PRESOL | A | Acquaint Phase II | 09-Aug-03

FBO DAILY ISSUE OF AUGUST 09, 2003 FBO #0618
SOLICITATION NOTICE

A -- Acquaint Phase II

Notice Date: 8/7/2003
Notice Type: Solicitation Notice
Contracting Office: DOI - NBC, Ft. Huachuca AZ P.O. Box 12924 ATTN: Nancy Johnson 520-538-0421 Fort Huachuca AZ 85670
ZIP Code: 85670
Solicitation Number: BAA-03-06-FH
Response Due: 9/24/2003
Archive Date: 8/6/2004
Point of Contact: Lawrence Carter Contracting Officer's Rep. 5205331213 lawrence.carter@hua.army.mil;
E-Mail Address: Email your questions to Point of Contact above, or if none listed, contact the IDEAS EC HELP DESK for assistance
(EC_helpdesk@NBC.GOV)
Small Business Set-Aside: N/A
Description: 1. INTRODUCTION: The National Business Center, acting as the contract agent for the Advanced Research and Development Activity in Information Technology (ARDA), is soliciting proposals for Phase 2 of the AQUAINT (Advanced Question Answering for Intelligence) Program. ARDA (<A HREF=http://www.ic-arda.org/>http://www.ic-arda.org/</A>) is a research organization jointly established in December 1998 by the Department of Defense and the U.S. Intelligence Community. While the ARDA office is organizationally part of the National Security Agency, ARDA's mission is to incubate revolutionary Research and Development (R&D) activities within the broad field of Information Technology. ARDA's individual research programs are organized around broad research topics called "Research Thrusts". One of these Research Thrusts is Information Exploitation and one of the research programs under this Thrust is the AQUAINT Advanced R&D Program. 2. AQUAINT PROGRAM GOALS: Phase 2 of the AQUAINT Program is the next two-year effort in the six-year question-answering (Q&A) program; it will be initiated at the end of 2003/early 2004 and run through the end of 2005/early 2006. The AQUAINT Program seeks proposals for innovative, creative, and high-risk research, continuing to advance the state-of-the-art in technologies and methods for advanced, automated question answering. The overall AQUAINT Program goal is to explore and develop innovative technical designs, algorithms, methods, processes, technologies or techniques with the potential of significantly advancing, directly and materially, one or more of the following facets of question answering: * Accept complex "Questions" in a form natural to an information analyst * Translate these questions into multiple queries appropriate to the multiple data sources to be searched and to the analyst's information-gathering task * Find relevant information in multiple, disparate data sources * Integrate features of the analyst's larger information-gathering task into the information identification process to help define what constitutes an adequate answer * Analyze, fuse, evaluate and summarize the relevant information extracted or otherwise derived from these multiple data sources into a coherent answer * Provide access to related, adjunct information uncovered in the process of searching for answers to specific questions * Provide an answer back to the information analyst in a form natural to and preferred by him or her in a timely fashion 2.1. Phase 1 Research Goals and Accomplishments: The Phase 1 technical goals focused on the following three Functional Components and Enabling Technologies in support of the AQUAINT Program goals: 1. Question Understanding and Interpretation 2. Determining the Answer 3. Formulating and Presenting the Answer And/Or Cross-Cutting/Enabling/Enhancing Technologies directly and materially supporting the goals of the AQUAINT Program and one or more of the areas 1 to 3 listed above. 2.2. Phase 2 Research Goals: During the life of the AQUAINT Program, advances are expected to demonstrate significant increases in capability over existing technologies and over the successes of previous phases of the Program. Therefore, in addition to pursuing the goals identified in Phase 1 of the Program, Phase 2 will encourage efforts, which focus on the following challenges: * Question Answering as Part of a Larger Information-Gathering Process * Accessing, Retrieving and Integrating Diverse Data Sources * Exploring Boundaries / Combinations of Knowledge-based, Statistical and Linguistic Approaches to Question Answering * Evaluating, Validating and Presenting an Answer 2.2.1. Question Answering as Part of a Larger Information-Gathering Process. During Phase 2, the AQUAINT Program seeks to focus on the development of algorithms and methods addressing question answering as a cooperative, information-gathering process between a user and the system. In standard interactions, questions are rarely asked in isolation; instead, they are linked to a larger information need and sequenced to explore an information space. The information need helps to define how a question should be interpreted and what information would be considered adequate as an answer. The