The GEODES Seminar aims to bring together members of the group in a semi-formal setting with a semi-regular timing.
We'd especially love to see students at all levels reporting their progress, entertain presentations on core or adjacent topics, host dry-runs before conferences, etc.
These events follow an in-person format. By that, we hope to re-establish the everyday social aspects of the research group in a post-COVID world.

  • Format: In person
  • Location: Pavillon Andre-Aisenstadt

Upcoming talks

  • Photo of Mouna Dhaouadi
    End-to-End Rationale Reconstruction

    The logic behind design decisions, called design rationale, is very valuable. In the past, researchers have tried to automatically extract and exploit this information, but prior techniques are only applicable to specific contexts and there is insufficient progress on an end-to-end rationale information extraction pipeline. Here we outline a path towards such a pipeline that leverages several Machine Learning (ML) and Natural Language Processing (NLP) techniques. Our proposed context-independent approach, called Kantara, produces a knowledge graph representation of decisions and of their rationales, which considers their historical evolution and traceability. We also propose inconsistency checking mechanisms to ensure the correctness of the extracted information and the coherence of the development process. We conducted a preliminary evaluation of our proposed approach on a small example sourced from the Linux Kernel, which shows promising results.

    2022-10-07@14h00 PA3195, Pav. André-Aisenstadt
  • Photo of Istvan David
    Blended Modeling in Commercial and Open-source Model-Driven Software Engineering Tools: A Systematic Study

    This talk reports on the results of a systematic multivocal study on the emerging topic of blended modeling. Blended modeling aims to improve the user experience of modeling activities by prioritizing the seamless interaction with models through multiple notations over the consistency of the models. To understand the potential of current commercial and open-source modeling tools to support blended modeling, we have reviewed nearly 5,000 academic papers and nearly 1,500 entries of grey literature. We have identified 133 candidate tools, and eventually selected 26 of them to represent the current spectrum of modeling tools. This work has been carried out in collaboration with Malvina Latifaj (Mälardalen University), Jakob Pietron (Universität Ulm), Weixing Zhang (Chalmers University of Technology, University of Gothenburg), Federico Ciccozzi (Mälardalen University), Ivano Malavolta (VU Amsterdam), Alexander Raschke (Universität Ulm), Jan-Philipp Steghöfer (Chalmers University of Technology, University of Gothenburg), and Regina Hebig (Chalmers University of Technology, University of Gothenburg).

    2022-10-14@14h00 PA3195, Pav. André-Aisenstadt

Previous talks

  • Migrating Enterprise applications into Software Product Lines

    Software Product Lines represent one of the most exciting paradigms shift in software development in the last two decades. The main objective is to improve reuse by focusing not only on the development of a single software product but on a family of related product variants. Over the years, multiple research approaches have been proposed to address the different activities of variability design and manipulation, reusable assets implementation, or product derivation. However, implementing SPL variability is still a major challenge and represents a risk for a company. Indeed, compared to single-system development, SPL variability management implies a methodology that highly impacts the life cycle of the products as well as the processes and roles inside the company. To break these barriers to SPL adoption in the industry, we propose a new software platform that assists developers in migrating existing applications into SPL. This tool is built with a language-independent engine to support any programming language. In this talk, we present the research foundations related to the implementation of tour platform. The talk will be guided with live demos on real applications.

    2022-08-26@14h00 PA3195, Pav. André-Aisenstadt
  • Uncertainty in Software Models [presentation]

    When designing and operating with software and systems, stakeholders and modelers often face situations in which they need to deal with uncertain data and information. These uncertainties may originate from different reasons such as vagueness, imprecision, incompleteness or inconsistencies. In many cases, neglecting these uncertainties may have disastrous consequences. In this talk, we will explain the different types of uncertainty that can appear when dealing with software models, paying particular attention to measurement uncertainty and belief uncertainty. We will explain an approach to capture, operate and propagate these uncertainties as well as a strategy and methodology to make informed decisions when different stakeholders are involved.

    2022-07-22@14h00 PA3195, Pav. André-Aisenstadt
  • Learning from Code Repositories to Recommend Model Classes

    With the growing popularity of machine learning algorithms, dramatic advances have been made for code completion, and specifically method-call completion. These advances were also possible thanks to the availability of large code repositories to learn from and to the well-defined boundaries of the method-call completion problem. This is, however, not the case for design completion, where model repositories are scarce and the space of possibilities for design completion is theoretically infinite. We propose in this presentation an approach that learns numeric representations of domain concepts and their relations from code repositories in order to recommend classes for UML class diagrams.

    2022-07-15@14h00 PA3195, Pav. André-Aisenstadt
  • Process modeling, analysis and enactment with the FTG+PM formalism [presentation]

    Originally developed for the sound composition and analysis of model transformation chains, the Formalism Transformation Graph + Process Model (FTG+PM) formalism provides versatile means to reason about the properties of complex engineering processes. This talk introduces the formalism in detail, samples from the numerous engineering cases the formalism has been successfully applied in before, and opens up a discussion about a related challenge in our research on digital twins for cyber-biophysical systems.

