IARPA-BAA-24-03_BENGAL_Amend 003 FINAL 11 29 23 C.pdf

Broad Agency

Announcement for

Bias Effects and Notable Generative AI Limitations

(BENGAL) Targeted Super Seedling Research Topics

IARPA-BAA-24-03

BAA Release Date:

November 17, 2023

Amendment 001: November 20, 2023

ALL CHANGES FOR AMENDMENT 001 ARE HIGHLIGHTED IN YELLOW

Amendment 002: November 22, 2023

ALL CHANGES FOR AMENDMENT 002 ARE HIGHLIGHTED IN LIGHT BLUE

Amendment 003: November 29, 2023

ALL CHANGES FOR AMENDMENT 003 ARE HIGHLIGHTED IN PINK

Table of Contents

1 Overview Information

2 Award Information

3 Eligibility Information

4 Application and Submission Information

5 Evaluation of White Papers/Proposals

6 Award Administration Information

7 APPENDIX A - BAA Attachments

1 Overview Information

• Federal Agency Name – Office of the Director of National Intelligence/Intelligence Advanced

Research Projects Activity (IARPA)

• Funding Opportunity Title – Bias Effects and Notable Generative AI Limitations

(BENGAL) Targeted Super Seedling Research Topics

• Announcement Type – Initial

• Funding Opportunity Number – IARPA-BAA-24-03

• Questions - Submit questions on administrative, technical, or contractual issues by email to

DNI-IARPA-BAA-24-03@iarpa.gov. All requests must include the full name and affiliation of a point of contact. Do not send questions with proprietary content. A consolidated

Question and Answer response will be posted on SAM.gov for Contract Opportunities website (https: sam.gov/) and linked from the IARPA website

(https://www.iarpa.gov/index.php/research-programs). No answer will go directly to the submitter. IARPA will accept questions until December 14, 2023, 4:00 PM EST.

• Dates o Posting Date: November 17, 2023 o White Papers (abstracts) Due Date: December 5, 2023 12:00 PM EST

(Offerors are strongly encouraged to submit white papers before submitting a proposal. The submission window for white papers will open on December 1, 2023

12:00PM EST and close on December 5, 2023 12:00PM EST.

o Proposal Due Date: January 29, 2024 4:00 PM EST. (A BAA amendment will be issued to announce subsequent rounds of selections, if any)

• Anticipated individual awards – Multiple awards

• Types of instruments that may be awarded – Procurement Contracts

• Agency Points of contact –

ATTN: IARPA-BAA-24-03

Office of the Director of National Intelligence

Intelligence Advanced Research Projects Activity

Washington, DC 20511

Electronic mail: dni-IARPA-BAA-24-03@iarpa.gov.

Phone: Contracting Officer (CO), 703-275-2195 (email is preferred)

• Program Website - https://www.iarpa.gov/index.php/research-programs

• BAA Summary – IARPA is soliciting proposals for early-stage research (which IARPA refers to as “seedlings”) for Bias Effects and Notable Generative AI Limitations (BENGAL).

Initially, interested parties are encouraged to submit a white paper so the Government can mailto:dni-IARPA-BAA-21-02@iarpa.gov https://www.iarpa.gov/index.php/research-programs review their ability to meet the intent of the BAA. Formal proposals will then be encouraged or not encouraged based on this white paper review. (See BAA Sections 4 and 5).

• Program Overview

The Intelligence Advanced Research Projects Activity (IARPA) invests in high-risk/high-payoff research programs that have the potential to provide our nation with an overwhelming intelligence advantage. IARPA seeks to develop new capabilities to enable the safe adoption and use of generative AI technologies to greatly enhance the effectiveness and efficiency of the Intelligence Community (IC).

Large Language Models (LLMs) exhibit impressive, human-seeming conversational capabilities. LLMs and applications that build ‘on top’ of these models are being rapidly adopted and are expected to transform work across diverse sectors. It is anticipated that the public will interact with a massive number of LLM-derivative technologies within this decade.

Even at this early stage in their adoption, however, the public has observed that LLMs can exhibit erroneous or potentially harmful behavior. The inherent characteristics of LLMs (ease of use, human-like dialogue, complexity, and lack of explainability) present vulnerabilities for benign applications and enable hostile applications. Models may conceal threats to users, including quick generation of mis/disinformation or elicitation of sensitive information. These threat modes may be unintended emergent artifacts of training complex models on vast and poorly understood training data, or they may be intentionally incorporated into models by their designers.

The IC is interested in safe uses of LLMs (multi-modal and text-only) for a wide variety of applications including the rapid summarization and contextualization of multilingual information relevant to the IC. These applications must avoid unwarranted biases and toxic outputs, preserve attribution to original sources, and be free of erroneous outputs. The US

Government is also interested in identifying and mitigating hazardous use of LLMs by potential nefarious actors.

The goal of the BENGAL targeted super seedling is to understand LLM threat modes, quantify them and to find novel methods to correct threats and vulnerabilities or to work resiliently with imperfect models. IARPA seeks to develop and incorporate novel technologies to efficiently probe large language models to detect and characterize LLM threat modes and vulnerabilities.