question and anticipated answer in a cooperative environment are: * Conditioned by a knowledge frame, which the analyst brings to the interaction; * Associated with a larger information-gathering task; * Shaped by previous questions and by additional topic-related information; and * Influenced by an analyst's expectations about the format, content, reliability of an answer to the question. The AQUAINT Program would expect, in the end, an advanced Q&A system able to generalize larger information-gathering tasks, anticipating the sorts of information which might be useful for such tasks; would decompose a less predictable task; would establish an extended interactive session to refine the task, and would identify domains of knowledge which the analyst brings to the task. An interactive session with the analyst would validate the system's understanding of the analysts' information needs, the scope of relevant information, additional paths of inquiry and the analyst's expectations about the adequacy, extent and format of the answer. Such knowledge of the context would then serve as a resource for selecting suitable retrieval strategies and for assigning value to the data discovered. 2.2.2. Accessing, Retrieving and Integrating Multiple, Diverse Data Sources. Rarely are the sources of information relevant to questions about global topics of one homogeneous type. Relevant information can be found in highly diverse and extremely large repositories of unstructured text; knowledge bases; structured databases; Web pages and other HTML and XML tagged documents; document images (e.g., scanned or faxed documents); voice recordings; images (both still and video); and other multi-media data objects. The data sources of interest to the Program should be viewed as widely distributed and subject to frequent and significant fluctuations in size and to continual updates in content. In addition to the diversity in the types of data, the human language content of these data sources can vary dramatically. Sources may be expressed in English but, more and more frequently, material is found in a foreign language or in multiple foreign languages, in a variety of scripts, a range of fonts, a multitude of formats. An advanced Q&A system will be able to address the broad range of data and language types, accessing and manipulating them and integrating answers or pieces of answers from these diverse sources into a cohesive response to a question. 2.2.3. Exploring Boundaries / Combinations of Knowledge-based, Statistical and Linguistic Approaches to Question Answering. Prior to and during Phase 1 of the AQUAINT Program, most question answering systems focused primarily on one technical approach, emphasizing knowledge-based reasoning, statistical, or linguistic techniques and theories. As the AQUAINT Program progresses and the question answering tasks become broader and more complex, it is unlikely that any one single approach will be able to handle all of the AQUAINT Program requirements for advanced question answering. In particular, we believe that three areas warrant additional attention. * We believe there are a number of challenging/unsolved problems involving knowledge acquisition, learning, representation and maintenance within an advanced question answering system as well as problems involving significantly more advanced and broadly applicable reasoning, influence and machine learning that builds on and utilizes these more expansive and comprehensive knowledge sources. * There are a number of challenging/unsolved problems whose solution might involve a combination of multiple approaches (e.g., decomposing complex questions into a series/sequence of simpler questions whose answers can be found; selecting the appropriate sources to search; giving understandable explanations of the "Plan", the "Reasoning Used" and the "Answers Found"). * Finally, each of the single disciplinary approaches has unique strengths and weaknesses. Fortunately for advanced question answering, they have significant areas where the strengths of one approach match up against weaknesses of another approach. The AQUAINT Program wishes to explore more fully and investigate the potential power gained by combined multiple approaches into a single advanced question answering system. 2.2.4. Evaluating, Validating and Presenting an Answer. Important progress has been made on formulating and presenting a natural-language answer to a question. Presentation of documents or document snippets containing the answer is giving way to formulation of appropriate answers in the form of well formed, natural-language sentences, lists, descriptions and so forth. Continued movement toward enhanced answer formulation will also depend on being able to generate answers to questions in a form natural to the information analyst when the answer is found in multiple documents, in different formats, in various media, in multiple languages and in varying styles. Not only is it important for advanced Q&A systems to be able to present information appropriately but the system must also be able to assess the likely relevance, value and reliability of the information. Reliability indices will also address the challenges of identifying when data are missing from the sources searched, when information is likely to be false and when no on-topic information was discovered. 