Big Data 2026

Keynote: Bootstrapping Coding Agents: The Specification Is the Program, Martin Monperrus, KTH Royal Institute of Technology

A coding agent can bootstrap itself. This reproduces, in the domain of AI coding agents, the classical bootstrap sequence known from compiler construction, and instantiates the meta-circular property known from Lisp. The result carries a practical implication: the specification, not the implementation, is the stable artifact of record. In cutting-edge agentic coding, the implementation is, in principle, regenerable at any time.

Martin Monperrus is Professor of Software Technology at KTH Royal Institute of Technology. Martin Monperrus is widely recognized as a pioneer of AI-assisted coding. His 2009 paper "Learning from Examples to Improve Code Completion Systems", co-authored with Bruch and Mezini, received the ACM SIGSOFT Impact Paper Award (2024) "for pioneering the use of machine learning to assist developers in generating code snippets, paving the way for today's intelligent code assistance tools." Monperrus is IEEE Fellow 2026 "for pioneering machine learning use in code assistance and program repair, impacting millions of developers." 

Automating Legacy Code Translation with Multi-Agent LLM Pipelines, Tibo Bruneel, Softwerk, Linnaeus University

Legacy codebases written in languages like Delphi and COBOL continue to power critical industrial systems, but maintaining them is increasingly costly. In this talk, I present how we use orchestrated multi-agent LLM pipelines to automatically translate millions of lines of Delphi into C# in an ongoing industrial project. I cover the practical challenges of applying LLMs at this scale, including partitioning code into translatable chunks, validating and debugging LLM outputs through specialised agents, and reassembling a working codebase on the other end. Early results show a 10x acceleration over manual translation.

Tibo Bruneel is Head of AI and an industrial PhD student at Softwerk AB in Växjö, Sweden, where his doctoral research is part of the DIA graduate school at Linnaeus University. His research specializes in LLM-based automatic code translation and MLOps, bridging academic research with industrial application. At Softwerk, he leads the company's AI division, building applied AI and large language model solutions for enterprise clients across both the public and private sectors.

Software on Demand: Embracing the Era of Autoevolving Systems, Miroslaw Staron, Chalmers, University of Gothenburg

For decades, software engineering has been a human-intensive discipline, with software evolution treated as a costly chore — a necessity we tolerated rather than enjoyed. The rise of agentic AI triggers a fundamental paradigm shift: the transition to Software Engineering 4.0 (SE 4.0), where software systems continuously and autonomously generate, evaluate, and integrate new implementations of themselves. SE 4.0 represents the most important leap in developer productivity since the compiler, unlocking a future where software evolves at the speed of developer's thought. This keynote introduces the vision of AutoEvolve and Software Engineering 4.0, a new paradigm where a system’s design space remains open-ended and mutable throughout its lifecycle. We will explore how software can adapt to changing hardware, environments, and user intents in real-time. Finally, we will examine how engineering roles, architectures, and processes must adapt to govern and collaborate with these self-improving systems.

Miroslaw Staron is a Professor of Software Engineering at Chalmers | University of Gothenburg, Sweden. He previously served as the Deputy Dean of the IT Faculty and the Chairman of the AI Coordination Team at the university. An active leader in the European AI ecosystem, he has served as a board member for AI Sweden and AI Competence for Sweden. His research sits at the intersection of software architecture, agentic systems, and embedded software, supported by over 300 peer-reviewed scientific publications and a long-term role as a Visiting Scientist at Ericsson. He is the Editor-in-Chief of the Information and Software Technology journal and heads the Software Engineering for Embedded and Automotive Systems (SEAS) group.

Improving national forest attribute maps with machine learning, Dag Björnberg, Softwerk, Linnaeus University

Remote sensing techniques are widely used for mapping and monitoring forest attributes, providing valuable information on forest cover, biomass, and overall forest health. In recent years, national airborne laser scanning (ALS) campaigns have been conducted in several countries to map forest resources. Combining ALS with field inventory data enables the development of nationwide models for important forest attributes, such as tree height and stem volume. In this talk, I will present our latest findings on using machine learning techniques to improve Swedish forest maps and discuss how the methods can be transferred and adapted to forest mapping efforts in other countries. 

Dag Björnberg is an industrial PhD student at Softwerk AB and a member of the industry graduate school on data intensive applications (DIA) at Linnaeus University. His work spans several forestry-related machine learning applications, including log traceability using computer vision, large-scale forest attribute mapping, and tree volume estimation. 

Speaker: Interactive 3D visualization of Forest Point-clouds, Benjamin Powley, Linnaeus University

This talk presents ForestVis, an interactive web-based 3D visualization tool developed as part of a collaborative research project. Forest data can be used to evaluate the health, structural, and physiological attributes of trees, and its visualization can inform forest management practices as well as research on ecology and biodiversity. Analysing forest data involves combining information from multiple sources and reasoning about multivariate data across different spatial scales. This talk includes an introduction to forest visualization, techniques for visualizing forest data using point clouds, and methods for processing forest point cloud data for web-based analysis.

