As I have been exposed to interdisciplinary educational training, I conducted research in multiple areas. This page includes the different research activities and projects I worked on and implemented at the different institutions I have been affiliated to.
PySeqLab: an open source Python package for sequence labeling and segmentation
Python Sequence Labeling (PySeqLab) is an open source package for performing supervised learning in structured prediction tasks. It implements conditional random fields (CRFs) models from (1) first-order to higher-order linear-chain CRFs, and from (2) first-order to higher-order semi-Markov CRFs (semi-CRFs). The package supports multiple training methods (perceptron- and gradient- based algorithms) in addition to multiple decoding methods (i.e. Viterbi and Viterbi A*). We demonstrate the use of the package in three different domains: (1) biomedical Natural language processing (NLP), (2) predictive DNA sequence analysis, and (3) Human activity recognition (HAR).
The package’s documentation and tutorials could be accessed through this link. An accompanying application note was published in Bioinformatics journal. The development of the package was funded by grant (number 161635) as part of the “Early Postdoc.Mobility” funding scheme from the Swiss National Science Foundation (SNSF).
A-DISCERN: Identifying high-quality online information on medical treatment options
The fundamental intention of this research project is to create a system that provides automatically a quality prediction based on the evaluation of an established tool for quality assessment (DISCERN) and to make this assessment easily available to everybody while they are looking at health webpages discussing health information (i.e. medical treatment options).
Theoretically, this means adding to the informative message of the webpage a meta-message with the intent to reduce or eliminate acceptance of messages originating from low-quality webpages.
A first article out of the project got published in the Journal of the American Medical Informatics Association reporting the characteristics/features of the corpus and implementing/discussing model-based methods (probabilistic models) for aggregating multiple annotations (provided by multiple annotators) to obtain a reference corpus.
The project was funded by grant (number 161635) as part of the “Early Postdoc.Mobility” funding scheme from the Swiss National Science Foundation (SNSF).
PhD Dissertation: The Role of Search Engines and Potentials of Gamification in Consumer Health Informatics: Experiments in Health Information Seeking and eHealth Interventions
My dissertation consisted of four studies that aimed at advancing and presenting new ideas in two tracks; online health information seeking and eHealth interventions, which both fall within the consumer health informatics discipline. It was conducted at the Institute of Communication and Health (ICH) under the supervision of Prof. Peter J. Schulz.
I designed and conducted two randomized controlled trial experiments on Internet- and Web- based interventions hosted on a newly redeveloped platform called ONESELF.
In the first study, I investigated the effect of interactivity (operationalized by the access to different interactive features of the intervention) on empowerment and self-management behavior for patients with chronic low back pain.
In the second intervention, I studied the effects of online social support and gamification (“the application of game design elements and mechanics in a non-game context”) on health outcomes (health care utilization and medication overuse), patients’ behavior (physical activity), empowerment and knowledge for patients with rheumatoid arthritis. The results provided the first piece of empirical evidence on the potentials of gamification in the health care domain.
Online health information seeking:
In this line of research, I explored the online health information seeking process by conducting two experiments manipulating Google search engine. The experiments demonstrated the influence of selection and sorting/ranking criteria operating in search engines (Google) on users’ knowledge, beliefs and attitudes towards a controversial health topic (i.e. vaccination).
A secondary analysis of the user behavior, navigation and preference of retrieved webpages led to preparing an article that was published in Computers in Human Behavior journal. It reported the study of the contribution of various quality markers and attributes to the overall quality of health websites, taking into consideration the perception of online information seekers toward these visited and evaluated webpages.
Manipulating Google’s knowledge graph box to counter biased information processing during an online search on vaccination. Application of a technological debiasing strategy.
The aim of this research was to test a technological debiasing strategy to reduce the negative effects of biased information processing when using a general search engine on people’s knowledge and attitude towards vaccination. This was realized by manipulating the content of a knowledge graph box integrated in the interface of a general search engine (Google). The manuscript reporting on the conducted experiment and the results was published in the Journal of Medical Internet Research. This research project was done in collaboration with Ms. Ramona Ludolph.
Cognitive Dissonance and MMR Vaccination Information: Influence of Attitudes and Beliefs on Selective Exposure, Perception and Retention
In this study, I participated in the design of the experiment and developed the online experimental manipulation (pre-post design) to better understand how people with pro- or anti-vaccination stances consume online information. This research is guided by cognitive dissonance theory and is an examination of the impact of MMR vaccine attitudes on selective exposure to and retention of online MMR information. It was done in collaboration with Dr. Anne-Linda Camerini.
Judging the credibility of online health information: the interaction between user health literacy and website quality
In this research project, I worked on the implementation of an online experiment in which users were asked to rate a predefined set of health websites discussing chronic low back pain based on set of quality criteria. The aim of the experiment was to investigate the interaction between the quality of websites and people’s health literacy in explaining their ability to evaluate online health information. The study was done in collaboration with Dr. Nicola Diviani.
Discussions on childhood vaccinations: A content analysis of Italian online forums
In this project, we investigated and analyzed forum posts from three famous Italian forums discussing childhood vaccination. The goal was to understand users’ role and posting patterns on the forums. The manuscript reporting the findings of this research was published in Vaccine. Moreover, the results of this research were highlighted and featured on a well-known Italian newspaper “il sole 24 ore” at this link. This research was done in collaboration with Ms. Marta Fadda.
Master thesis: Video Code Identification
Under the supervision of Prof. Marco Tagliasacchi, I worked on my Master thesis that dealt with a setting in which a video content was passed in a processing chain of two coding steps and the goal was to identify the first type of codec used in the first step. By exploiting the coding-based footprints of the video stream, my thesis reported a method in which it was feasible to detect and identify the first encoding. The application of this thesis lies within the field of multimedia forensics (i.e. to identify the device that generated the original video stream or detect collages of different sequences). A publication based on my Master thesis was published in the IEEE International Conference on Acoustics, Speech, and Signal Processing 2012.
Green Cloud: Advanced Energy Management in Cloud systems
The aim of the Green Cloud project (project 5) was to develop a systematic set of methods towards the design of novel resource allocation policies for energy-aware Clouds. I was part of a team consisting of 5 members who were working on the project. By constructing and developing a mathematical model (optimization-based model) and using linear programming, formulated and implemented in AMPL modeling language, the model showed to reduce the energy consumption in the Cloud system, taking into consideration the communication network and the energy generation methods, and emphasizing on the use of green and renewable energies. Exhaustive simulations established the efficiency and efficacy of the proposed model that showed to minimize significantly the energy cost in both Data Centers and in the network.