TRUST/UFRPE investigates methods for designing, evaluating, and operating computing and AI-enabled systems so that they behave as intended and remain dependable throughout their lifecycle. Our research encompasses software and AI engineering; dependability, performance, and system modeling; operational resilience; and the optimization and deployment of AI models in resource-constrained environments.
Our research addresses trustworthiness at three complementary stages of the system lifecycle.
01
We investigate software and AI engineering practices, including quality assurance, testing, maturity models, software engineering for AI (SE4AI), and AI for software engineering (AI4SE), to reduce faults during design and development.
02
We use modeling, simulation, and experimentation to assess dependability, performance, energy, cost, and resource needs, and to guide the optimization and deployment of AI models in environments with limited resources.
03
We study software aging and operational degradation and develop monitoring, anomaly detection, and predictive maintenance techniques to strengthen operational resilience and support adaptation and recovery.
The group investigates these challenges across AI systems and LLM-based applications, cloud platforms, edge environments with limited computational resources, databases, industrial equipment and manufacturing systems, UAVs, IoT, robotic telesurgery, and other cyber-physical systems and critical infrastructures.