Performers will focus on one or more of the topic domains below, clearly articulate a taxonomy of threat modes within their domain of interest and develop technologies to detect, defeat or mitigate these threats.

1.1 Technical Challenges and Objectives

IARPA seeks novel research ideas from multidisciplinary teams pursuing advanced research topics capable of supporting the interests described below:

• Biases and induction of diverse analytical perspectives: The IC seeks capabilities to enhance awareness of LLM output biases (cognitive, demographic, ideological, cultural, temporal, etc.) that might suppress or promote relevant content with the effect of misleading a user. IC users also greatly benefit from exposure to contrasting perspectives on the same event or situation. IARPA is interested in novel technologies to accurately and automatically characterize and detect biases, as well as leverage LLMs to induce diverse perspectives on events or states of affairs.

• AI hallucinations and inferences: Generative LLMs are known to produce spurious, ungrounded outputs (‘hallucinations’ or ‘confabulations’) that can cause erroneous analysis and decision making. Successful methods for reducing hallucinations constrain LLM outputs to those that are in some fashion corroborated by ground truth (e.g., quotations from a trusted document). However, constraints enforcing grounded outputs are over-restrictive and block

LLMs from drawing correct or plausible inferences, since good inferences are often not

(straightforwardly) based on ground truth evidence. IARPA is interested in capabilities that detect spurious hallucinations while maximizing and inducing correct and plausible inferences.

• Safe information flow in sensitive environments: The IC’s classification system limits access to sensitive information, which if disclosed to unauthorized individuals poses a grave threat to national security. However, restricting information to certain people and systems can pose a threat to national security by preventing critical collaborations and timely sharing of critical information between IC organizations. IARPA is interested in LLM technologies that increase the flow of information while minimizing the likelihood of sensitive information disclosure, enabling safe use of LLMs over the broadest range of data and tasks. (Please note that performers will not be permitted access or test systems on sensitive data; thus, offerors must propose unclassified testing and evaluation schemes that credibly simulate environments where sensitive information is stored but not all users are authorized to access it.)

• Working resiliently with imperfect or poisoned sources: The IC is interested in novel LLM techniques to evaluate the reliability of specific information sources (e.g., news organizations, individual content producers from around the world), especially sources whose content is used to train LLMs or to interpret developing situations using LLMs in scenarios where only sparse information is available. The IC is also interested in technologies that enable analysts and other users to work resiliently with imperfect or malicious sources, identifying within their content reliable information where possible.

Efforts addressing these topic areas align well with needs of the intelligence and national security communities and are, therefore, under the purview of IARPA’s research mission.

Successful technological solutions will require creative, multidisciplinary methods, paradigm changing thinking, and transformative approaches. Preference will be given to research with the ability to revolutionize capabilities or demonstrate that revolutionary change is possible in the coming decade.

This BAA solicits short-term, limited scope research in topic areas that are not addressed by emerging or ongoing Government programs or other published solicitations. It is primarily, but not solely, intended for early-stage research that may lead to larger, focused programs through a separate BAA in the future.

1.2 Program Phases

This Targeted Super Seedling effort is structured as a Phase A (Base Period) with a Phase B

(Option Period). Phase A represents an initial Proof of Concept of the proposed approach. There is no requirement that the Phase A capability include a user interface or that it be useable by persons who are not programmers or system administrators. Phase B, if exercised, will build upon the Proof-of-Concept research in Phase A to deliver a demonstration capability for use by potential

Government users who may not have computer programming or system administration expertise.

Phase A shall be of a duration of 12 months to demonstrate a prototype Proof of Concept, with preliminary software deliverables and Performance Self-Evaluation Milestone reports due at months 4, 7 and 10.

BENGAL performers are expected to propose and implement their own testing and evaluation protocols. An independent testing and evaluation (T&E) team will verify performer results and validate software performance. Independent evaluation teams have not been selected for this effort, but organizations serving in this capacity may include Government agencies, Federally Funded

Research and Development Corporations (FFRDCs), University Affiliated Research Centers

(UARCs), or Department of Energy Labs. At the conclusion of Phase A, performers shall submit a Final Report. Reports and deliverables shall be used in evaluation of projects for continuation to

Phase B. Phase B will be 12 months in duration. Shorter duration projects, if appropriate for the subject matter, may be considered. See Figure 1 for a proposed project timeline.

Phase A (12 Months) 1 2 3 4 5 6 7 8 9 10 11 12

Program meeting (program kickoff, PI meeting, demos)

Gov’t visits performer site

Performer Self-Evaluation

Milestone/ T&E validation of results and system Performer deliver final report and technical products

Phase B (12 Months) 13 14 15 16 17 18 19 20 21 22 23 24

Program meeting (phase kickoff, PI meeting, demos)

Gov’t visits performer site

Performer Self-Evaluation

Milestone T&E validation of results and system Performer deliver final report and technical products

Figure 1: Proposed Phase A + B timeline with key activities.

