Professor, Architecture, Interior Architecture, and Designed Objects (1989). BA, 1971, University of Chicago; BArch, 1976, University of Illinois at Chicago. Exhibitions: Chicago Athenaeum; Betty Rymer Gallery; I-Space, Gallery 400, Chicago. Publications: Architectural Record; Progressive Architecture; Architectural Digest. Editor: Inland Architect, Chicago Architectural Journal. Collections: Art Institute of Chicago, Chicago Historical Society, Deutsches Architecture Museum. Awards: Chicago Chapter AIA Awards; Graham Foundation Grant.
Experience at SAIC
SAIC is unique and protean. It has a culture unlike any other art school, and it can be hard to get a fix on it. But if you have a vision for how you are going to tack across the disciplines, there is no better place to work.
Loose Neural Nets in Architecture
The human brain can be understood as a massively complex net of neurons, and consciousness can be understood as emergent behavior in that net. In a dramatically scaled down way, intelligence in buildings can be assembled by putting together lots of relatively stupid microprocessors, in a net of influence and propagation, using light impulses and physical contact as signals, as well as typical electrical impulses.
These recently constructed research projects construe loose networks of analog integrated circuits as neural networks, and make the case that emergent behavior in these neural nets forms a kind of low level intelligence. Systems of sensors, emitters, and analog processors embedded into building components like ceilings and skins, can function in a way that is similar to primitive neurons in an organism, propagating signals through adjacent processors, ultimately affecting the characteristics of interior space and the performance of exterior building systems.
In the design of these systems, it is useful to ascribe some autonomy to the embedded components of the net, and also to the human inhabitants of the building. In these projects, a human doesn't overtly control the state of the net and its impact on the inhabitant, but instead the net is tuned to "watch" for certain behaviors and then to respond. The net's ultimate response is not a directly programmed reply to that behavior. Instead, the net's response and the subsequent realization and responses of the inhabitants can be understood as symbiotic emergent behavior.
Flocking and schooling behavior provides a useful simple example of emergent behavior. In those, birds and fish navigate with two levels of goal, in a kind of subsumption architecture - a high level, seeking of polarized light or water without big shadows - and a low level, seeking to avoid collisions with neighbors. If a neural net cycles through these assessments fast enough, both goals are simple to achieve, and the overall behavior gives the impression of a greater intelligence that may actually be built in.
A set of independent mobile lights in a ceiling are designed to move about, scan below for their subject to illuminate, and scan each other to avoid collision. Simple analog comparators and phototransistors end up functioning in a way similar to the destabilized oscillations of a flip-flop - the precursor to a stable bit of computer memory - and they produce apparent goal orientation and the avoidance of collisions.
A set of photovoltaic collectors in an exterior wall are designed to collect energy whenever they can, but to release that energy in some very specific specific ways. Two types of simple analog "cells" are designed to co-exist in the skin. One set that responds to faster diurnal variations in illumination and provides timed pulses of electricity that adjust shading "scales" in the skin to shade the interior as an effective seasonal bris-soleil, opened wide on the side away from the sun, and angled for carefully metered in-solation on the sides facing the sun. The other set of "cells" responds to slower, gross variations in diurnal illumination, and at a certain point releases all of its stored electricity in pulses that closes the skin 's shading and insulation scales tighter in the evening.
Anders Nereim is a Professor of Architecture at the School of the Art Institute of Chicago, former Department Chair and Director of the professional Master of Architecture program there, and Chairman of the Faculty Senate. During high school he worked summers building analog electronic instrumentation at the Borg Warner Research Center. He studied bio-psychology with Dr. Harold Moltz at the University of Chicago, and he learned Fortran as a student of architecture at the University of Illinois at Chicago. Between 1975 and 1995 he practiced architecture, winning several AIA awards. Since 1996 he has been actively engaged in the construction of research prototypes that use networks of simple processors to achieve the appearance of greater intelligence in building components.
I am interested in zoning envelopes that encourage cities to shape the air that cloaks them, and gather their own wind energy. I am interested in efficient building envelopes that gather low voltage direct current for use inside, completely ignoring the alternating current power grid.
Disclaimer: All work represents the views of the INDIVIDUAL ARTISTS & AUTHORS who created them, and are not those of the school or museum of the Art Institute.