Research
Postdoc
During my first year as a postdoc, I worked as a research fellow at the italian National Research Center (CNR-IEIIT): my activity led to a contribution to the development of the Interomics flagship project as well as to innovative extensions of the Rulex machine learning software suite.
Rulex
The Rulex software suite is a prescriptive analytics machine learning platform, comprising both ETL tools, well-known and propertary, exclusive machine learning algorithms, such as the Logic Learning Machine. Rulex technology is available as a portfolio of products: as a standalone software equipped with a graphical user interface (GUI), as OEM modules for a quick integration into pre-existing business analytics softwares, or as APIs for application over the cloud.
My contributions to the suite, included in the official release since the 3.2 version, mainly involve the development of both the front-end and the back-end of tasks in the area of association rules mining. The back-end was developed in C++, while the GUI is Python-based. Together with the implementation of well-established techniques such as the Eclat algorithm, the development of innovative tools, especially oriented towards complexity reduction and assortment optimization was also addressed.
Interomics
The Interomics Flagship project is oriented towards the development of an integrated platform for the application of "omic" sciences to biomarker definition and theranostic, predictive and diagnostic profiles.
Reference papers:
Analyzing gene expression data for pediatric and adult cancer diagnosis using logic learning machine and standard supervised methods , Damiano Verda, Stefano Parodi, Enrico Ferrari, Marco Muselli, BMC Bioinformatics, November 2019
Logic Learning Machine and standard supervised methods for Hodgkin’s lymphoma prognosis using gene expression data and clinical variables , Stefano Parodi, Chiara Manneschi, Damiano Verda, Enrico Ferrari, Marco Muselli, Health Informatics Journal, March 2018
PhD
During the PhD, my research activity focused on robotic perception. More specifically, I worked on three projects, briefly described in the following, which make use of a laser scanner, a Microsoft Kinect and a camera. All these sensors are used in robotics, given that each of them exhibits both desirable features and downsides:
- a laser scanner ensures to get a fixed number of reliable distance measurements, but it gathers these measurements only on the scanning plane and it is limited by a maximum distance range,
- a camera is not affected by any of these restrictions, but the perceived image requires computation in order to be associated to a (variable) amount of recognizable features,
- the Microsoft Kinect is strongly limited by its maximum scanning range (4m) and cannot be reliably used outdoor (given its sensitivity to lighting conditions), but it is a low-cost sensor providing 3D data.
Micro Air Vehicle (MAV) position tracking
During the period I spent at the Robotics and Perception Group at the University of Zurich (led by prof. Davide Scaramuzza) I contributed to the development of a purely vision-based position tracking technique for Micro Air Vehicles (MAVs).
Reference papers:
Air-ground Matching: Appearance-based GPS-denied Urban Localization of Micro Aerial Vehicles, Andràs L. Majdik, Damiano Verda, Yves Albers-Schoenberg, Davide Scaramuzza, Journal of Field Robotics (JFR), October 2015
Micro Air Vehicle Localization and Position Tracking from Textured 3D Cadastral Models, Andràs L. Majdik, Damiano Verda, Yves Albers-Schoenberg, Davide Scaramuzza, International Conference on Robotics and Automation (ICRA) 2014
Structure-based object category recognition
A Microsoft Kinect sensor is used as an input for an object category recognition application, the Kinect is mounted on a ground robot.
Reference paper: Structure-based object representation and classification in mobile robotics through a Microsoft Kinect, Antonio Sgorbissa, Damiano Verda Robotics And Autonomous Systems (RAS), December 2013
Human localization and mapping
A laser scanner is mounted on a helmet (together with a 6-DOF Inertial Measurement Unit, IMU) to build a map-generation and self-localization application based on wearable sensors.
Reference paper: Human navigation and mapping with a 6DOF IMU and a laser scanner, Marco Baglietto, Antonio Sgorbissa, Damiano Verda, Renato Zaccaria Robotics And Autonomous Systems (RAS), December 2011