White papers and proposals must explicitly address how the offeror’s technical approach will enable the safe adoption and use of generative AI technologies within the IC. Offerors shall demonstrate that the proposed effort has the potential to make revolutionary, rather than incremental, improvements to current capabilities. Research that primarily results in evolutionary improvement to the existing state of practice is specifically excluded.

White papers and proposals must include offeror-defined objectives, as well as milestones and performance metrics as task-driven intermediate steps towards the objectives. Offerors must clearly articulate tasks, quantitative metrics, and metric targets. Good metrics for the purpose of this effort maximize interpretability, allow easy implementation/replicability, allow calculation of confidence bounds, furnish results that are comparable over time, and are actionable (i.e., inform changes in technical approach).

Testing and evaluation of performer systems will occur 3 times during each of the 12-month phases. As noted above, successful submissions will clearly articulate a testing and evaluation protocol that can be either run on the performer site and validated by the T&E team or replicated by T&E using T&E’s own infrastructure. Metric targets should enable comparison with state-of-the-art or a well-justified baseline. Offerors must state target values for each milestone and justify why these values are challenging given the current state of the art. Proposals that do not provide explicit, feasible, and replicable protocols to measure progress will not be considered for award.

Successful projects must also contain the following elements:

• Developed capabilities must generalize across LLM text generation models and their different versions.

• Delivery of turn-key containerized software (Phase A) with user interface (UI) components and thorough documentation (Phase B). The containerized software shall be compatible with

Docker technology. An independent test and evaluation (T&E) team will affirm that software can be successfully deployed with minimal developer effort. (Ease of deployment will be one of the criteria for advancement into the second program phase.)

• Explicitly stated model access limitations and, where relevant, provide justification why a particular method (e.g., black box) cannot be used.

1.3 Description of Topic Areas

The following is a list of suggested subtopics. These are intended to provide to the offeror additional information concerning the Government’s interests in the BENGAL topic areas. These are not considered an exhaustive list and offerors are free to propose projects which address one or more topics/subtopics.

Topic #1 Biases and induction of diverse analytical perspectives:

Subtopics:

• Methods for objectively quantifying bias (e.g., relative to a specific collection of texts and other content)

• Computational techniques to characterize perspective spaces and measure the differences between perspectives.

• Using human-LLMs interactions to identify analysts’ blind spots and induce perspectives representative of those blind spots.

• Induction of outputs representing diverse perspectives (e.g., “How would a particular group or organization interpret this event?”)

• Deriving insights from simulated dialogue between LLMs with different biases/perspectives

Topic #2 AI hallucinations and inferences:

Subtopics:

• Methods to maximize the LLM’s ability to produce valuable inferences in the absence of ground-truth evidence.

• Novel and explainable approaches to quantifying confidence of generative model output

(e.g., to ensure trustworthiness for the user or enable generation of high-quality synthetic training data to reduce reliance on sensitive, sparse, or noisy data sources)

• Methods to investigate theoretical bases for LLM hallucinations or grounding (e.g., are hallucinations inevitable?)

Topic #3 Safe information flow in sensitive environments:

Subtopics:

• Targeted “unlearning” in pre-trained or fine-tuned LLMs (e.g., methods to remove from an

LLM: information about an individual or information derived from a particular document deemed sensitive without otherwise affecting the performance of the model). IARPA is not interested in filtering outputs.

• Decoupling of sensitive information: Given a description of information deemed sensitive

(e.g., source/method of collection), sanitizing a document or collection of documents such that sensitive information or ancillary information that could be used to infer sensitive information is verifiably removed, while retaining the meaning of the original document(s).

• Methods to identify when an aggregation of innocuous facts can be used to derive specific sensitive information. Given a set of queries, quantify the likelihood that the user is trying to access a particular piece of sensitive information from an LLM. Alternatively, given a set of LLM responses, quantify the likelihood that the LLM is trying to access sensitive information from the user.

Topic #4 Working resiliently with imperfect or poisoned sources:

Subtopics:

• LLM techniques to evaluate the reliability of a given information source (e.g., individuals or organizations) either for the purpose of ensuring the integrity of training data or for evaluating incoming information

• Automated and explainable techniques for inferring source intentions

• Quantifying source corroboration

• Extracting reliable intelligence from incomplete or biased content

The following topics are out of scope for this targeted super seedling effort:

• research into approaches that do not generalize across LLM text generation models and their different versions;

• research focused on systems integration or engineering of existing approaches or instruments;

• cybersecurity research not primarily focused on LLM technology; research that will not result in functional prototype technology;

• approaches requiring access to classified data; and/or

• research which are resubmissions of work already awarded by the National Science

Foundation, National Institutes of Health, Department of Defense, Intelligence

Community, or other federal agencies.

