the Human Interface.

Accordingly, with a view toward demonstrating that an exciting undergraduate program can be developed, we lay out a sample undergraduate curriculum. We defer the tactical question of how such a program can be adjusted so as to conform to reality without losing vitality.

Our sample curriculum brings together subjects and subject fragments from various existing MIT departments, including Electrical Engineering and Computer Science, Brain and Cognitive Science, Linguistics, and perhaps Media Arts and Technology. We understand that our illustration is too heavily weighted toward what we have come to know and find essential. We further understand that the route to a better curriculum lies in working with representatives from other contributing disciplines, starting from a collection of such drafts.

We believe the sample curriculum, or one like it, would prepare participating students for exciting futures, both applied and academic. The curriculum will attract some of the gifted students who now decide against technically oriented institutions, because they perceive that institutions with strong offerings on the mind's artifacts must have only weak offerings on the mind itself.


Students will be attracted to the Information, Mind, and the Human Interface program, in quantity, for several reasons:

Novel Ideas


The subjects below include traditional subjects plus required short subjects, exam subjects, and a major opus. It is obvious, therefore, that our work toward developing a curriculum has only begun, inasmuch as the list is thin in the Mind and Human Interface dimensions and inasmuch as we have listed more than eighteen subjects, which is the number that traditionally constitutes a full undergraduate major.

Computing Subjects

The subjects in the core of the EE&CS curriculum contain material that is essential for just about anyone who is supposed to be a technically trained person. Two of these subjects develop familiarity with the basics of computer programming and hardware design. The computing subject is essential both because it supplies a language for talking about computing and because it develops skill in programming. The hardware subject is essential because it provides a context for looking at nature's hardware.

Subject Need MIT Approximation
Programming 6.001
Computing Hardware 6.004

Engineering Subjects

Students need to develop a sound understanding of the modeling process and the standard techniques for producing models. In this area, the EE&CS core provides subjects that, as usually taught, develop such understanding:

Subject Need MIT Approximation
Modeling Electrical Circuits 6.002
Modeling Linear Systems 6.003

Algorithms and Information Theory

An understanding of algorithms provides guidance in the design of experimental systems. An understanding of information theory provides ingredients for theories about how brain mechanisms solve communications problems.

Subject Need MIT Approximation
Algorithms ?
Information Theory ?

Mathematics Subjects

Beyond the general institute requirements, there are mathematics subjects that every veteran of the program must master. Candidates favored at the moment include:

Subject Need MIT Approximation
Probability 6.041
Statistics for Computation and Cognition ?
Differential Equations18.03
Abstract Algebra18.701

Core Artificial Intelligence Subjects

To build practical systems, students need to understand traditional problem solving and learning paradigms. To understand the computational basis for intelligence, they need to understand the contributions of at least the vision, language, and motor faculties. Such subjects are at the center of the information part of the program:

Subject Need MIT Approximation
Introduction to AI 6.034
Project Laboratory 6.916
Problem Solving in Perceptual Systems ?
Great Papers ?

There are, of course, many other AI subjects that provide a rich reservoir of potential electives. For example, subjects such as Embodied Intelligence (6.836) and Medical Informatics (6.894) present specialized points of view or highly focused, practically oriented material.

Core Brain and Cognitive Psychology Subjects

We are, of course, less able to assess what subjects should be at the center of the mind part of the program, but we imagine that the following are good candidates:

Subject Need MIT Approximation
Introduction to Psychology 9.01
Experimental Methods ?
Cognitive and Developmental Psychology ?
Linguistics and Cognition ?
Neurophysiology and Brain Anatomy 9.011

The Human Interface

Our biggest need is to think through the Human Interface part of the program. Potential subjects include:

Subject Need MIT Approximation
Human Factor Psychology ?
Human Interface Technology ?

The human-interface subject might, conceivably cover elements of speech and graphics systems.

Interdisciplinary Subjects

Subject Need MIT Approximation
The Neuron Viewed as a Device ?
The Brain Viewed as a Computational Mechanism ?
Computational Cognitive Science ?
Computational Biology ?

Short Subjects

Certain skills are best taught in intensive, hands-on bursts, in short subjects. Candidates include the following:

Information Search
Building Experimental Apparatus
Commercial Programming Tools and Methods

Exam Subjects

With a view toward helping students learn to learn on their own, certain subjects, with well-worked out and circumscribed content, will be taught via a reading list. Candidates include elementary aspects of the following:

Numerical Methods
Modeling and Simulation
Scientific Literacy

The subject in scientific literacy will include a treatment of estimation, dimensional analysis, units, and useful-to-know physical quantities.