In February 2019, the Software Institute started a Seminar Series. Every Thursday afternoon, a researcher of the Institute will publicly give a short talk on a software engineering argument of her choice. Examples include, but are not limited to, novel interesting papers, seminal papers, personal research, preliminary research ideas, tutorials, and small experiments.

Below you can find more details on the next seminar, the upcoming seminars, and an archive of the past speakers.

Everyone is welcome to attend the seminars organized by the Software Institute.

Next Speaker: Mohammad Rezaalipour

Date: September 17, 2020 @ 16:30
Room: SI-003
Deep Neural Network Bugs and the Challenges of Repairing Them

Deep Neural Networks (DNNs) have become quite popular these days. Many software developers are employing DNNs to add learning capabilities to their software products. As a result, Software Engineering (SE) for DNNs has become quite important. Generally, the presence of bugs in software products is a significant SE problem. On the other hand, the nature of bugs in DNN based software products might be different from those appearing in traditional software products.
Therefore, the first step towards addressing this problem (i.e., the presence of bugs) is to understand the characteristics of bugs in DNN based software products. What kinds of bugs are more frequent in these software products? What are the root causes of such bugs? What is the impact of such bugs? Which stages of a deep learning pipeline are more prone to bugs? Answering these questions helps researchers better understand the nature of bugs in DNN based software products. The second step can be to understand the challenges of repairing DNN bugs.
In this talk, I will review two recent papers about bugs in DNN based software products. The first paper presents a comprehensive study on the characteristics of DNN bugs, and the second paper continues this work by focusing on the repair patterns that can fix these bugs and the challenges of repairing DNN bugs.

Upcoming Seminars

Past Seminars

  • Michael Weiss - Detecting Uncertainty in Deep Learning (February 27, 2020)
  • Christoph Treude - Uncovering the best parts of software documentation (January 28, 2020)
  • Bhargav Bhatt - DroidPLUMB: Repairing Resource-Leak bugs with Static Analysis (December 5, 2019)
  • Jesper Findahl - TypeScript - what is that and why should I care? (November 28, 2019)
  • Francesco Magagnino - Envisioning the future of the customer interaction (November 21, 2019)
  • Armin Heinzl - How Pair Programming Influences Team Performance: The Role of backup-behavior, shared mental models, and task novelty (November 7, 2019)
  • Davide Paolo Tua - Time Evolving Voronoi Treemaps for Metrics Visualization (October 31, 2019)
  • Bin Lin - Program Comprehension at ICSME 2019 (October 24, 2019)
  • Ana Ivanchikj - Discovering Imgur API – Controlled Experiment (October 17, 2019)
  • Marco D'Ambros - Dashboarding your inbox for fun and profit (October 3, 2019)
  • Emad Aghajani - Software Documentation: How far we've come, and challenges ahead (September 26, 2019)
  • Andrea Stocco - Black-box Confidence Estimation for Misbehavior Prediction in Autonomous Driving Systems (September 19, 2019)
  • Jacopo Tagliabue - Less (Data) Is More: Why Small Data Holds the Key to the Future of Artificial Intelligence (June 24, 2019)
  • David Clark - The Theory of Testing Programs - An Information Theoretic View (June 19, 2019)
  • Jan Vitek - Getting everything wrong without doing anything right! (June 13, 2019)
  • Hridesh Rajan - Software as Data (June 12, 2019)
  • Alejandro Mazuera Rozo - SOFIA: An Automated Security Oracle for Black-Box Testing of SQL-Injection Vulnerabilities (May 23, 2019)
  • Jevgenija Pantiuchina - On the Naturalness of Buggy Code (May 16, 2019)
  • Richard Torkar - Why do we encourage even more missingness when having missing data? (May 9, 2019)
  • Fengcai Wen - Neural-Machine-Translation-Based Commit Message Generation: How Far Are We? (May 2, 2019)
  • Vincenzo Riccio - A Day in the (Activity) Lifecycle (April 18, 2019)
  • - Casting about in the Dark (April 11, 2019)
  • Gunel Jahangirova - Mutation Testing of Deep Learning Systems (April 4, 2019)
  • Andrea Mocci - The Tale of 'Quattro Tabelle' (March 28, 2019)
  • Carlo Alberto Furia - Why You Should Use Bayesian Statistics for Empirical Software Engineering (March 7, 2019)
  • Csaba Nagy - Beauty and the Beast: True Stories of Evolving Software Systems (February 28, 2019)
  • Andrea Gallidabino - Liquid Software: Multi-Device Adaptation with Liquid Media Queries (February 21, 2019)