1.4 Proposal Information and Structure

The Government anticipates that proposals submitted in response to this BAA will be

UNCLASSIFIED. Proposals must address two independent and sequential project phases: Phase

A – Initial Proof-of-Concept (Base Period) and Phase B – Demonstration (Option Period). The periods of performance for these phases shall not exceed 12 months for Phase A and 12 months for Phase B. Combined Phase A and Phase B shall not exceed 24 months. Specific technical objectives to be achieved within the topic areas listed above, task descriptions, intellectual property rights, milestone schedule, and deliverables shall be addressed in the proposal. Detailed proposal preparation instructions are provided in Section 4. The total award value for the combined Phase

A (Base Period) and Phase B (Option Period) shall be less than $4,000,000.

2 Award Information

2.1 General Award Information

The BAA shall result in awards for both phases of the targeted super seedling. Exercise of the

Phase B (Option Period) shall depend upon performance during Phase A (Base Period) as well as program goals, the availability of funding, and IARPA priorities. Exercising of the Phase B option is at the sole discretion of the Government.

Multiple awards are anticipated. The amount of resources made available under this BAA shall depend on the quality of the proposals received and the availability of funds.

The Government reserves the right to select for negotiation all, some, one, or none of the proposals received in response to this solicitation and to make awards without discussions with Offerors.

The Government also reserves the right to conduct discussions if determined to be necessary.

Evaluation and award of proposals will follow FAR Part 35 processes as described herein.

Proposals selected for negotiation may result in a procurement contract.

Awards under this BAA shall be made to Offerors on the basis of the Evaluation Factors listed herein, as well as successful completion of negotiations.

This announcement constitutes the full solicitation package. This solicitation will be conducted in two steps:

STEP ONE – Submission of white papers. The Government will review white papers and encourage or not encourage submission of a full proposal.

STEP TWO – A reasonable number of Offerors apparently qualified to respond to the intent of this BAA will be encouraged to submit technical and cost proposals. Offerors not encouraged to submit or Offerors that chose not to submit a white paper are still permitted to submit a proposal, but may not be prioritized for review, see BAA Section 5. All proposals must be submitted by the due date listed in Section 1 of this BAA.

See Section 4 of this announcement for further details regarding the proposal requirements.

Upon completion of evaluation, the Government will contact Offerors whose proposals are selected for negotiations and may request additional information required for award. The

Government may establish a deadline for the close of fact-finding and negotiations that allows a reasonable time for the award of a contract. Offerors that are not responsive to Government deadlines established and communicated with the request may be removed from award consideration. Offerors may also be removed from award consideration should the parties fail to reach agreement within a reasonable time on contract terms, conditions, and cost/price.

2.2 Multiple Submissions to the BAA

Organizations may participate as a prime or subcontractor in more than one submission to the

BAA. However, if multiple submissions to the BAA which include a common team member are selected, such common team members shall not receive duplicative funding (i.e., no one entity can be paid twice to perform the same task).

3 Eligibility Information

3.1 Eligible Applicants

All responsible sources capable of satisfying the Government's needs may submit a proposal.

Historically Black Colleges and Universities, Small Businesses, Small Disadvantaged Businesses and Minority Institutions are encouraged to submit proposals and team with others to submit proposals; however, no portion of this announcement shall be set aside for these organizations’ participation due to the impracticality of reserving discrete or severable areas for exclusive competition among these entities. Other Government Agencies, Federally Funded Research and

Development Centers, University Affiliated Research Centers, Government-Owned, Contractor-

Operated facilities, Government Military Academies, and any other similar type of organization that has a special relationship with the Government, that gives them access to privileged and/or proprietary information or access to Government equipment or real property, are not eligible to submit proposals under this BAA or participate as team members under proposals submitted by eligible entities. An entity of which only a portion has been designated as a UARC may be eligible to submit a proposal or participate as a team member subject to an organizational conflict of interest review.

Offerors can request use of a unique facility, specialized equipment or technical service that is not otherwise obtainable and the Government will determine if the resource can be made available to all Offerors as Government Furnished Property/Equipment/Information Service. If the resource requested cannot be provided directly by the Government, the Government may consider an

Offeror’s request for limited use as a procured service not otherwise available, and only after an

OCI review and determination. It is advised that the Offeror have an alternate plan in its proposal in case the Government does not accept the proposed participation. Requests for such resources can be submitted during the BENGAL BAA Q&A period.

3.1.1 U.S. Academic Institutions

According to Executive Order 12333, as amended, paragraph 2.7, “Elements of the Intelligence

Community are authorized to enter into contracts or arrangements for the provision of goods or services with private companies or institutions in the United States and need not reveal the sponsorship of such contracts or arrangements for authorized intelligence purposes. Contracts or arrangements with academic institutions may be undertaken only with the consent of appropriate officials of the institution.”

Offerors shall submit a completed and signed Academic Institution Acknowledgment Letter for each U.S. academic institution that is a part of their team, whether the academic institution is serving in the role of a prime, or a subcontractor or a consultant at any tier of their team with their proposal. Each Letter must be signed by a senior official from the institution (e.g., President, Chancellor, Provost, or other appropriately designated official). A template of the Academic

Institution Acknowledgment Letter is enclosed in Appendix A of this BAA. Note: It is highly recommended that this letter(s) be submitted with the Offeror’s proposal. In any case, IARPA shall not enter into negotiations with an Offeror whose team includes a U.S. academic institution until all required Academic Institution Acknowledgment Letters are received.

