No matter what machine the software engineers build, the requirements are located in the environment" [Jackson'97]. This environment is part of the real world in which the machine is installed and the machine's effect is observed and evaluated. The re-emergence of AI (especially the black-box deep learning solutions) and the unstoppable penetration of AI-based systems across industries, public sectors, and all walks of life make it important and timely for the requirements engineering (RE) community to discuss the role of environment in driving various activities: elicitation, modeling, implementation, testing, deployment, and evolution. With the machine becoming more intelligent and embedded, the environment is more open and dynamic. This year's workshop objectives are to bring the interested researchers and practitioners together, exchange ideas, discuss ongoing work, and work together as a community to explore the capabilities and limitations of ChatGPT as a requirements modeling helper. In particular, the workshop participants will work in small groups on an agreed set of RE problems (scenarios), and interact with ChatGPT to build a requirements model in problem frames, goal models, UMLs, etc. Throughout this working session, we will share our experience as a community in the kinds of good and bad questions that ChatGPT responds, the resulting model's quality or the lack thereof, etc. [top]
EnviRE'23 will feature a keynote, paper presentations, and a participant working session.
We are excited to announce that Professor Didar Zowghi will deliver a keynote on, "Diversity and Inclusion: The Bedrock for Responsible AI". Didar is one of the most prominent researchers in RE. Join us to learn her thoughts and discuss with her in the days of AI.
Here is a tentative program; all times are local, Hannover, Germany times.
9:00am-9:15am Welcome and participant introduction [Chair: Organizers]
9:15am-10:15am Keynote by Professor Didar Zowghi on ""Diversity and Inclusion: The Bedrock for Responsible AI"" [Chair: Zhi] (Zoom link: https://ucincinnati.zoom.us/j/96299169440)
10:15am-10:30am Keynote Q&As [Chair: Zhi]
10:30am-11:00am Coffee Break
11:00am-12:30pm Paper presentations (about 20 minutes per paper, including Q&As) [Chair: Nan]
Requirements Modeling Aided by ChatGPT: An Experience in Embedded Systems by Kun Ruan, Xiaohong Chen, and Zhi Jin
Augmenting the Problem Frames Approach with Explicit Data Descriptions: Prioritizing Requirements in Conjunction with ChatGPT by Ling Xie, Hongbin Xiao, and Zhi Li
A Model Checking based Software Requirements Specification Approach for Embedded Systems by Xiao Yang and Xiaohong Chen
Environmental Variations of Software Features: A Logical Test Cases' Perspective by Md Rayhan Amin, Tanmay Bhowmik, Nan Niu, and Juha Savolainen
12:30am-2:00pm Lunch
2:00pm-3:30pm Paper presentations (about 20 minutes per paper, including Q&As) and working session setup [Chair: TBA]
Automating Extraction of Problem Diagrams from Natural Language Requirements Document by Dongming Jin, Chunhui Wang, and Zhi Jin
A multimedia-approach to problem descriptions for fine-grained detail characterization by Bowen Zheng, Zhi Li, and Hongbin Xiao
PF-HCPS: Extending Problem Frames for Supporting Human-Cyber-Physical System Collaboration by Yilong Yang, Bingjie Zeng, Zhiching Chen, and Juntao Gao
Working session topics identification, grouping, prioritization, and selection
3:30pm-4:00pm Coffee Break
4:00pm-5:30pm Working session, reporting, and workshop summary
[Download the one-page Call for Papers in PDF format: here.]
Modeling the environment will be more and more important in RE when the systems will situate in the open world and with the human in the loop. For example, IoT-enabled systems, cyber-physical systems, AI-based systems, etc. are expected to be able to perceive the changes of an open and dynamic environment, respond to changes through architectural transformations, and exhibit context-aware, adaptive, and trustworthy behaivors.
Specifically for the AI-based systems, the components built by machine learning in fact are black boxes. It is not possible to structuring their functions by examining their architectures (consisting of a hierarchical neural networks). Their functions can only be represented by the effects imposed on their operational and interacting environment. These effects can in turn define the tasks of model training, validation, testing, deployment, and operation. When mapping the requirements into the environment properties or assertions, the benefits include natural decomposition and structuring of the problem. The "Environment-Driven Requirements Engineering" workshop thus solicits position, short, and full papers to provoke the discussions about:
Other emerging topics are encouraged and welcomed. We expect each position or short paper to be up to 4 pages long, whereas a long paper to be up to 8 pages long. However, if you need more or less space, feel free to contact the EnviRE'23 organizers.
This year, the EnviRE workshop will organize a working session to use ChatGPT to elicit and model the requirements for a specific problem. The problem scenario is adopted from the paper, and is given as follows:
Transport for London (TfL) is considering the deployment of drones to deliver medical assets such as blood and organs between hospitals. In the past, blue-light vehicles were used by the London Ambulance Service to deal with such medical emergencies. Blood and organs need to be delivered efficiently and reliably, which is not always viable in a traffic-congested mega city. Therefore, motor-bikes are often used instead. Cost aside, motor-bikes are noisy and one of the causes of traffic accidents. The use of drones for this purpose has therefore been proposed. However, when drone-related incidents happen, they must be analysed quickly, including checks for their compliance to the regulations of the airspace they navigated.The workshop participants will work in small groups to figure out requirements for such a system with the help of ChatGPT. Sample tasks include: working with ChatGPT to build a goal model, a problem frames model, a UML model, etc. for the problem scenario. Throughout the working session, we will share our experience as a community in the kinds of good and bad questions that ChatGPT responds, the resulting model's quality or the lack thereof, etc. It is our intention that EnviRE as a community can begin establishing an exemplar such as the TfL scenario that will facilitate knowledge sharing and reuse. Interested authors are encouraged to apply their approaches to the TfL scenario listed above, and/or share their insights of using ChatGPT to perform requirements elicitation and modeling tasks. [top]
Please refer to the RE'23 page for formatting instructions. and submit your EnviRE'23 papers to: https://easychair.org/conferences/?conf=envire2023 . [top]
All deadlines are 23:59 Anywhere on Earth (Standard Time). [top]
Mounifah Alenazi, University of Hafr Albatin, Saudi Arabia
Carina Alves, Universidade Federal de Pernambuco, Brazil
Tanmay Bhowmik, Mississippi State University, USA
Xiaohong Chen, East China Normal University, China
Fuyuki Ishikawa, National Institute of Informatics, Japan
Eunsuk Kang, Carnegie Mellon University, USA
Emmanuel Letier, University College London, UK
Tong Li, Beijing University of Technology, China
Zhi Li, Guangxi Normal University, China
Gunter Mussbacher, McGill University, Canada
Kenji Tei, Waseda University, Japan
Tao Yue, Simula Research Laboratory, Norway
Jianzhang Zhang, Hangzhou Normal University, China