You are here

Past Seminars

19/03/2015 - 16:00

Abstract: I will present a paper that I am currently writing with Angelo for the next ER conference (deadline on March 29th), and which focuses on what expert finding techniques should consider to identify experts. This research is in a preliminary state, thus the conceptual model are far to be complete, but already provide an interesting top-down approach to help designing and evaluating expert finding techniques.

12/03/2015 - 16:00

The task of specifying requirements is not achieved by simply uttering the words "hocus pocus" but if we select the right elicitation and analysis technique the requirements become true. I will present how the elicitation technique focus group has been applied on a case study about argumentation on online discussions of an OSS community. This focus group was led by Daniele Belluta with the collaboration and supervision of Ornella Mich, Anna Perini and myself. The work was done by Daniele during his stage in our SE group. 

05/03/2015 - 16:00
This seminar is about the paper:
Aaron K. Massey, Paul N. Otto, Annie I. Antón, "Evaluating Legal Implementation Readiness Decision-making".
19/02/2015 - 16:00
This seminar is about the paper:
"Generating hard satisfiability problems" by Bart Selman, David G. Mitchellb, Hector J. Levesqueb.
12/02/2015 - 16:00

I will present the paper:

05/02/2015 - 16:00

Oracles are key artifacts that determine the effectiveness of test cases in revealing faults in the software. I will present techniques that address the problem of (1) assessing the quality of oracles through metrics; (2) improving the quality of oracles by detecting brittle assertions and unused inputs. The presentation is based on the following two papers:

29/01/2015 - 16:00
22/01/2015 - 16:00


Test data generation has been extensively investigated as a search problem, where the search goal is to maximize the number of covered program elements (e.g., branches). Recently, the whole suite approach, which combines the fitness functions of single branches into an aggregate, test suite-level fitness, has been demonstrated to be superior to the traditional single-target at a time approach.