Schedule

Opening

PhD Survival Strategies

In this tutorial we will present and discuss a wide set of strategies (a.k.a. tips, tricks and secrets) to maximize your chances of succeeding both during and after your PhD. In essence, we will tell you all the things you always wanted to know but never dared to ask. This tutorial is based on our own personal experiences, which is known to be an empirically sound approach.

Gabriele Bavota , Università della Svizzera italiana, Switzerland

Gabriele Bavota received (cum laude) the Laurea in Computer Science from the University of Salerno (Italy). He received the PhD in Computer Science from the University of Salerno in 2013. From January 2013 to October 2014 he has been a research fellow at the Department of Engineering of the University of Sannio. From November 2014 to August 2016 he was Assistant Professor at Free University of Bolzano-Bozen. He joined the Faculty of Informatics of USI as an Assistant Professor in September 2016. His research interests include software maintenance and evolution, mining software repositories, and empirical software engineering. He is author of over 120 papers appeared in international journals and conferences. He is the recipient of the 2018 ACM Sigsoft Early Career Researcher Award for outstanding contributions in the area of software engineering as an early career investigator. He received four ACM Sigsoft Distinguished Paper awards at the three top software engineering conference: ASE 2013 and 2017, ESEC-FSE 2015, and ICSE 2015. He also received the best paper award at SCAM 2012 and ICSME 2018. He served as a Program Co-Chair for ICPC 2016, SCAM 2016, and SANER 2017. He also serves and has served as organizing and program committee member of international conferences in the field of software engineering, such as ICSE, ESEC/FSE, ASE, ICSME, MSR, ICPC, SANER, SCAM, and others.

Michele Lanza , Università della Svizzera italiana, Switzerland

Michele Lanza is full professor, vice-rector, and director of the Software Institute at the Università della Svizzera italiana (USI) in Lugano, Swizterland. His doctoral dissertation, completed in 2003 at the University of Bern, received the Denert award for best thesis in software engineering of 2003. Prof. Lanza received the Credit Suisse Award for best teaching in 2007 and 2009. At USI Prof. Lanza directs the Software Institute of USI, founded in 2017. He leads the REVEAL research group, working in the areas of software visualization, evolution, and analytics. He co-authored more than 200 peer-reviewed articles and the book "Object-Oriented Metrics in Practice".

Performance-driven Software Model Refactoring

Software refactoring is a common practice aimed at addressing requirements or fixing bugs during the software development. While refactoring related to functional requirements has been widely studied in the last few years, non-functional-driven refactoring is still critical, mostly because non-functional characteristics of software are hard to assess, and appropriate refactoring actions can be difficult to identify. In the context of performance, antipatterns represent effective instruments to tackle this issue, because they document common mistakes leading to performance problems as well as their solutions. However, in order to effectively reuse the knowledge beyond performance antipatterns, automation is required to detect and remove them.

Goal of this tutorial is to introduce the audience to problems related to performance-driven refactoring of software models and to solutions based on the performance antipattern concept.
After shortly introducing Software Performance Engineering, the concept of performance antipattern will be defined, and existing techniques for representing, detecting and removing such antipatterns will be illustrated. A tool, based on the Epsilon platform, to support such activities will be also shown at work. The final part of this tutorial will be dedicated to the key open issues in this field, as well as to widen the scope mostly in two directions: (i) moving from models to code and code traces, (ii) moving from pure performance analysis to tradeoff analysis of multiple non-functional attributes.

Vittorio Cortellessa , Università dell'Aquila, Italy

Vittorio Cortellessa has received his M.S. degree in Computer Science from University of Salerno (Italy, 1991), and his Ph.D. degree in Computer Engineering from University of Roma "Tor Vergata" (Italy, 1995). Between 1996 and 1999 he has been post-doc fellow at the European Space Agency and at University of Roma "Tor Vergata". Since January 2000 through October 2001 he was Research Assistant Professor at CSEE, West Virginia University. In November 2001 he moved to University of L'Aquila, where he is Full Professor since March 2018. His main research interests are in the areas of Software Performance Engineering, Model-Driven Engineering, and more in general in the analysis of non-functional attributes of software systems. He has published more than 100 conference and journal papers on these topics. He has served and serves in program committees of conferences and as journal reviewer in these areas. He is currently member of the Steering Committee of ICPE (ACM/SPEC International Conference on Performance Engineering), and Editorial Board Member of ACM Transactions on Software Engineering and Methodology and of Empirical Software Engineering journal (Springer).

Natural Language Processing in Software Engineering

Natural Language Processing (NLP) is a mature research discipline, at the intersection of artificial intelligence and computational linguistics. Nowadays, NLP is regarded as crucial for software engineering, with applications including automatic summarization and generation of software documentation, requirements elicitation, and implementation of software development bots. Furthermore, NLP techniques have been used to mine the developers’ emotions and opinions from their communication channels. Applying NLP techniques in software engineering poses new challenges due to the presence of technical jargon, code fragments, URLs, and platform-specific lexical features such as hash-tags, emoticons or @-mentions. As such, traditional approaches may require SE-specific fine tuning and adaptation to be successfully exploited in this domain. In this tutorial we will use the task of sentiment analysis, i.e. the recognition of the positive, negative, or neutral semantic orientation of a text, as a working example. We will overview supervised and unsupervised approaches to text categorization, including preprocessing techniques, lexical analysis, semantics, and pragmatics. The tutorial will be combined with demo sessions in which the participants will have the opportunity to apply the learned techniques and jointly discuss their results.

