DCSA_SSN_24_AID_Education.pdf

SOURCES SOUGHT NOTICE

Defense Counterintelligence and Security Agency

Automated Interview Development – Educational Institutions

1.0 Background

This is a Sources Sought (SS) notice issued by the Defense Counterintelligence and Security Agency (DCSA) for Market Research purposes only. As permitted by Federal Acquisition Regulation (FAR) 10 “Market Research”, and FAR 15.201(e), responses to this notice are not offers and cannot be accepted by the U.S. Government to form a binding contract. This SS notice is issued solely for information and planning purposes; it does not constitute a Request for Proposal (RFP), Request for Quote (RFQ), or a promise to issue an RFP or RFQ. DCSA is not at this time seeking proposals, and as such, responders are advised that the Government will not pay any cost incurred in response to this notice. All costs associated with responding to this notice will be solely at the responding party’s expense. Not responding to this notice does not preclude participation in any future RFP or RFQ. All responses to this request become the property of the Federal Government and will not be returned. This SS notice is being posted as an update to DCSA notice HS0021-24-Q-AID1 due to acquisition strategy updates.

DCSA’s Office of Research and Innovation (ORI) delivers premier credibility assessment capabilities to strengthen and improve national security, law enforcement and intelligence functions. ORI is seeking information on research capabilities that may serve to enhance credibility assessment functions and capabilities within the Federal Government to identify threats to national security, particularly when coupled with state-of-the-art physiological sensors and contactless imaging during credibility assessment interviews.

ORI is seeking information from Colleges, Universities, and Professional Schools regarding capabilities to provide and maintain advanced and sophisticated engineering, research and development in support of the DCSA ORI mission.

2.0 Overview of Requirements:

ORI has a potential requirement for the design and development of a system capable of collecting physiological measures during a credibility assessment interview and using this information, along with novel natural language processing (NLP) features, to direct credibility assessment interviews. The primary objective is to automate the collection of both physiological measures and NLP features, and then integrate these data streams to develop new information that scientists and computer engineers can use for algorithm development. The theoretical and computational foundations surrounding the use of these new NLP technologies as an aid to credibility assessment will also be explored by experts in artificial intelligence (AI), language graphs, and large language models.

Various aspects of an interviewee’s speech, including sentiment (negative, positive, and neutral), emotion (happy, sad, anger, disgust etc.), intent (question, acknowledgement, opinion, referral etc.) will be recognized and classified. These features will be combined in productive ways using artificial intelligence and machine learning techniques. ORI is interested in capabilities for the research and development of a software that can convert examiner and examinee speech into a written transcript and identify deceptive utterances. Physiological features based largely on novel cardiovascular information, to include mean arterial pressure, pulse transit time, and left ventricular ejection time (LVET), along with physiological data channels such as electrodermal activity and respiration, will be extracted and stored in a separate database. ORI is seeking information from potential university, college and professional school partners who may be capable of exploring the utility of these novel physiological features for the development of improved credibility assessment algorithms.

3.0 Critical Task Areas/Elements:

1. Task Areas.

a. Natural Language Processing.

i. Develop Speech-to-text Data Collection System: Includes the software and hardware necessary to acquire speech data during credibility assessment interviews, potentially including acoustic sensors (microphones), amplifiers, processors, mixers, filters, and Digital Audio Workstations. Software should be able to record multiple channels of audio and store speech data.

ii. Develop the Computational and Theoretical Foundations of NLP and Credibility Assessment: Develop a computational method for creating large deception-related corpora founded on linguistic theory. Identify and develop corpora that can be used for development and evaluation. Develop proof of concept detection deception algorithms that can flag and explain suspicious text episodes in a manner that is comprehensible to an interviewer. Employ expert interviewer reactions to evaluate the algorithms and provide input for improving model development and training.

iii. Develop Data Processing and Data Storage Systems: Develop software to recognize the speech data format and appropriate language for transcription.

Azure Cognitive AI Services will be utilized to transcribe multiple channels of acquired speech data from interviews to a text format. Audio should be synthesized, processed, and formatted in near real-time. Develop a database system for storing and retrieving raw audio files, models used for speech-to-text encoders, and final output transcription.

iv. Develop Pattern and Feature Selection System: Automate the manual process of scanning text documents and highlight linguistic features. The system will process raw text files developed in by a speech-to-text system and perform methods in data mining, pattern recognition, and linguistic feature selection using algorithms

v. Develop Graphical User Interface (GUI) for deception characteristics: A user interface will be developed that will display raw speech-to-text input in a scripted format. Deceptive characteristics discovered will be highlighted and displayed in a user-friendly format.

b. Physiological/Thermal.

i. Assess existing Thermal hardware and software capability: Assess and evaluate current software and hardware capabilities necessary to acquire Optical and Thermal Imaging (OTI) data during credibility assessment interviews. Provide recommendations on upgrades to hardware, to include thermal and optical cameras, necessary for blood pressure (BP) estimation.

ii. Evaluate Thermal decision logic for BP Estimation: Assess and evaluate the current thermal imaging system at ORI for accuracy and consistency in contactless BP estimation. Algorithm development will be conducted in parallel to extract variables from the thermal measurements. Specific metrics of accuracy will include: error in mean arterial pressure (MAP), diastolic pressure (DP), and systolic pressure (SP) between the thermal imaging estimation and ground truth measurements. Ground truth will be established using traditional BP measurements. Investigate the effect of spatial resolution and temporal resolution on the analysis. Accuracy will also be quantified during repeat experiment within the same session to ensure data consistency.

iii. Develop algorithms for response variable definition/extraction from OTI:

Algorithm design will involve image segmentation, image analysis, and a statistical framework for maximum likelihood estimation of blood pressure.