Benjamin Powley is a postdoctoral researcher in the Department of Computer Science and Media Technology at Linnaeus University in Växjö, Sweden. His research interests include the visualization of geospatial and environmental data, human-computer interaction, virtual reality, immersive analytics, and the automatic optimization of visual designs. His work applies user-centred design methods to real-world environmental and geospatial data visualization problems. 

Keynote: From Contacts to Contagion: Visualizing Infection Spread on Temporal Networks, Jean R. Ponciano, University of São Paulo 

How does an infection spread in an environment such as a school, a hospital, or a private company? Who infected whom, and when? Answering these questions requires not only epidemiological models but also effective methods for representing and exploring complex, time-evolving interaction data. In this talk, I will present how temporal network visualization can go beyond structural pattern identification to assist experts in extracting meaningful epidemiological insights related to transmission paths, epidemic outbreaks, and the role of key individuals. 

In particular, I will introduce and contextualize two key visualization strategies — the Massive Sequence View (MSV) and the Temporal Activity Map (TAM) — as well as DyNetVis, the interactive system that implements them. I will show how they enable analysts to understand the dynamics of infection propagation and the impacts of different simulation scenarios on spread speed and magnitude when considering, for example, social distancing compliance and infection probabilities.

Jean Ponciano is currently an assistant professor at the Institute of Mathematics and Computer Sciences (ICMC), University of São Paulo (USP), Brazil. Before becoming a professor, he was a postdoctoral fellow with ICMC/USP (2022-2024) and previously with the School of Applied Mathematics of Fundação Getulio Vargas, Brazil (2020-2022). He received his PhD (2020) in Computer Science from the Federal University of Uberlandia, Brazil, where he also worked as a faculty member for almost three years. His research interests include information visualization, visual analytics, network science, and data streams, with particular interest in interdisciplinary collaborations that aim to optimize data exploration through visualization and visual analytics, e.g., in the legal and healthcare domains.

Speaker: Alessio Arleo, TU Eindhoven

Alessio Arleo is an Assistant Professor in the Department of Mathematics and Computer Science at Eindhoven University of Technology (TU/e), the Netherlands. Previously, he was a postdoctoral researcher at the Centre for Visual Analytics Science and Technology (CVAST) at TU Wien, Austria. He earned his Bachelor’s, Master’s (cum laude), and PhD degrees from the University of Perugia, Italy. His research focuses on graph drawing and information visualization, with an emphasis on distributed algorithms for large-scale networks. Over time, he has expanded his work to visual analytics, applying his expertise across multiple domains, including social sciences, finance, and healthcare. He is the Principal Investigator of the cross-national project SANE: ViSual ANalytics for Event-based Diffusion on Networks, in collaboration with TU Wien and the University of Cologne (Germany), which focuses on advanced visualization techniques for information diffusion. In collaboration with the University of Newcastle (UK), he is also developing visualization methods for dynamic networks without timeslices, including an algorithm adopted by the Scottish COVID-19 response team. Currently, he is actively contributing to visualization research in healthcare, serving as guest editor for special issues in international journals and as general and paper chair for workshops in the field.

Speaker: Visualization of Epidemics in Sweden, Claudio Linhares, Linnaeus University

This presentation introduces an interactive visualization approach for studying the spread and societal impacts of epidemics in Sweden by integrating spatial, temporal, demographic, and network-based data across diverse urban and rural contexts, enabling analysis of infection dynamics, mobility, healthcare accessibility, and structural inequalities, while also supporting scenario-based exploration of interventions such as vaccination, testing, and mobility restrictions to inform more equitable and evidence-based public health decision-making.

Claudio Linhares is an Associate Professor at Linnaeus University, in the Department of Computer Science and Media Technology, Faculty of Technology, in Växjö, Sweden. My research interests include information visualization, network visualization, human-computer interaction, visual analytics, and human-in-the-loop AI, with applications especially in forestry and healthcare.

Speaker: Learning Latent Epidemic Dynamics from Cross-Sectional Diagnostic Data, Tin New Aye and Sebastian Zeng, Linnaeus University

PCR cycle threshold (Ct) values, widely collected during infectious disease outbreaks, contain indirect information about infection age. We develop a framework linking cross-sectional Ct distributions to epidemic dynamics through an infection-age representation of the infected population. Renewal dynamics connect this latent state to incidence, while viral kinetics map infection age to observable Ct values, yielding a state-space model. We study identifiability to determine which features of epidemic trajectories can be inferred from Ct data, highlighting both the opportunities and fundamental limitations of epidemic monitoring based on routinely collected diagnostic measurements.

Tin Nwe Aye is a lecturer with a research background in mathematical biology and epidemiology. Her work focuses on the development and analysis of mathematical models for understanding the dynamics of infectious diseases and biological systems. She is particularly interested in the use of differential equations and stochastic modeling to study disease transmission and related biological processes, with the aim of improving theoretical understanding of epidemic behavior and control.

Sebastian Zeng received his Ph.D. in Computer Science from the University of Salzburg, Austria, in 2025. He is currently a postdoctoral researcher in the Department of Mathematics and Physics at Linnaeus University, Växjö, Sweden. His research is situated at the intersection of mathematics and machine learning, with a focus on learning stochastic and dynamical systems from large-scale real-world data.