Doron Drusinsky
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Doron Drusinsky has a total of 12 co-authors. Here are the top co-authors by number of shared publications:
- James Bret Michael
Naval Postgraduate School
14 shared publications - Matthew Litton
U.S. Africa Command
7 shared publications - Duminda Wijesekera
George Mason University
2 shared publications - Man-Tak Shing
Department of Computer Science, Naval Postgraduate School, Monterey, CA, USA
2 shared publications - Armon Barton
Department of Computer Science, Naval Postgraduate School, Monterey, CA, USA
1 shared publication - Christopher Bud Daniel
Naval Post-graduate School, Monterey, CA, USA
1 shared publication
Latest Publications
- 1 October 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in Computer
- Vol. 59 (10) , 123-129
An Algorithm for Search and Rescue Using Uncrewed Vehicles
- 1 October 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in Computer
- Vol. 59 (10) , 123-129
An Algorithm for Search and Rescue Using Uncrewed Vehicles
- 1 July 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in Computer
- Vol. 59 (7) , 132-145
From Shortest Path to Safest Path
- 1 July 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in Computer
- Vol. 59 (7) , 132-145
From Shortest Path to Safest Path
- 31 March 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in Computer
- Vol. 59 (4) , 145-155
Protecting AI From Training Data Contamination
With the advent of generative artificial intelligence (GAI) there is risk that training data for GAI can become contaminated, resulting in skewing of GAI output. We propose a cryptocurrency verification network to armor-plate GAI against external contamination.
- 31 March 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in Computer
- Vol. 59 (4) , 145-155
Protecting AI From Training Data Contamination
With the advent of generative artificial intelligence (GAI) there is risk that training data for GAI can become contaminated, resulting in skewing of GAI output. We propose a cryptocurrency verification network to armor-plate GAI against external contamination.
- 24 September 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in Computer
- Vol. 58 (10) , 122-130
Generative AI Regulation Is an Undecidable Problem
Using computability theory, this article demonstrates that while single-prompt generative artificial intelligence (GAI) responses can be regulated, GAI can circumvent regulation using compound multi-user conversations that cannot be regulated.
- 24 September 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in Computer
- Vol. 58 (10) , 122-130
Generative AI Regulation Is an Undecidable Problem
Using computability theory, this article demonstrates that while single-prompt generative artificial intelligence (GAI) responses can be regulated, GAI can circumvent regulation using compound multi-user conversations that cannot be regulated.
- 11 April 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Systems Journal
- Vol. 19 (2) , 413-424
Discovering Decision Manifolds to Assure Trusted Autonomous Systems
Developing and fielding complex systems requires proof that they are reliably correct with respect to their design and operating requirements. Especially for autonomous systems, which exhibit unanticipated emergent behavior, fully enumerating the range of possible correct and incorrect behaviors is intractable. Therefore, we propose an optimization-based search technique for generating high-quality, high-variance, and nontrivial data, which captures the range of correct and incorrect responses a system could exhibit. This manifold between desired and undesired behavior provides a more detailed understanding of system reliability than traditional testing or Monte Carlo simulations. After discovering data points along the manifold, we apply machine learning techniques to quantify the decision manifold's underlying mathematical function. Such models serve as correctness properties, which can be utilized to enable both verification during development and testing, as well as continuous assurance during operation, even amidst system adaptations and dynamic operating environments. This method can be applied in combination with a simulator in order to provide evidence of dependability to system designers and users, with the ultimate aim of establishing trust in the deployment of complex systems. In this proof-of-concept, we apply our method to a software-in-the-loop evaluation of an autonomous vehicle.
- 11 April 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Systems Journal
- Vol. 19 (2) , 413-424
Discovering Decision Manifolds to Assure Trusted Autonomous Systems
Developing and fielding complex systems requires proof that they are reliably correct with respect to their design and operating requirements. Especially for autonomous systems, which exhibit unanticipated emergent behavior, fully enumerating the range of possible correct and incorrect behaviors is intractable. Therefore, we propose an optimization-based search technique for generating high-quality, high-variance, and nontrivial data, which captures the range of correct and incorrect responses a system could exhibit. This manifold between desired and undesired behavior provides a more detailed understanding of system reliability than traditional testing or Monte Carlo simulations. After discovering data points along the manifold, we apply machine learning techniques to quantify the decision manifold's underlying mathematical function. Such models serve as correctness properties, which can be utilized to enable both verification during development and testing, as well as continuous assurance during operation, even amidst system adaptations and dynamic operating environments. This method can be applied in combination with a simulator in order to provide evidence of dependability to system designers and users, with the ultimate aim of establishing trust in the deployment of complex systems. In this proof-of-concept, we apply our method to a software-in-the-loop evaluation of an autonomous vehicle.
- 25 March 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in Computer
- Vol. 58 (4) , 130-146
Multiagent Aircraft Flight Route Planning in a 3D Threat-Contested Environment Using Cross-Entropy Search
We present cross-entropy search for multiple-aircraft facing inter- and intra-aircraft constraints, a problem that cannot be solved by traditional Dynamic Programming algorithms such as A*.
