Bid Office, SPAWARSYSCEN, Code D21B, 53570 Silvergate Avenue, Bldg. A33, Room 0061, San Diego, CA 92152-5112

A -- RDTE/OTHER SCIENCES -- BASIC RESEARCH SOL N66001-00-Q-A153 DUE 031300 POC Purchasing Agent, Melissa Dolan (619) 553-6336; Contracting Officer, Sylvia Proffit (619) 553-3292 This is a combined synopsis/solicitation prepared in accordance with the format in the Federal Acquisition Regulation, supplemented with information included in this notice. This solicitation incorporates provisions and clauses at FAR 52.213-4 Terms and Conditions -Simplified Acquisitions (Other than Commercial Items) (FEB 2000), and any others that are applicable and in effect through the current Federal Acquisition Circular (FAC) and the Defense Acquisition Circular. It is the contractors responsibility to be familiar with all applicable clauses and provision. No RFQ will be prepared or made available for distribution. The government intends to award, on a Sole Source basis, an order to Neurosciences Research Foundation, Inc., for the following task: Background: At the core of the total information awareness (TIA) architecture, in the "transition zone," is a box. Such boxes are often seen in advanced R&D projects. It might as well be labeled "a miracle happens here." In this box, somehow, multiple incoming sensor streams from the near and far field are fused together, presumably aided by models and substantial prior knowledge of the target and environment, to produce an integrated view of the situation. TIA correctly recognizes that this is likely to remain an intractable problem for traditional automated approaches to information processing, and assigns, instead, this task to humans. But the TIA vision cannot be realized unless humans are augmented in a substantial way by machines in this task: there are simply too many data streams to integrate, too many persons and scenarios to attend to. Large-scale computer models that capture the principles of the underlying neuroanatomy and physiology of the brain have been utilized to investigate several aspects central to the question of sensory integration, perceptual categorization and the development of perceptual invariance. In addition to the modeling of large brain networks a number of statistical measures, derived from information theory, have been developed to illustrate and quantify some essential characteristics of brain activity. In addition to their use in characterizing the global statistical properties of brain activation these information theoretic measures can also be used as variables to optimize in abstract networks. There are two such studies currently underway. The first utilizes a selectionally based optimization routine and elements of graph theory to investigate the outcome of optimizing for various global statistical parameters, including neural complexity. In a second study the use of complexity matching as a way to produce networks that perform novelty detection is being investigated. Networks, consisting of multiple input modalities, are trained, using a selection-based algorithm, to optimize complexity matching across a stimulus set. Comparing values of complexity matching obtained when different inputs are presented to such trained networks are used to detect novelty. SCOPE: The idea of a Perception-Turing machine proposes to combinethese multiple avenues of research into a system that can integrate sensor information from multiple streams into a single, coherent worldview. Models, based on functional mechanism of the brain, would be utilized to produce a system that performs perceptual categorization in multiple sensory modalities and simultaneously produces a global integrated state corresponding to the current view of the world. The various statistical measures, or derivatives thereof, could then be used to provide a bridge between these perception-based models and traditional computer algorithms to detect novelty, abnormality or salience in the current environment. TECHNICAL REQUIREMENTS: The contractor shall develop a conceptual but actionable definition of a Perception-Turing Machine. The contractor shall explain how such a conceptual definition could contribute to the goals of TIA. The contractor shall identify the components of a four year research and development program aimed at delivering operational functionality in the TIA transition zone as well as contribute insights to the architectural evolution of the TIA concept. The contractor shall specify areas of experimentation necessary to confirm the results to date as well as validate the proposed R& D program. The contractor shall identify key uncertainties that would be subject to future basic scientific investigation. GOVERNMENT FURNISHED PROPERTY: The Government will provide the contractor access to facilities, computer networks, computer equipment, data or services required for the performance of this task. TRAVEL: To the maximum extent possible, tasks shall be accomplished by electronic media over existing high bandwidth or other communication links rather than by contractor travel. It is estimated that the following trips will be required for the performance of this task: four trips of three days for one person to Washington, DC to meet with DARPA/ISO. OTHER: Security: NONE. Place of Performance: The Neurosciences Institute, 10640 John J. Hopkins Drive, San Diego, CA 92121. Inspection and acceptance: Dr. William S. Lapp, SPAWARSYSCEN, Code D44202, Building 606, Room 223, Phone: 619-553-2757. Period of Performance: Period of 12 months from date of receipt of award. Facsimile quotations may be sent to (619) 553-1062. Questions may be directed to Melissa Dolan (619) 553-6336, email: dolanm@spawar.navy.mil. Note 22 applies, however, all offers received within 15 days (in lieu of 45 days) after date of publication of this synopsis will be considered by the government. See numbered notes from any Monday edition of the CBD. The applicable Standard Industrial Classification Code and Size Standard is: 8733/$5.0. Note: The full text of the Federal Acquisition Regulation (FAR) can be accessed on the Internet at http://farsite.hill.af.mil/. Offerors must complete and submit with their quote, a copy of 52.219-1 Alt I/II, Small Business Program Representations. Central Contractor Registration is required for resulting award. Information may be found at http://www.ccr2000.com/. Posted 02/23/00 (W-SN427692). (0054)

Loren Data Corp. http://www.ld.com (SYN# 0004 20000225\A-0004.SOL)