3.1.2 Foreign Entities

Foreign entities and/or individuals may participate but only as part of a U.S. based team. The prime contractor must be a U.S. company. Foreign entities and individuals may participate as subcontractors or employees of a U.S. company however; all foreign participation must comply with any necessary Non-Disclosure Agreements, Security Regulations, Export Control Laws and other governing statutes applicable under the circumstances. Offerors are expected to ensure that the efforts of foreign participants do not either directly or indirectly compromise the laws of the

United States, nor its security interests. As such, domestic Offerors should carefully consider the roles and responsibilities of foreign participants as they pursue teaming arrangements.

3.2 Organizational Conflicts of Interest

According to FAR 2.101 “Organizational Conflict of Interest” (OCI) means that because of other activities or relationships with other persons, a person is unable or potentially unable to render impartial assistance or advice to the Government, or the person’s objectivity in performing the contract work is or might be otherwise impaired, or a person has an unfair competitive advantage.

In accordance with FAR 9.5, Offerors are required to identify and disclose all facts relevant to potential OCIs involving the Offeror’s organization and any proposed team member (subawardee, consultant). Under this Section, the Offeror is responsible for providing this disclosure with each proposal submitted pursuant to the BAA. The disclosure must include the Offeror’s, and as applicable, proposed team member’s OCI mitigation plan. The OCI mitigation plan must include a description of the actions the Offeror has taken, or intends to take, to prevent the existence of conflicting roles that might bias the Offeror’s judgment and to prevent the Offeror from having an unfair competitive advantage. The OCI mitigation plan will specifically discuss the disclosed OCI in the context of each of the OCI limitations outlined in FAR 9.505-1 through FAR 9.505-4.

IARPA generally prohibits Contractors from concurrently providing Scientific Engineering

Technical Assistance (SETA), Advisory and Assistance Services (A&AS), or similar support services and being a technical Contractor. Therefore, as part of the FAR 9.5 disclosure requirement above, address whether an Offeror or an Offeror’s team member (e.g. subawardee, consultant) is providing SETA, A&AS, or similar support (e.g., T&E services) to IARPA under: (a) a current award or subaward; or (b) a past award or subaward.

If SETA, A&AS, or similar support is or was being provided to IARPA, the proposal must include:

The name of the IARPA program or office receiving the support;

The prime contract number; and

Identification of proposed team member (subawardee, consultant) providing the support.

As part of their proposal, Offerors shall include either (a) a copy of their OCI notification including mitigation plan or (b) a written certification that neither they nor their subcontractor teammates have any potential conflicts of interest, real or perceived. A sample certification is provided in

Appendix A.

The Government will evaluate OCIs and potential OCIs to determine whether they can be avoided, neutralized or mitigated and/or whether it is in the Government’s interest to grant a waiver. The

Government will make OCI determinations, as applicable, for proposals that are otherwise selectable under the BAA Evaluation Factors.

The Government may require Offerors to provide additional information to assist the Government in evaluating OCIs and OCI mitigation plans.

If the Government determines that an Offeror failed to fully disclose an OCI; or failed to provide the affirmation of IARPA support as described above; or failed to reasonably provide additional information requested by Government to assist in evaluating the Offeror’s OCI and proposed OCI mitigation plan, the Government may reject the proposal and withdraw it from consideration for award.

4 Application and Submission Information

STEP ONE – White Paper Submission: Interested Offerors are encouraged to submit white papers. White papers will be accepted until December 5, 2023, 12:00 PM EST.

The Government anticipates that Offerors will receive a response to their white paper within 15 days of submission stating whether the Government encourages a proposal or does not encourage a proposal. The purpose of the white paper is to encourage proposals from a reasonable number of apparently qualified sources with technical approaches/solutions of interest to the Government.

STEP TWO – Proposal Submission: Upon review of the white papers, the Government anticipates encouraging a reasonable number of apparently qualified offerors, whose proposals are of interest to the Government, to submit full proposals (Volume I, Technical and Management

Proposal, and Volume 2 Cost/Price Proposal). In order to receive consideration for award, compliant proposals should be received by the proposal due date in Section 1 of the BAA.

Proposals received after this date will be considered late and may not be reviewed. If there are any subsequent rounds of selection, the BAA will be amended to notify Offerors and to provide the proposal due date for the next round of selections. Selection for award remains contingent on the technical and funding availability evaluation factors.

The Government intends to use Booz Allen Hamilton, Navstar Inc., SERCO, Inc., Airlin

Technologies Inc., ALKU, Advanced Technology Consulting, Site Works, TekMasters, TENICA

Global Solutions, the Infusement Group, Bluemont Technology and Research, Crimson Phoenix, MBO Partners, Inc. to provide expert advice, regarding portions of the white papers and proposals submitted to the Government and to provide logistical support in carrying out the evaluation process.

