Spain

Data-driven requirements engineering is becoming increasingly widespread in the development of today&#39;s software systems, services and apps. The exploitation of data coming from the user through several sources may indeed become an extremely useful input to requirements elicitation and management, but it does not come for free. Techniques such as NLP and ML are difficult to master and require high-quality data, whilst their generalization remains a challenge. Also, understanding the consequences into the companies&#39; development practices is still an open issue. In this keynote, I summarize the main concepts behind data-driven requirements engineering, then I provide an overview of the state of the art, recapitulate lessons learned and open challenges, and outline future research areas especially related to the impact of this approach into the full software development process.

ENASE 2020

Data-Driven Requirements Engineering: The Way Ahead

**BRACIS** originated from the combination of the two most important scientific events in Brazil in Artificial Intelligence (AI) and Computational Intelligence (CI): the Brazilian Symposium on Artificial Intelligence – **SBIA** (22 editions), and the Brazilian Symposium on Neural Networks – **SBRN** (13 editions). **ENIAC** is a national forum for researchers, practitioners, educators and students to present and discuss innovations, trends, experiences, developments and work in progress in the fields of Artificial and Computational Intelligence. **STIL** is the bi-annual Language Technology event.
All events are supported by the Brazilian Computer Society – Special Interest Groups: **CE-IA**, **CE-IC** and **CE-PLN**.

To access the 10th Brazilian Conference on Intelligent Systems (**BRACIS**), you need to register for the conference.  Please visit the Conference official site to register

[Register now](http://c4ai.inova.usp.br/bracis2021/)

Register now

BRACIS 2021

BRACIS originated from the combination of the two most important scientific events in Brazil in Artificial Intelligence (AI) and Computational Intelligence (CI): the Brazilian Symposium on Artificial Intelligence – SBIA (22 editions), and the Brazilian Symposium on Neural Networks – SBRN (13 editions). 
ENIAC is a national forum for researchers, practitioners, educators and students to present and discuss innovations, trends, experiences, developments and work in progress in the fields of Artificial and Computational Intelligence. 
STIL is the bi-annual Language Technology event.
All events are supported by the Brazilian Computer Society – Special Interest Groups: CE-IA, CE-IC and CE-PLN

Data-driven requirements engineering is becoming increasingly widespread in the development of today's software systems, services and apps. The exploitation of data coming from the user through several sources may indeed become an extremely useful input to requirements elicitation and management, but it does not come for free. Techniques such as NLP and ML are difficult to master and require high-quality data, whilst their generalization remains a challenge. Also, understanding the consequences into the companies' development practices is still an open issue. In this keynote, I summarize the main concepts behind data-driven requirements engineering, then I provide an overview of the state of the art, recapitulate lessons learned and open challenges, and outline future research areas especially related to the impact of this approach into the full software development process.

keynote

The mission of ENASE (Evaluation of Novel Approaches to Software Engineering) is to be a prime international forum to discuss and publish research findings and IT industry experiences with relation to novel approaches to software engineering.

Online consumers are constantly seeking experiences, such as vacations, restaurant outings and exciting jobs in order to improve their well-being. However, e-commerce search engines only support searches for experiences to a very limited extent -- you can search on the objective attributes of a service (e.g., hotel price and location), but the experiential aspects are buried in online reviews. E-commerce sites make some effort to surface comments from reviews, but users can still not specify experiential aspects (e.g., romantic hotel in a quiet Mediterranean town) in their queries. There has been considerable work in the NLP community to recognize and extract subjective text, but that’s only the first step towards querying.
To address this challenge, we introduce OpineDB, a subjective database. OpineDB is based on a data model that carefully balances the richness and bottom-up nature of natural language and the top-down design principles of databases. OpineDB is able to answer queries that combine multiple subjective conditions and aggregate subjective data. Unlike a traditional database system, there may not be a 1-1 mapping between query terms and the database schema. In some cases, OpineDB needs to find the closest attribute (or combination of attributes) that answers a user query, and in some cases it may have to fall back to retrieval directly from the review text.
Joint work with Yuliang Li, Jinfeng Li, Vivian Li, Aaron Feng, Saran Mumick and Wang-Chiew Tan from Megagon Labs.

Subjective Databases

Similarities are inherent in software - Who has not adapted existing code to speed up writing new programs? While simplistic and not be very effective in long run, copy-paste-modify is a common reuse practice. It produces software clones - recurring in variant forms similar code fragments, classes, source files or even bigger program modules. Software cloning phenomenon has been investigated by researchers for decades. Whether clones are good or bad depends on the context. Sometimes software clones hinder program understanding and maintenance, and are considered a sign and measure of decaying software structure and quality. Such software clones should be avoided or eliminated, whenever possible. In other situations, software clones are created intentionally and play some useful role in a program. Application of company standards and design patterns leads to clones, but we do not question the value of these practices because of that. In the talk, I will analyse the multifaceted phenomenon of software similarities, particularly focusing on the situations when software clones - not necessarily good - explode because we can’t contain their explosion with conventional programming techniques.
Software systems can comprise 10’s of millions LOC (WINDOWS is well over 100 millions LOC), with thousands of inter-related components, reaching the limits of what today’s technology can handle. We’ll be surely challenged by even larger and more complex systems-of-systems of the future. How do we cope with such systems if their complexity grows proportionally to their size? Especially software maintenance, which is almost exclusively done at the level of code, exposes developers to such complexity. Not surprisingly, up to 80% of software costs go to maintenance. I will present a view that software similarities have yet unexploited potential to help us reduce software complexity. But to tap on that potential we must reach beyond the current software reuse paradigm.

Premium content

Data-Driven Requirements Engineering: The Way Ahead

Downloads

Next from ENASE 2020

Subjective Databases

Similar lecture

Classificacão multimodal de produtos que não podem ser vendidos no marketplace

Stay up to date with the latest Underline news!

PRESENTATIONS

CONFERENCES

COMPANY

RESOURCES