Nicole Novielli , University of Bari, Italy

Nicole Novielli is an Assistant Professor at the University of Bari, where she received a Ph.D. in Computer Science in 2010. Her research interests lie at the intersection of software engineering and affective computing with a specific focus on emotion mining from software repositories and natural language processing of developers’ communication traces. In 2016, she started the workshop series on Emotion Awareness in Software Engineering co-located with the International Conference on Software Engineering (ICSE). She was the principal investigator of the three-year project 'EmoQuest - Investigating the Role of Emotions in Online Question and Answering Sites', funded by MIUR under the SIR (Scientific Independence of young Researchers) program.

Mining and Analyzing Source Code Changes

Current versioning systems track code changes on a pure textual level. While tools like “diff" allow for quickly storing and generating the textual differences, they suffer from one major problem: their line-based structure is coarse grained and does not consider syntax nor semantic information. It is therefore the task of the software developer to figure out what actually has changed in the source code and, moreover, what the impact of a code change is. This is often a difficult and time consuming task, and it is not surprising that developers still spend most of their time on understanding source code and its changes. In this seminar, I will introduce you to tree-differencing techniques and in particular to our IJM approach and tool. IJM is capable of extracting source code changes on the level of abstract syntax trees (ASTs). This way, IJM obtains detailed syntactic information about code changes that developers and researchers can leverage to faster and better understand code changes and their potential impact. Furthermore, the seminar will include a lab session in which you will do some mining of the information extracted with IJM to answer questions, such as what has changed in a class in the last release, which kinds of changes were performed most frequently, can we identify certain change patterns, etc. Of course, you can pose your own questions that you would like to answer in the lab session.

Martin Pinzger , Alpen-Adria-Universität Klagenfurt, Austria

Martin Pinzger is a full professor of software engineering and the head of the Software Engineering Research Group at the University of Klagenfurt, Austria. His research interests are in software engineering with focus on software evolution, mining software repositories, program analysis, software visualization, and automating software engineering tasks. He received a PhD (Dr. techn.) in 2005 and a MSc (Dipl. Ing.) in 2001 from the Vienna University of Technology. He was a Postdoc at the University of Zurich and a Visiting Researcher at Microsoft Research in Redmond. In 2008, he became an Assistant Professor at the Delft University of Technology from which he received tenure in 2012. He has authored and co-authored more than 85 scientific articles in renowned international journals and conference proceedings. In 2012, he won an NWO Vidi grant, one of the most prestigious Dutch individual research grants for his research proposal on recording and analysis of fine-grained code changes. In 2013, he received an ICSE 2013 ACM SIGSOFT distinguished paper award and the ICSM 2013 most influential paper award. In 2014, he won the ICSME 2014 Best Paper Award. Martin Pinzger is a senior member of IEEE and ACM.

Deep Learning for Software Engineering

Bridging the abstraction gap between concepts and source code is the essence of software engineering (SE). SE researchers regularly use machine learning to bridge this gap, but there are two fundamental issues with traditional applications of machine learning in SE research. Traditional applications are too reliant on human intuition, and they are not capable of learning expressive yet efficient internal representations. Ultimately, SE research needs approaches that can automatically learn representations of massive, heterogeneous, datasets in situ, apply the learned features to a particular task and possibly transfer knowledge from task to task.
Improvements in both computational power and the amount of memory in modern computer architectures have enabled new approaches to canonical machine learning tasks. Specifically, these architectural advances have enabled machines that are capable of learning deep, compositional representations of massive data depots. This rise of deep learning has led to tremendous advances in several fields. Given the complexity of software repositories, deep learning has the potential to usher in new analytical frameworks and methodologies for SE research and the practical applications it reaches. Conversely, the development of deep learning algorithms and models represents an entirely new type of software engineering that is still evolving.
This tutorial examines how deep learning algorithms can enhance or automate several critical SE tasks, including program generation and program repair among others. We demonstrate that deep learners significantly outperform state-of-practice canonical machine learning approaches for these tasks. These examples illustrate the tremendous potential that deep learning can have on the science and practice of software engineering by moving SE research from the art of feature engineering to the science of automated discovery. Additionally, we explore how advancements in software engineering tools and practices can enable further progress in making deep learning frameworks more accessible and useful for researchers, programmers, and data scientists.

Denys Poshyvanyk , The College of William and Mary, USA

Denys Poshyvanyk is a Class of 1953 Term Distinguished Associate Professor in the Computer Science Department at William and Mary. His current research is in the area of software engineering with specific focus on understanding how machine learning can improve the research and practice of software development and maintenance. His papers received 12 Best and/or Distinguished Paper Awards. He also received two Most Influential (“Test of Time”) Paper Awards at from ICSME'06 and ICPC'07. He is a recipient of the NSF CAREER award (2013). Denys Poshyvanyk currently serves on the editorial board of IEEE Transactions on Software Engineering (TSE), Empirical Software Engineering Journal (EMSE, Springer) and Journal of Software: Evolution and Process (JSEP, Wiley). He served on the steering committee of the IEEE International Conference on Software Maintenance and Evolution (ICSME) and IEEE International Conference on Program Comprehension (ICPC). He also served as a PC Co-chair for five software engineering conferences. He is also a present member of the CRA-E committee.

September 3

Opening

Mining and Analyzing Source Code Changes

Coffee Break

Mining and Analyzing Source Code Changes

Lunch

Students' Talks

Coffee Break

Guided Tour

September 4

PhD Survival Strategies

Coffee Break

PhD Survival Strategies

Lunch

Apero

September 5

Deep Learning for Software Engineering

Coffee Break

Deep Learning for Software Engineering

Lunch

Performance-driven Software Model Refactoring

Coffee Break

Performance-driven Software Model Refactoring

Social Dinner

September 6

Natural Language Processing in Software Engineering

Coffee Break

Natural Language Processing in Software Engineering

Lunch