All Government and Contractor personnel shall have signed and be subject to the terms and conditions of Non-Disclosure Agreements. By submission of its white paper and/or proposal, an

Offeror agrees that its white paper and/or proposal information may be disclosed to employees of these organizations for the limited purposes stated above. Offerors who object to this arrangement shall provide clear notice of their objection as part of their submittal as indicated in the white paper/proposal preparation instructions. If Offerors do not include a notice of objection to this arrangement, the Government shall assume consent to the use of contractor support personnel in assisting the review of submittal(s) under this BAA. Only Government personnel will make evaluation and award determinations under this BAA.

All administrative correspondence and questions regarding this solicitation shall be directed by email to dni-IARPA-BAA-24-03@iarpa.gov. White papers and proposals shall be submitted in accordance with the procedures stated in the BAA.

4.1 White Paper Preparation Instructions

The white papers shall not exceed 3-pages summarizing Offeror qualifications and the Offeror’s intended technical approach/solution to the BAA Topics. The Offeror’s response to the first

Heilmeier question (“What are you trying to do?”) should be written in plain language, with minimal to no technical jargon. This statement should clearly articulate what the current state of the art is and how the proposed project, if successful, would advance beyond current capabilities.

White papers must concisely answer all the following:

Qualifications:

1. Summarize your organization’s qualifications to perform research and development in this specific field of science and technology.

2. Provide a short description of present and past performance of similar work and/or professional stature and reputation in this specific field of science and technology.

Heilmeier questions (Address in relation to the technical approach/solution for your intended proposal):

1. What are you trying to do?

2. How is it done at present? Who does it? What are the limitations of present approaches?

3. What is new about your approach? Why do you think that you can be successful at this time?

4. If you succeed, what difference will it make?

5. How will you evaluate progress during and at the conclusion of the effort? (i.e., what are your proposed milestones and metrics?)

The white paper shall not describe management nor detailed cost/price information. All white papers shall be written in English. Additionally, text should be black and paper size 8-1/2 by 11-inch, white in color with 1” margins from paper edge to text or graphic on all sides. Submissions should also use Times New Roman font (or equivalent) with font size not smaller than 12 point.

Additionally, the font size for figures, tables and charts should not be smaller than 10 point. All contents shall be clearly legible with the unaided eye or the white paper may not be considered.

White papers shall be submitted in a PFD format.

The Government anticipates white papers submitted under this BAA will be UNCLASSIFIED.

All white papers shall be in the format given below. The Government reserves the right to reject a white paper without review if the information requested below is not adequately addressed.

4.1.1 White Paper Structure

- Cover Sheet - Offerors will be prompted by IARPA’s electronic proposal submittal system, IDEAS (see BAA Section 4.3), to complete a white paper cover sheet. It will also prompt Offerors to insert a cost/price. In this case, please include the proposal limit stated in BAA paragraph 1.2, $3,999,999, as a placeholder as cost/price will not be assessed as part of the white paper review.

Offeror’s will also be prompted to indicate whether they have any objections to non-Government personnel reviewing their white paper (see BAA Section 4). This system generated cover sheet is not included in the white paper page count. Note: In addition to the system generated cover sheets, proposals require additional cover sheets for the Technical/Management and Cost

Volumes, which are included in BAA Appendix A. These additional cover sheets are not required for the white paper.

- Qualifications

- Heilmeier questions (Indicate the proposed Topic(s), see BAA Section 1.3)

White Paper not to exceed 3 pages.

4.2 Proposal Preparation Instructions

All proposals shall be in the format given below:

Proposals shall consist of Volume 1 - Technical and Management Proposal and Volume 2 –

Cost/Price Proposal. All proposals shall be written in English. Additionally, text should be black and paper size 8-1/2 by 11-inch, white in color with 1” margins from paper edge to text or graphic on all sides. Submissions should also use Times New Roman font (or equivalent) with font size not smaller than 12 point. Additionally, the font size for figures, tables and charts should not be smaller than 10 point. All contents shall be clearly legible with the unaided eye or the proposal may not be considered. Proposals shall be submitted in PDF version.

The Government anticipates proposals submitted under this BAA will be UNCLASSIFIED.

Proposals shall be valid for 120 days unless the Offeror proposes a shorter validity period.

The Government reserves the right to reject a proposal without review if the information requested below is not adequately addressed.

Each proposal submitted in response to this BAA shall consist of the following:

4.2.1 Volume 1 – Technical & Management Proposal

- Cover Sheet (not included in page count)

- Transmittal Letter (limited to 1 page, not included in page count)

- Proposal (not to exceed 15 pages)

- Attachments (the following attachments are not included in the page count)

1 – Academic Institution Acknowledgment Letter, if required (See para 3.1.1 and Appendix

A)

2 – Intellectual Property Rights (estimated not to exceed 2 pages) (See Appendix A)

3 – OCI Notification or Certification (See para 3.2 and Appendix A)

4 – Bibliography

5 - Human Use Documentation (See Section 6). Only applicable if Human Subject Research is anticipated to be part of the technical approach/plan. Claims for exemptions from Federal regulations for human subject protection must be accompanied by a strong defense of the claims.