    2022-07-08@14h00 PA3195, Pav. André-Aisenstadt
  • Systematic Reviews with ReLiS [presentation]

    Systematic reviews are an important endeavor that shows the maturity of a discipline, research theme, or topic. In this talk, we will understand what systematic reviews are and discuss how they are conducted. As for tool support, we present ReLiS, a tool to automatically install and configure systematic reviews projects to conduct them collaboratively and iteratively on the cloud. We will also reflect on lessons learned from conducting systematic reviews in software engineering. This seminar ends with an open discussion on conducting a systematic review on the use of reinforcement learning in software engineering.

    2022-06-30@10h00 PA3195, Pav. André-Aisenstadt
  • Fine-grained Analysis of Similar Code Snippets

    Code recommendation approaches such as code search or code repair may rely on code snippets or code templates extracted from existing projects to provide these suggestions. In this context, extracting and characterizing reusable and recurring code structures beforehand is thus essential. In this work, we characterize recurring code structures through parametrizable code templates. Code templates can outline the common structure in code snippets along with their variation points, hence providing a convenient way to define their structural similarity. Pattern Structure is a mathematical data analysis framework for organizing objects depending on their similarity: it produces a structure supporting clustering, analysis, and knowledge discovery tasks. We propose an approach leveraging this framework and similarity defined through code templates to highlight and organize groups of similar snippets.

    2022-06-17@13h00 PA3195, Pav. André-Aisenstadt
  • The ongoing problem of reuse and evolution in software architectures: illustration through four research projects

    These last years, my research work has been mainly oriented towards the needs of companies and industries, looking for social and economic impacts. I will briefly introduce four different research issues and adopted approaches to address them: 1) Microservice migration, 2) Asynchronous business-event choreography, 3) Business-IT Alignment and 4) Eco-design in shipbuilding. The objective of my presentation is not to detail each work but to give you an overview of my current research interests, and, ultimately, to open the discussion with you about a question that has been bothering me for a long time: is it possible to move from craft to industrialization in Software engineering?

    2022-06-10@13h00 PA3195, Pav. André-Aisenstadt
  • Moose – A modular, composable software analysis environment

    Advanced reverse engineering tools are required to cope with the complexity of software systems and the specific requirements of different tasks (re-architecturing, migration, evolution). Consequently, reverse engineering tools should adapt to a wide range of situations. Yet, because they require a large infrastructure investment, being able to reuse these tools is key. Moose is a reverse engineering environment answering these requirements. While Moose started as a research project 20 years ago, it is also used in industrial projects, exposing itself to all these difficulties. In this talk we present the new version of Moose. It revolves around a composable meta-model; interacting micro-tools; and an open architecture supporting the synchronization and interaction of tools per task.

    2022-06-03@13h00 PA3195, Pav. André-Aisenstadt
  • Introduction to Reinforcement Learning

    Reinforcement learning is a mathematical framework to infer and represent knowledge. Reinforcement learning has shown promising results in the recent years in the quest for building a general learning agent that act as human do. This seminar provides a gentle introduction to the topic of reinforcement learning, introduces some of the recent advancements, and outlines future challenges.

    2022-05-27@13h00 PA3195, Pav. André-Aisenstadt
  • Multi-paradigm modeling of complex heterogeneous systems – Are we there yet?

    Nowadays engineered systems have reached a previously unprecedented complexity. Systems are becoming faster, more autonomous, and more safety-critical. To cope with this increasing complexity, engineering is best approached through modeling. Throughout the design, engineering, and assembly stages of a system, modeling offers many benefits. For example, the design-space of the system can be explored easily by multiple stakeholders, or properties can be formally verified at the early stages of system construction. Model-driven engineering also enables the partial synthesis of the system, eliminating many sources of human error from the overall process. Multi-paradigm modeling (MPM), specifically, advocates modeling aspects of the system using the most appropriate formalisms. The diverse nature of these formalisms allows for increased descriptive power but complicates coordination between stakeholders. Pertinent examples are cyber-physical and mechatronic systems which feature various mechanical, electrical, and software components. In such settings, a common vocabulary among stakeholders cannot be assumed. This, in turn, gives rise to inconsistencies between stakeholder models, which might result in faulty or suboptimal products. This talk introduces three lines of research addressing major challenges in MPM. First, the PROxIMA inconsistency management framework focuses on optimizing and intelligently orchestrating the overall system engineering process. Second, the Lowkey framework allows for real-time collaborative modeling between heterogeneous stakeholders. Third, a technique for the automated inference of simulation models by reinforcement learning to further automate system engineering tasks.

    2022-05-20@13h00 3195, Pav. André-Aisenstadt
  • Improving online visibility for you and your research [presentation]

    Visibility is a key success factor for researchers to increase their impact in academica. A particular challenge of academic visibility is that it requires effort over time. This talk will provide the audience with motivation and specific recommendations for increasing their academic visibility online. Specifically, online 'channels' such as ResearchGate, Google Scholar, and personal websites are discussed, along with best practices and tips for maintaining up-to-date information. This talk is for early-stage and mid-stage researchers who are interested in improving their online presence.

    2022-05-13@11h00 3195, Pav. André-Aisenstadt
  • AST-Probe: Recovering abstract syntax trees from hidden representations of pre-trained language models

    2022-05-06@14h30 3195, Pav. André-Aisenstadt


Got an interesting topic you'd like to present? Reach out to your favorite postdoc for details.