3. SCOPE: The AQUAINT Program is a 24-month program, open to all research and development organizations located in the U.S., including large and small businesses; academic and eligible non-profit and not-for-profit institutions; collaborative ventures from mixed sources; and Federally Funded Research and Development Centers (FFRDCs). All AQUAINT Program Phase 1 participants must submit proposals in response to this BAA to be selected for award under Phase 2. Funding for Phase 2 of the AQUAINT Program will utilize funds from FY 2004-2006 and is anticipated to total $22,000,000. The Government anticipates funding approximately 15-20 proposals at varying levels of effort ranging from highly focused studies utilizing 1 to 1 1/2 FTE researchers/investigators and aimed at a single, key technical sub-problem to component / system-level investigations utilizing three to four FTE researchers / investigators seeking to address larger, more comprehensive, component wide and/or end-to-end question and answering issues. It is anticipated any resultant contract(s) will be funded for two years, with each year's level approximately $200,000 to $300,000 for a highly focused study up through $600,000 to $900,000 for a more comprehensive component / system level effort. Please note this is an approximation only and the cost proposal should reflect the realistic cost of the proposed work. Information in this BAA is supplemented by a Proposer Information Pamphlet (PIP) providing further information on the eligibility, support, research areas, multidisciplinary approaches, preparation and submission of proposals, and selection criteria. It is available at <A HREF=http://www.nbc.gov/acquaint.cfm>http://www.nbc.gov/acquaint.cfm</A>. This pamphlet may also be obtained by fax, electronic mail, or official mail. Send requests for PIP to the Department of Interior, Fort Huachuca by email to <A HREF=mailto:Lawrence_H_Carter@nbc.gov>Lawrence_H_Carter@nbc.gov</A> or call 520-533-1213. This Broad Agency Announcement constitutes the entire solicitation for this effort. No additional information (beyond the PIP) is available, nor will a formal request for proposal or other solicitation regarding this notice be issued, requests for the same will be disregarded. 4. PROPOSAL PREPARATION INSTRUCTIONS: See the Proposer Information Pamphlet (PIP) which provides preparation and submission instructions for proposals. It is available on the web site listed above. 5. PROPOSAL SUBMISSION: Proposals are due on or before 4:00 PM, Mountain Standard Time, 24 September 2003, to the Department of the Interior, National Business Center, Acquisition Support Division, Southwest Branch, Post Office Box 12924, ATTN: BAA (N. Johnson), Fort Huachuca, Arizona, 85670-2924. See the Proposer Information Pamphlet (PIP) which provides detailed instructions on proposal submission procedures. It is available on the web site listed above. 6. PROPOSAL EVALUATION AND SELECTION. Proposals will be selected through a technical/scientific/business decision process using the following criteria in descending order of importance: (1) Overall scientific and/or technical merit, including technical feasibility, degree of innovation, understanding of the technical and operational approach, and experimental approach. If a proposal lacks overall scientific and/or technical merit, it will not be further considered for award. (2) The effort's potential contributions to the Intelligence Community's Information Assurance efforts. (3) Cost reasonableness and realism. Awards under this BAA will be made to responsible offerors on the basis of the evaluation criteria above and a BEST VALUE approach to the Government. Awards will be subject to the availability of funds. Awards may take the form of a procurement contract. The Government reserves the right to 1) select for award all, some, one, or none of the proposals received, and 2) incrementally fund any award instrument. The Government also reserves the right to fund all or any part of a proposal evaluated to be eligible for award. Awards are subject to the availability of Government funds.
Web Link: Please click here to view more details.
(http://ideasec.nbc.gov/ecprod/owa/ec$cbd.sypfirstcount?P_SERVER_ID3=NB1401FH&P_OBJ_ID1=124912)
Record: SN00395225-W 20030809/030807213938 (fbodaily.com)
Source: FedBizOpps.gov Link to This Notice
(may not be valid after Archive Date)

| FSG Index | This Issue's Index | Today's FBO Daily Index Page |

Loren Data's SAM Daily™

A -- Acquaint Phase II