DataIngestion-RFI FINAL.pdf

PILOT IRS- Request for Information (RFI) 2032H8-20-N-00032

Internal Revenue Service (IRS), Criminal Investigation (CI) Cyber Crimes Data Ingestion Initiative

1.0 High-Level Summary/Future State

This RFI is part of the Pilot IRS Program. Internal Revenue Service, Criminal Investigation Division (CI) is requesting information about innovative data ingestion technologies. Specifically, technologies that will allow digital files with arbitrary layouts and formats to be parsed and their data ingested into a structured, user-defined ontology automatically while preserving data relationships, without user intervention. The technology must be able to handle file layouts from new sources that have not previously been identified, without modification. The technology must operate autonomously within CI's technology environment without any interaction with the vendor, and little to no interaction with the user.

For example, corporate records of financial transactions obtained through due legal process come in a virtually endless variety of formats. Possible formats include listing out and labeling rows and columns of transactional and account information, unstructured narrative text reports, and other formats. The technology needs to be able to handle this variation without prior preparation for the specific layout presented, through artificial intelligence or other means, and correlate each data element with existing data ontologies despite different word choice, labeling, formatting, or organization. The technology must also preserve any relationships between data elements being ingested, for example sequence, hierarchy, or links to other data elements. This requirement is not limited to corporate financial transaction records but could include other financial records, online transaction logs, communications, and any other files obtained in the course of a criminal investigation.

CI is the largest Federal law enforcement agency in the United States Department of the Treasury, consisting of sworn law enforcement officers (Special Agents) and support personnel focused solely on matters related to criminal violations of law. CI is responsible for investigating potential criminal violations of the U.S. tax, money laundering and bank secrecy act laws and related financial crimes.

CI is a global leader in cyber-criminal investigations involving cryptocurrency and various digital assets. CI has played a leading or pivotal role in the takedown of numerous major Dark Net Marketplaces, virtual currency exchanges and other transnational criminal organizations facilitating stolen identify refund fraud (SIRF), narcotics trafficking, money laundering, terrorist financing, sex trafficking, and child prostitution.

We are primary interested in: 1) a self-service tool hosted entirely on CI computers; 2) algorithms and self-contained code capable of implementing the above requirement; and 3) a rough estimate of cost relative to return on investment (ROI). We are not interested in solutions that require sensitive data to be sent outside of CI or that rely on an external service-based model.

2.0 Background/Current State

CI Special Agents routinely encounter a wide variety of data file types, formats, and layouts in the course of conducting criminal investigations. We have automated pathways for handling certain commonly used data files from specific companies, but this requires us to manually construct a new automation pathway for each new file format or import the data manually. For one-off file formats or spreadsheet layouts that have not been encountered before and are not likely to be encountered in volume again, it is generally not worth the resource investment to create a dedicated automation pipeline. The data ontologies we use for analysis change and evolve over time to meet evolving mission requirements. When this occurs, these data pathways must also be modified to adapt to the ontology changes and existing data must be remapped. In addition, many of the data extract, transform, load technologies generally commercially available do not preserve the relationships between extract data items (i.e. hierarchy within a file, sequence on a spreadsheet, or links to related data in other parts of the file).

Investigative data is highly sensitive and subject to strict legal, regulatory policy and privacy requirements.

3.0 Specific Areas of Input Requested/Questions

CI respectfully requests input from industry partners and other parties regarding the following inquiries/topics:

1. Provide summary background on any existing instances of data ingestion algorithms, code, and tools capable of automated ingestion of arbitrary unknown unstructured and semi-structured data into a user-defined ontology. What sort of file types would be supported? Are there any identified limitations to supportable file types?

2. Describe the user and data load supported and any dependencies in these calculations.

3. Are there known technologies that will make this effort easier if utilized? i.e. GPUs, OCR software, etc.

4. Provide recommendations regarding our primary areas of focus and recommendations for implementation.

5. Provide any overarching recommended approaches for automated data ingestion and mapping in this space.

6. Identify where a Rough Order of Magnitude (ROM) for contract cost to support this initiative would fall, with any notations for development, hosting, training and operations and maintenance.

7. Provide potential ROI summary information that CI could reasonably expect to attain with the adoption and implementation of selected technologies in this space.

4.0 Disclaimer

This RFI is issued solely for information and planning purposes. This notice should not be construed as a commitment by the Government for any purpose other than market research. This announcement does not commit the Government to any contractual agreement. The Government is not seeking proposals and will not accept unsolicited proposals. No reimbursement will be made for any costs associated with providing information in response to this announcement or any follow-up information requests.

Respondents will NOT be notified of the results of the analysis. All data received in response to this Sources Sought synopsis that is marked or designated as corporate or proprietary information will be fully protected from release outside the Government. The Government shall not be liable for or suffer any consequential damages for any proprietary information not properly identified. Proprietary information will be safeguarded in accordance with all applicable Government regulations. All documentation shall become the property of the Government and will not be returned.

The deadline for responses to this RFI is 08:00AM EST, July 14, 2020. Responses should be forwarded via email to both the primary and secondary points of contact below. Responses shall not exceed 5 pages and shall not contain any brochures, advertising or any other type of extraneous, graphic literature or documents that have not been requested and are not relevant or essential in demonstrating the company’s ability to provide the required services. Vendors may provide links to relevant websites, as appropriate.

Vendors shall reference “RFI_Data_Ingestion” in the subject line of their response as well as in the subject line of any other e-mail correspondence referencing this notice.

** THE GOVERNMENT WILL NOT RESPOND TO PHONE CALLS. **

Thank you for your interest.

Contracting Office Address:

IRS, 1111 Constitution Ave. NW Washington, District of Columbia 20024

Primary Point of Contact:

Marcy Almeida Marcela.a.almeida@irs.gov 202-317-4087

Secondary Point of Contact:

Cyber Crimes Program Manager pilotirscicyber@ci.irs.gov mailto:Marcela.a.almeida@irs.gov mailto:pilotirscicyber@ci.irs.gov