- 25 March 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in Computer
- Vol. 58 (4) , 130-146
Multiagent Aircraft Flight Route Planning in a 3D Threat-Contested Environment Using Cross-Entropy Search
We present cross-entropy search for multiple-aircraft facing inter- and intra-aircraft constraints, a problem that cannot be solved by traditional Dynamic Programming algorithms such as A*.
- 23 September 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in Computer
- Vol. 57 (10) , 118-130
Machine-Learned Correctness Properties, Runtime Verification, and Advance-Warning Oracles for Autonomous Systems
We describe machine-learned classifiers that act as correctness properties for runtime verification of autonomous systems. Additionally, we show an advance collision detection system using such machine learning correctness properties that provides an uncertainty measure alongside a prediction.
- 23 September 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in Computer
- Vol. 57 (10) , 118-130
Machine-Learned Correctness Properties, Runtime Verification, and Advance-Warning Oracles for Autonomous Systems
We describe machine-learned classifiers that act as correctness properties for runtime verification of autonomous systems. Additionally, we show an advance collision detection system using such machine learning correctness properties that provides an uncertainty measure alongside a prediction.
- 2 April 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in Computer
- Vol. 57 (4) , 119-130
From P Versus NP to Probabilistic and Zero Knowledge Proof Systems
I present an overview of the field of zero knowledge proof systems, which has evolved considerably over the last four decades, with applications ranging from secure authentication, secure communications, crypto-currency, and online privacy.
- 2 April 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in Computer
- Vol. 57 (4) , 119-130
From P Versus NP to Probabilistic and Zero Knowledge Proof Systems
I present an overview of the field of zero knowledge proof systems, which has evolved considerably over the last four decades, with applications ranging from secure authentication, secure communications, crypto-currency, and online privacy.
- 30 January 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Reliability Magazine
- Vol. 1 (1) , 37-46
Reliable Autonomous Vehicles: How Do We Get There?
On 10 August 2023, the California Public Utilities Commission (CPUC) voted to allow Cruise LLC to provide fared passenger services throughout San Francisco without a safety driver or time restrictions [1]. Ten years after the company was founded, Cruise seemed poised to spur the autonomous vehicle (AV) industry to turn the long-promised fleets of AVs into reality. Then, two major incidents occurred. On 19 August 2023, a Cruise vehicle collided with a San Francisco Fire Department truck that was responding to an emergency, injuring the rider inside the Cruise vehicle [2]. Two months later, a human-driven vehicle hit a pedestrian, knocking them into a Cruise AV’s path; the AV stopped near the pedestrian before restarting again, dragging her approximately 20 ft at low speed before parking on top of her leg [2]. Soon after, on 24 October 2023, the California Department of Motor Vehicles (CA DMV) indefinitely suspended Cruise’s AV deployment and driverless testing permits, stating that “the Department determines the manufacturer’s vehicles are not safe for the public’s operation” [3].
- 30 January 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Reliability Magazine
- Vol. 1 (1) , 37-46
Reliable Autonomous Vehicles: How Do We Get There?
On 10 August 2023, the California Public Utilities Commission (CPUC) voted to allow Cruise LLC to provide fared passenger services throughout San Francisco without a safety driver or time restrictions [1]. Ten years after the company was founded, Cruise seemed poised to spur the autonomous vehicle (AV) industry to turn the long-promised fleets of AVs into reality. Then, two major incidents occurred. On 19 August 2023, a Cruise vehicle collided with a San Francisco Fire Department truck that was responding to an emergency, injuring the rider inside the Cruise vehicle [2]. Two months later, a human-driven vehicle hit a pedestrian, knocking them into a Cruise AV’s path; the AV stopped near the pedestrian before restarting again, dragging her approximately 20 ft at low speed before parking on top of her leg [2]. Soon after, on 24 October 2023, the California Department of Motor Vehicles (CA DMV) indefinitely suspended Cruise’s AV deployment and driverless testing permits, stating that “the Department determines the manufacturer’s vehicles are not safe for the public’s operation” [3].
- 16 June 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in Computer
- Vol. 56 (7) , 121-130
Machine-Learned Verification and Advance Notice Oracles for Autonomous Systems
The assurance of cyberphysical systems that employ some degree of autonomy is a difficult problem in that the underlying system is artificial-intelligence-based with almost no accompanying human-written correctness specifications. This article proposes a two-pronged approach using a fast simulation-based search technique named hybrid pair cross-entropy.
- 16 June 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in Computer
- Vol. 56 (7) , 121-130
Machine-Learned Verification and Advance Notice Oracles for Autonomous Systems
The assurance of cyberphysical systems that employ some degree of autonomy is a difficult problem in that the underlying system is artificial-intelligence-based with almost no accompanying human-written correctness specifications. This article proposes a two-pronged approach using a fast simulation-based search technique named hybrid pair cross-entropy.