6 - Privacy Plan – A plan that comprehensively describes the efforts the Offeror will take to protect personally identifiable information (PII) and United States Persons (USP) information obtained for performance of any contract resulting from this BAA, and safeguard the security of any personal data collected and that of any devices, applications, networks, or services involved in collection, transmission, processing, and storage of such data. This plan shall include but is not limited to: 1) All performers shall provide to the

Government annual privacy training certificates of completion for all personnel on this effort to document and instill the importance of data protections, 2) creation of a data protection plan that addresses the performer’s safeguarding approach i.e, for encrypting the data at rest and in motion for any PII and/or USP data, and 3) Notifying the IARPA Program

Manager and Contracting Officer when there is any data spill or leak. Any claims that data are anonymous must be based on evidence and supported with sufficient information regarding how the data have been anonymized.

7-Quad Chart Summary of Proposal (See Appendix A)

The Cover Sheet template is included in Appendix A. Complete all sections and include as the proposal cover. The Cover Sheet is not included in the page count.

The Transmittal Letter shall include the following (not to exceed one page):

Introduction of Offeror and team (subcontractors and consultants), the BAA number, Offeror’s

Program name, the proposal validity period, the type of contract vehicle being requested

(Cost/Cost-Plus-Fixed-Fee/FFP/Cost-Sharing procurement contract) with a short rationale, any non-negotiable conditions on which the offer is based (such as contract type, Intellectual Property

(IP ) restrictions, etc.), any restrictions from review (e.g., Government Eyes Only or restricted from certain entities, see paragraph 4) and the Offeror’s points of contact information including: name, email and phone number for both technical and administrative issues.

The Proposal shall include the mandatory elements specified in sections A. through C. below.

A. Proposal Overview

• A technical overview of the proposed research and plan. Effectively and succinctly convey the main objective, key innovations, expected impact, and other unique aspects of the proposed research project. The overview must include a paragraph on the relevance of the proposed research to the Intelligence Community mission, as well as a realistic timeframe for implementation of results. This shall include a description of the key technical challenges, a concise review of the technologies proposed to overcome these challenges and achieve the project’s goal, and a clear statement of the novelty and uniqueness of the proposed work are required. This section shall address in detail the following questions:

What is the proposed work attempting to accomplish or do? How is it done today, and what are the limitations? Who or what will be affected and what will be the impact if the work is successful?

• Summary of the products, transferable technology and deliverables associated with the proposed research results. Describe measurable deliverables that show progress toward achieving proposed milestones and goals. (All proprietary claims to the results, prototypes, IP, or systems supporting and/or necessary for the use of the research, results, and/or demonstration shall be detailed in proposal Attachment 2, Should no proprietary claims be identified in Attachment 2, the Government shall receive Unlimited Rights, as defined in

FAR 52.227-14, to all technology and deliverables resulting from or delivered under this

BAA.)

• Schedule and milestones for the proposed research. Summarize, in table form the schedule and milestones for the proposed research. Do not include proprietary information with the milestone chart. (The milestone chart may become part of the resultant contract.)

• Related research. Include a brief summary of other research in this area, comparing the significance and plausibility of the proposed innovations against competitive approaches to achieve proposed objectives.

• Project contributors. Include a clearly defined organizational chart of all anticipated project participants and affiliations (e.g. subcontractor, consultant), organized under functional roles for the effort, along with the associated task number responsibilities for each. Provide a summary of expertise of the proposed team, including any subawardees/consultants and key personnel who will be executing the work. Identify a principal investigator (PI) for the project. Demonstrate that the proposed Key Personnel/Principal Investigator are sufficiently dedicated to the project.

• Facilities. Describe the facilities and resources that will be used for the proposed effort, including computational and experimental resources.

• Resource Share. Include the type of support, if any, the Offeror might request from the

Government, such as facilities, equipment, materials, or any such resources the Offeror is willing to provide at no additional cost to the Government to support the research effort.

(Cost-sharing is not required from Offerors and is not an evaluation criterion but is encouraged where there is a reasonable probability of a potential commercial application related to the proposed research and development effort).

• The names of other federal, state or local agencies or other parties receiving the proposal and/or funding the proposed effort. If none, so state. Concurrent submission of the proposal to other organizations will not prejudice its review but may impact IARPA’s decision to fund the effort.

B. Statement of Work (SOW)

This section shall provide a detailed, clearly defined plan for the technical tasks/subtasks to be performed, by phase, their durations and the dependencies among them. For each task/subtask, provide:

• A general description of the objective;

• A detailed description of the approach to be taken, developed in an orderly progression and in enough detail to establish the feasibility of accomplishing the goals of the task;

• Identification of the primary organization responsible for task execution (prime, sub-contractor, team member, etc.) by name;

• The exit criteria for each task/activity (i.e., a product, event or milestone that defines its completion); and

• Identification of all deliverables (e.g. reports, software) to be provided to the

Government.

Note: Do not include any proprietary information in the SOW (The SOW will be incorporated into any resultant contract).

C. Technical Proposal Attachments (Not included in page count):

• Attachment 1: Academic Institution Acknowledgment Letter (see BAA Section 3.1.1. and sample letter in Appendix A).

