|
Executive
Summary
November 14, 2003 MIT's Center for Bits and Atoms is an ambitious interdisciplinary initiative that is looking beyond the end of the Digital Revolution to ask how a functional description of a system can be embodied in, and abstracted from, a physical form. These simple, profound questions date back to the beginning of modern manufacturing, and before that to the origins of natural science, but they have revolutionary new implications that follow from the recognition of the computational universality of physical systems. We can no longer afford to ignore nature's capabilities that have been neglected by conventional digital logic; it is at this boundary between the content of information and its physical representation that many of science's greatest technological, economic, and social obstacles and opportunities lie. CBA was founded by Profs. Isaac Chuang, Neil Gershenfeld, Joseph Jacobson, and Scott Manalis, with Marvin Minsky. It was launched by a National Science Foundation award in 2001 [1] that is supporting the creation of a unique shared experimental resource that enables the creation of form and function across nine orders of magnitude in length scales [2], as well as an associated intellectual community drawn from across MIT's campus [3] spanning the historical divisions that have emerged between the study of computer science and physical science, and between the development of software and hardware. CBA's government funding is complemented by corporate sponsorship for technology development and transfer [4]. The origins of the CBA program lay in seminal studies by Leo Szilard introducing the bit as a unit of information about the location of a gas molecule (1929), Claude Shannon showing that encoding information digitally can create a threshold allowing for perfect communication over a noisy channel (1948), by John von Neumann (1952), and Shmuel Winograd and Jack Cowan (1963), extending this result to prove that reliable digital computation can be done by unreliable analog components, and by von Neumann (1957) on self-reproducing machines. The principles underlying modern digital technology grew out of these studies of physical systems. However, if nature can compute it can also be understood using the language of computation, an insight that was pioneered by Rolf Landauer's (1961) explanation of the origin of physical dissipation in logical erasure. John Archibald Wheeler (1989) described this aim as "it from bit," meaning that our world (and us) are most fundamentaly understood as a manifestation of information. This has served as a provocative philosophical principle; in CBA it has become applied engineering practice. One of CBA's grand-challenge goals is quite literally to create "it from bit," seeking to realize von Neumann's vision of a universal assembler. Analogous to the earlier results in communications and computation, if logic can be introduced into the process of physical assembly then perfect macroscopic structures could be built out of imperfect microscopic parts. Biological proteins are in fact produced in exactly this way, by programs run by cellular molecular machinery; CBA researchers have shown how nanocluster antennae can be attached to these proteins in order to provide for radiofrequency control over cellular signaling pathways [5]. This promises to create a "digital" technology for molecular manufacturing, with implications for atoms as profound as they have been for bits. And just as personal computers made the capabilities of mainframes accessible to ordinary people, CBA researchers are now doing the same with industrial machine tools, developing means for "personal fabrication" that will bring the programmability of the digital world to the rest of the world [6]. In the opposite direction, the CBA program can be understood as seeking to create "bit from it," abstracting and applying logical descriptions of physical systems. This effort starts with studies of transduction at the interface between digital and physical systems [7], includes investigation of quantum [8] and classical [9] mechanisms for manipulating information, and encompasses the development of architectural principles for organizing these resources into scalable [10], "fungible" [11] systems. Ultimately, these activities are leading towards a design practice for "Avogadro-scale" engineering that can bring rigor to the control of enormously complex systems without requiring an explicit specification of their internal configuration [12]. CBA's development of devices and algorithms in this "statistical-mechanical" limit extends early work on information-processing in graphical networks by Warren McCullough and Walter Pitts (1943), and Marvin Minsky and Seymour Papert (1969) [13]. The academic home for CBA is the MAS Program (Media Arts and Sciences), directed by William Mitchell, broadly encompassing the study of design across intellectual and physical scales [14]. Along with teaching and research, CBA outreach activities include meetings and events run with partner institutions [15], a growing network of field "fab labs" that are bringing tools for technological design and production to remote parts of the world that have been beyond the reach of conventional technical solutions [16], and industrial coordination such as the "Internet 0" initiative that is extending the Internet's "end-to-end" principle of internetworking all the way down to the device level [17]. References
|