Research and quantify MAP, SP, and DP accurately and consistently and define facial regions for which these measurements are least prone to error.

iv. Develop AI and Decision support platforms: Design and/or develop an AI engine capable of analyzing imaging data utilizing non-contact technologies and extract/identify metrics for CA.

c. Feasibility Studies.

i. A feasibility study will be conducted, following approved IRB protocols, to test the workflow of the speech-to-text transcription system. Appropriate methodology will be developed to test the transcription system's functionality, robustness, and accuracy.

ii. A feasibility study will be conducted, following approved IRB protocols, to test the developed system and algorithm design. Appropriate methodology will be developed to test the system’s accuracy and consistency in contactless BP estimation.

4.0 Proposed NAICS and PSC: The NAICS code contemplated for this requirement is 541715 – Research and Development in the Physical, Engineering and Life Sciences. This particular NAICS code has a Business Size Standard of 1,000 employees. The PSC contemplated for this requirement is A31 - National Defense R&D Services; Defense-Related Activities; Basic Research.

5.0 Source Sought Details: This SS notice is issued solely for informational and planning purposes – it does not constitute a RFP or a promise to issue a RFP in the future. This notice does not commit the Government to contract for any supply or service whatsoever. Further, DCSA is not at this time seeking proposals and will not accept unsolicited proposals.

Responders are advised that the U.S. Government will not pay for any information or administrative costs incurred in response to this notice; all costs associated with responding will be solely at the interested Vendor’s expense. Not responding to this notice does not preclude participation in any future RFP, if any is issued.

This SS notice, DCSA_SSN_24_AID_Education, is intended for educational institution responses. A separate SS notice is posted for industry responses.

6.0 Requested Information:

Interested Vendors are requested to respond to this Sources Sought no later than 5:00 PM EST on 17 May 2024 by emailing the Contract Specialist, Jibril Cox, at jibril.h.cox.civ@mail.mil, and the Contracting Officer, Kelsie Kautz, at kelsie.kautz.civ@mail.mil. Verbal questions will not be accepted. DCSA will respond to questions at its discretion. The Government will not respond to further questions regarding the anticipated date of solicitation release.

Proprietary information and trade secrets, if any, must be clearly marked on all materials. All information received that is marked “Proprietary” will be handled accordingly. Please be advised that all submissions become Government property and will not be returned. All Government and Contractor personnel reviewing responses will have signed non-disclosure agreements and understand their responsibility for mailto:jibril.h.cox.civ@mail.mil mailto:kelsie.kautz.civ@mail.mil proper use and protection from unauthorized disclosure of proprietary information as described 41 USC

423. The Government shall not be held liable for any damages incurred if proprietary information is not properly identified.

Responses shall include the following information, at a minimum:

1. Entity Information

a) Company/Institute Name, address, point of contact name, telephone, and email address.

b) CAGE Code and Unique Entity Identifier (UEI).

c) Company/Institute Website.

d) Entity Type.

e) A statement as to whether your company/institute is domestically or foreign owned (if foreign, please indicate the country ownership).

f) Current status in regards to:

Adequate Accounting System in accordance with FAR 16.301-3(a)(3).

DCMA approved Accounting System in accordance with DFARS clause 252.242-7006.

2. Contract Vehicles

Indicate any existing contracting vehicles a future DCSA contract award may be placed against.

3. NAICS and PCS Code Recommendations

Provide confirmation if the existing codes are sufficient, or recommend new codes.

4. Requirements Response

In addition to the business information requested in items 1-3 above, interested Vendors are requested to complete the attached excel spreadsheet by marking an ‘X’ in either the YES or NO column to indicate a positive or negative response to the corresponding element within that line.

There is a column for additional comments where further explanation may be provided to elaborate as necessary. The description column is also available for the equipment section to provide any pertinent details regarding the make/model information for each item, if applicable.

Information regarding similar equipment may also be detailed within the comments section.

DCSA will also accept capability statements up to three (3) pages in length.

To facilitate a timely and comprehensive review of all submitted responses, interested vendors must respond using the format requested in this Notice. Any deviation from this format may lead to the rejection of the response.

All correspondences must include the subject line: “DCSA_SSN_24_AID_Education_ (Vendor name)”

5. Formatting

Please provide the business information addressing items 1-3 in a searchable .PDF format.

Responses must include a completed copy of Attachment 1: DCSA_SSN_24_ Education Vendor Response Microsoft Excel Spreadsheet. Optional capability statements must not exceed 3 pages in length, and must also be submitted in a .PDF format.

7.0 Vendor follow-on meetings: DCSA representatives may or may not choose to meet with potential vendors. Such meetings would only be intended to obtain further clarification of potential capability to meet future or existing requirements.

8.0 Summary: This is a request for information only to identify potential sources that can provide support pertaining to the subject requirement. The information provided in the Sources Sought is subject to change and is not binding on the Government. DCSA has not made a commitment to procure any of the items discussed, and release of this Sources Sought should not be construed as such a commitment or as authorization to incur costs for which reimbursement would be required or sought. All submissions become Government property and will not be returned.