• Attachment 2: Intellectual Property and Data Rights Assertion (estimated not to exceed 2 pages, see template in Appendix A).

• Attachment 3: OCI notification including mitigation plan or Certification stating no OCI

(see BAA Section 3.2 and sample certification letter in Appendix A.)

• Attachment 4: Bibliography. A brief bibliography of relevant technical papers and research notes (published and unpublished) which document the technical ideas on which the proposal is based.

• Attachment 5: Human Use documentation (if applicable)

• Attachment 6: Privacy Plan (if applicable)

• Attachment 7: Quad Chart Summary of Proposal

4.2.2 Volume 2: Cost/Price Proposal (No page Limit)

IARPA anticipates awarding Cost type procurement contracts. However, Offerors may request other than a Cost type procurement contract (e.g. FFP, cost-share).

Regardless of the type of contract, the Offeror’s cost/price proposal shall contain sufficient supporting information to establish the Offeror’s understanding of the project, the perception of project risks, the ability to organize and perform the work and to support the realism and reasonableness of the proposed cost/price, to the extent appropriate.

Offerors shall provide the detailed cost supporting information addressed below and in the Volume

2 Cost Element Spreadsheet, Appendix A.

Offerors may submit alternative cost/price supporting information or information in a different format; however, this will be subject to a CO determination of acceptability. If alternative information and formatting are not found acceptable, the CO will request the Offeror provide appropriate cost supporting information during negotiations. Examples where alternative cost/price supporting information and formatting may be found acceptable are when submitted by non-traditional contractors such as commercial entities that do not typically accept FAR-based contracts, small businesses, start-up companies or foreign companies.

The Cost/Price Volume shall include the following:

A. Cost Element Breakdown and Total Cost Summary

Offerors shall submit an Excel document, in the format provided in Appendix A. It shall include intact formulas and shall not be hard numbered. The base and option period cost data should roll up into a total cost summary. The Excel files may be write-protected but shall not be password protected.

• Completed cost element breakdown for the base period, option period and the total project summary in the format provided in Appendix A.

• Total costs broken down by major task.

B. Narrative Supporting Information

In addition to the above, supporting cost and pricing information shall be provided in sufficient detail to substantiate the Offeror’s cost estimates. Include a description of the basis of estimate

(BOE) in a narrative for each cost element and provide supporting documentation, as applicable:

Direct Labor –Describe the basis of the proposed labor categories and rates and provide a copy of the most recent Forward Pricing Rate Agreement (FPRA) with the Government. If Offerors do not have a current FPRA with the Government, provide payroll records or contingency hire letters with salary data to support each proposed labor category, including those for key individuals, and the most recent Forward Pricing Rate Proposal Submission, if applicable. Offeror should also address whether any portion of their labor rates is attributable to uncompensated overtime.

Labor Escalation Factor – State the proposed escalation rate and the basis for that rate (e.g., based upon Global Insight indices, Cost Index or historical data). If the escalation rate is based upon historical data, provide data to demonstrate the labor escalation trend. Provide a sample calculation demonstrating application of the factor to direct labor.

Subcontracts (to include consultants and Inter-organizational Transfers (IOTs) – The Offeror is responsible for compiling and providing full subcontractor proposals with the Cost Volume.

Subcontractor cost element sheets shall be completed for the base period, option period and the total summary using the same format required for the prime contractor (See Appendix A).

Consultant letter(s) of commitment shall also be attached.

Information shall be presented in Excel with intact formulas using the format provided in

Appendix A. The Offeror shall also provide justification for why the subcontractor was selected and its determination that the cost/price is fair and reasonable (Reference FAR Part 44 and FAR clause 52.244-2). If subcontractors have concerns about proprietary cost information, subcontractors can submit their detailed cost proposal information directly to the CO during negotiations.

Materials and Equipment – Provide copies of quotes, bill of materials, historical data or any other information including Offeror’s analysis to support proposed costs.

Travel - The proposed travel supporting detail shall include destination and purpose of the trip, number of trips, number of travelers and days per trip and price per traveler in sufficient detail to verify the BOE. Limited travel is anticipated. Offerors may require travel to meet with team members. Offerors will not be required to travel to meet with IARPA. Proposed travel costs shall comply with the limitations set forth in FAR Part 31.

Conference travel will not be authorized under this contract.

Other Direct Costs (ODCs) – ODCs shall be listed separately and supported by quotes, historical data or any other information including the Offeror’s analysis.

Indirect Costs – The Offeror shall show indirect cost calculations, identify the proposed indirect rate by fiscal year and period (base, option) and provide information on indirect cost pools and allocation bases for each year and program period involved. If a Government agency recently audited the Offeror’s indirect rates, the Offeror shall identify the agency that conducted the audit, when the rates were approved and the period for which they are effective. Include a copy of this rate agreement. Absent current Government rate recommendations, it is incumbent on the Offeror to provide some other means of demonstrating indirect rate realism (e.g., 3 years of historical actual costs with applicable pools and bases).