ACM SAC 2027 · Technical Track
AI-enabled Resilience for
Autonomous Systems (AIRAS)
Keeping AI-driven autonomous systems resilient to faults and attacks
Autonomous systems (self-driving vehicles, aerial drones, cognitive robots, and industrial control platforms) are increasingly powered by AI and machine learning (ML) components that handle perception, planning, and control. This shift brings real capability, but also a new kind of fragility: AI/ML models can be fooled, manipulated, or pushed outside their training distribution in ways that classical fault-tolerance and security techniques were never designed to catch. As these systems take on more responsibility in transportation, healthcare, agriculture, and industrial settings, resilience failures carry consequences well beyond the system itself.
AIRAS (AI-enabled Resilience for Autonomous Systems) is a new track at ACM SAC dedicated to this challenge. We bring together researchers and practitioners working at the intersection of AI, ML, systems security, control theory, and dependable computing, anyone building, breaking, or hardening the AI/ML that autonomous systems rely on. Our scope spans adversarial robustness, anomaly detection, explainable AI, federated learning, and fault recovery, across autonomous vehicles, robotics, drones, and industrial cyber-physical systems. We invite original research, experience reports, survey and tutorials, and emerging ideas from academia and industry.
Topics of Interest
We invite original contributions on the security and
resilience of AI-driven autonomous systems broadly. Topics include, but are not limited to:
- Adversarial robustness and defenses across perception, planning, and control pipelines
- AI-driven anomaly/intrusion detection, and adaptive threat response
- Explainable AI (XAI) for resilience assurance and safety verification
- Attack and defense mechanisms for learning-enabled real-time systems
- Resilient federated and privacy-preserving learning for distributed autonomous systems
- Reinforcement learning for adaptive and self-healing autonomous behavior
- Runtime assurance and uncertainty-aware monitoring
- Foundation and Large Language Models for cyber-physical and autonomous systems resiliency
- AI-enabled fault diagnosis, prognosis, and system recovery
- Trust and assurance frameworks for AI-driven autonomous systems
- Hardware/software co-design for resilient AI in embedded and edge platforms
- Resilience in autonomous vehicles, drones, robots, space, and industrial systems
- Secure deployment and integrity assurance of AI models
- AI security for safety-critical and resource-constrained autonomous platforms
- Data-centric ML resilience (poisoning detection, noise robustness, and dataset provenance)
- Multi-agent coordination and resilience under adversarial conditions
- Integration of quantum computing and quantum machine learning for autonomous systems
- Digital twin and simulation-based resilience analysis
- Testbeds, benchmarks, and evaluation methodologies for AI resilience
Submission Guidelines
Research Papers
AIRAS welcomes original submissions of up to 8 pages (two extra pages with an additional page charge for $80 per page) of technical content (including references, appendices, and acknowledgments). The author name(s) and address(es) must not appear in the body of the paper, and self-references should be in the third person. This is to facilitate a double-blind review. Only the title should be shown on the first page without the author information. Papers must be formatted according to the ACM SAC template. For full template and submission guidance, please see the ACM SAC 2027 website. Papers that have been concurrently submitted to other conferences or journals (double submissions) will be automatically rejected.
Submissions may take one of the following forms:
- Research papers: Original technical contributions that introduce new methods, systems, analyses, or evaluations for AI-enabled resilience in autonomous systems. Suitable papers may present new defenses, monitoring and assurance mechanisms, learning-enabled recovery methods, resilient system architectures, or principled analyses of attacks, failures, and necessary trade-offs.
- Empirical / Reproduction and Replication (R&R) papers: Empirical studies, experience reports, datasets, benchmarks, testbeds, systematization of knowledge, and/or reproduction and replication studies that confirm, question, clarify, or extend prior results in an AIRAS-relevant area. Such papers should go beyond simply re-running an artifact and should explain (a) what was reproduced or replicated, (b) what changed in the setting or assumptions, (c) what was learned, and (d) how the findings affect the design or evaluation of resilient autonomous systems. Strong submissions should identify technical gaps or inconsistencies in assumptions and synthesize evidence across communities that may reshape resilience in autonomous systems.
- Application / Technology Transfer / Industry Experience papers: Manuscripts describing innovative applications of AI-enabled resilience techniques in real-world autonomous systems across science, engineering, or business domains; accounts of technology successfully transferred from research into new problem domains or operational settings; and industry experience reports or demonstrations of novel, deployed systems. Such papers should articulate the problem context, the approach taken, and lessons learned, with attention to the industrial experience and practical constraints (e.g., deployment, integration, certification).
Student Research Competition (SRC)
Graduate students seeking feedback from the scientific community on their research ideas are invited to submit abstracts of their original unpublished and in-progress research work. Authors of selected abstracts will have the opportunity to share and discuss their research work through poster and oral presentations and compete for the three top winning places as selected by the SRC committee. The winners will receive cash awards and SIGAPP recognition certificates. Furthermore, invited authors are eligible to apply for the SIGAPP Student Travel Award Program (STAP) for support. SRC abstracts are limited to 4 pages and submitted via SAC-SRC 2027 webpage.
Please visit ACM SRC program and SAC SRC program for more information.
Submission Portal
- Paper submissions: TBD
- Student Research Competition (SRC): TBD
Important Dates
- Paper Submission Deadline: October 2, 2026
- Notification to Authors: TBD
- Camera-ready Due: TBD
- Author Registration Deadline: TBD
All deadlines will be announced in accordance with SAC 2027 guidelines.
Technical Program Committee
TBD
Web Chair
- Tamim Ahmed, Washington State University