Guaranteeing AI robustness against deception (gard) - develop a new generation of defenses against deception attacks on machine learning (ML).
Ended Definitive Contract Awarded Last action
- Contract number
- HR001120C0015
- Awardee
-
Massachusetts Institute Of Technology UEI E2NYLCDML6V1
- Awarding agency
-
Defense Advanced Research Projects Agency
Department of Defense
- Obligated
- $6,775,001
Award facts
- Opportunity
-
- Guaranteeing AI Robustness against Deception (GARD) Pre-Solicitation · HR001119S0026 · Posted Feb 12, 2019
- Guaranteed Robust Artificial InteLligence (GRAIL) Award Notice · Posted Feb 4, 2020
- Contract value 100% obligated
- $6,775,001.00 Current and potential, with all options
- Completion date 100% complete
- Current and potential, with all options
- Contracting staff
- NAICS code
- 541715 Research and Development in the Physical, Engineering, and Life Sciences (except Nanotechnology and Biotechnology)
- Set-aside
- No set-aside
- Competition
- Full and Open Competition · 48 offers Basic Research
- Pricing
- Cost No Fee
- Place of performance
- Cambridge, MA 02139, USA
- Legislative mandates
-
- Employment Eligibility Verification (E-Verify) FAR subpart 22.18
Award details come from SAM.gov/FPDS.
Transactions
| Action | Obligated | Date |
|---|---|---|
| Funding Only Action | $572,170 | |
| Funding Only Action | $1,450,000 | |
| Funding Only Action | $350,000 | |
| Other Administrative Action | $0 | |
| Exercise an Option | $100,000 | |
| Funding Only Action | $888,884 | |
| Funding Only Action | $500,000 | |
| Funding Only Action | $500,000 | |
| Exercise an Option | $500,000 | |
| Other Administrative Action | $0 | |
| Other Administrative Action | $0 | |
| Original award | $1,913,947 |
Show all 12
About this award
This federal contract award, valued at $6,775,001.00, was issued by the Defense Advanced Research Projects Agency (DARPA) to the Massachusetts Institute of Technology (MIT) to develop a new generation of defenses against deception attacks on machine learning (ML) under the Guaranteeing AI Robustness against Deception (GARD) program. The contract has an ultimate completion date of August 16, 2024 and does not utilize any set-aside designations.
The original solicitation sought innovative research proposals across two technical areas: defense theories and algorithms, as well as an evaluation framework, to enable ML systems to be robust against adversary deception. The contract work will involve multiple sensor modalities, such as still images, audio, and video, with the evaluation scenarios becoming more complex over three phases. Key deliverables include theoretical findings, defense implementations, test results, and publications.
MIT has been awarded this prime contract and has subcontracted with Aarno Labs LLC and the University of California, Berkeley to support the research and development efforts. The University of California, Berkeley also holds a separate $8.92 million Other Transaction (OT) Indefinite Delivery Vehicle (IDV) with DARPA for advanced simulation capabilities using machine learning.
Subawards
| Subcontractor | Amount | Awarded |
|---|---|---|
| The Regents Of The University Of California | $116,387 |
Quick answers
Who was awarded contract HR001120C0015?
HR001120C0015 was awarded to Massachusetts Institute Of Technology by Defense Advanced Research Projects Agency (Department of Defense) on January 31, 2020.
How much is contract HR001120C0015 worth?
The government has obligated $6,775,001 on HR001120C0015.
Was contract HR001120C0015 competed?
The extent of competition for HR001120C0015 was full and open competition. The government received 48 offers.
On GovTribe
See this award's full record
- Every transaction and funding action
- The awardee's other awards and teaming partners
- Similar awards and recompete opportunities
- Ask GovTribe AI about this award