Our model-driven engineering research develops intelligent approaches for creating, understanding,
assessing, and evolving software models.
Our uncertainty-aware software engineering research develops methods for understanding, specifying,
modeling, evolving, and testing uncertainties in software and cyber-physical systems.
Selected work spanning model-driven engineering and uncertainty-aware software engineering.
Man Zhang, Yunyang Li, and Tao Yue
This systematic literature review analyzes 235 primary studies covering 25 types of model-driven engineering tasks, with model generation as the dominant focus. It examines LLM techniques, benchmarking and evaluation practices, and the availability of tools and datasets. The findings show that UML, DSLs, and BPMN are the most frequently targeted formalisms. GPT-family models, zero-shot prompting, API-based access, and single-LLM architectures prevail, with fine-tuning and retrieval-augmented generation as common enhancements. Evaluations focus mainly on correctness, with limited realistic benchmarking and insufficiently comprehensive replication materials.
VenueACM Transactions on Software Engineering and Methodology
Man Zhang, Yunyang Li, and Tao Yue
This work introduces PSUM-SysMLv2, an extension that incorporates the OMG Precise Semantics for
Uncertainty Modeling metamodel into SysML v2. Seven case studies examine its completeness,
expressiveness, and applicability across SysML v2 elements and application domains; the paper also provides
an application methodology and two realization approaches.
Venue
Software and Systems Modeling
Man Zhang, Tao Yue, Nazareno M. Aguirre, Diego Garbervetsky, and Sebastian Uchitel
The paper presents TrustModel, a vision for agentic generation and evolution of living knowledge models.
Its Modeling, Conformance, and Evolution subsystems construct and update models, assess alignment with
systems and environments, and generate evolution guidance, with model-based testing illustrating the
approach.
VenuearXiv · Preprint
Man Zhang, Tao Yue, and Andrea Arcuri
The paper motivates uncertainty-driven, system-level fuzzing for microservices and identifies challenges
in modeling, injecting, propagating, and analyzing uncertainty across services. It outlines a continuous
testing architecture combining service virtualization, uncertainty simulation, adaptive test generation,
optimization, and causal reasoning for fault localization.
VenueACM Transactions on Software Engineering and Methodology
Man Zhang, Chongyang Shen, Andrea Arcuri, and Tao Yue
This study examines flakiness in REST API fuzzing through empirical investigations of 36 APIs and an
analysis of nearly 3,000 failing tests. Based on the identified sources of instability, it introduces
FlakyCatch to detect and mitigate flaky behavior in tests produced by white-box and black-box fuzzers.
VenuearXiv · Preprint
Yifan Wang, Tiexin Wang, and Tao Yue
This work investigates how uncertainty originating in autonomous-driving sensor data affects downstream
deep-learning models. It focuses on tracing uncertainty from model inputs to learning-component behavior,
supporting more dependable perception and uncertainty-aware evaluation of autonomous driving systems.
VenueInformation and Software Technology
Pablo Valle, Aitor Arrieta, Liping Han, Shaukat Ali, and Tao Yue
This work proposes a domain-specific language for defining multi-level, uncertainty-aware CPS test
oracles and an automated generator that packages them as a DevOps-compatible microservice. Evaluation on
two industrial systems and nine open-source CPSs assesses requirements coverage and practical generation
time.
VenueSoftware and Systems Modeling