Projects

In this page we summarize our past and current projects. Their associated software products are maintained in our FusionForge (migrating to GitLab…).

UCCare: Software Basado en Modelos para Automatizar la Gestión del Riesgo de Enfermedades Contagiosas (e.g., Covid-19) y Asistencia a Autoridades Delegadas Sanitarias en el Regreso a Actividades Presenciales

El objetivo principal de este proyecto es construir e implementar un modelo parametrizable, según las características de enfermedades altamente contagiosas, como la Covid-19, que permita automatizar la gestión del riesgo de esta clase de enfermedades, así como brindar soporte y asistencia a autoridades delegadas sanitarias de instituciones particulares en su regreso a actividades presenciales.

La concepción y desarrollo de UCCare se resume en esta presentación.

El sistema UCCare tiene funcionalidades que pueden usarse a través de un navegador web (https://www.icesi.edu.co/uccare/) aunque algunas funcionalidades requieren descargar la App respectiva: para Android y para iOS.

La política de uso de datos de UCCare se puede consultar en la misma página de UCCare.

UCCare fue presentado en el Foro Virtual “Investigaciones en Cali de Frente a la Pandemia”, organizado por la Secretaría de Salud de Cali, agosto 27 y 28 de agosto de 2020. Ver PDF.

Project Leader:Gabriel Tamura
Participants:Norha Villegas, Juan Carlos Muñoz, Yoseth Ariza, Domiciano Rincón
Research Line: Model-Driven Software Engineering

AVIOM: A Virtual Assistant for Controlling the Operation of the SITM-MIO

The main goal of this project is to develop a virtual assistant that, realized as a self-adapting and context-aware software system, increases the resilience and capacity of the mass transportation system of the city of Cali (SITM-MIO by its initials in Spanish). This software would assist the Operation Control Center (OCC) personnel in their daily mission for the SITM-MIO to comply with the defined Operational Service Plans, based on the application of autonomous systems, data analytics, data modeling techniques, machine learning and algorithms in general.

In the aspect of science, technology and innovation, with the development of the virtual assistant of the SITM-MIO, we expect to advance in the high-impact technologies convergence, applied in the domain of mass public transport management systems and mobility. Indeed, that advance would result from the combination of autonomous systems with internet of things and analytics / machine-learning components, incorporated as a whole in a software system, for the improvement of operational compliance of the SITM-MIO with respect to the plans of operational service.

Project Leader:Gabriel Tamura
Participants:Norha Villegas, Javier Díaz, Andrés Aristizabal, Juan Carlos Muñoz, Julio César Alonso
Research Line: Situation-Aware Self-Adaptive Software Systems, Big Data and Analytics

CAOBA: Colombian Center of Excellence in Big Data and Data Analytics

The main goal of this project was to create CAOBA, a Center of Excellence and  Appropriation in Big Data and Data Analytics for strengthen the generation of solutions based on data analytics and supported by software technologies in industrial sectors, governmental and academic institutions in Colombia.
More precisely, CAOBA is a collective initiative that groups 11 entities: 4 universities (Pontificia Universidad Javeriana, Universidad de Los Andes, Universidad EAFIT, and Universidad Icesi); 4 entities from the industrial/government sectors (Nutresa, Bancolombia, DNP, Cluster CreaTIC); and 3 IT/Big Data/Analytics market leaders (Dell-EMC, IBM, SAS).
As a Center of Excellence, CAOBA was formed with the sponsorship and financial support of Colciencias and the Ministry of ICT, to develop projects and solutions in the Big Data/Analytics area. Within this initiative, several professors from the ICT Department (Engineering Faculty) and from the Economics and Finance Departments (Administration and Economics Faculty) have developed several projects for Nutresa, Bancolombia, and Public entities (under the umbrella of the DNP).
Local Project Leader:Gabriel Tamura
Participants:Norha Villegas, Javier Díaz, Juan Manuel Salamanca, Julio César Alonso, Luis Berggrun
Research Line: Big Data and Analytics

Reliability on Self-Controlling Mechanisms in Software Systems (ReSeCo)

This research project focuses on the investigation of mathematical relationships between adaptation goals (i.e., quality attributes), adaptation mechanisms, and adaptation properties, with the goal of advancing toward the definition of a theory that helps designing reliable self-controlled software systems.
Project Leader:Gabriel Tamura
Participants:Norha Villegas, Juan M. Ropero, Karen L. Lara
Research Line:Situation-Aware Self-Adaptive Software

2018

Characterizing Geographic Areas Using Natural Language Processing

In data analysis, context information plays a significant role in enhancing the quality of the insight obtained. One type of context information that can be exploited is location. For instance, the location in which a series of events occurred can contribute to explain the event itself. Location context has been studied extensively in the literature, particularly in business applications. Furthermore, spatial analysis helps understand spatial relationships among entities. Nevertheless, one aspect that has been less explored is the characterization of geographic areas based on user generated content, such as opinions. Findings of a comprehensive and Systematic Literature Review (SLR) conducted as part of this thesis suggest that there are no studies focused on exploiting natural language to characterize geographic areas. This thesis focused on investigating how to combine and exploit geographic information with natural language to detect geographic clusters of textual events, and infer relationships between each cluster and a fixed set of categories. We implemented a series of models capable of predicting product classes from a product review text corpus, and then characterizing the obtained geographic clusters based on their aggregated text corpus. The output of our system is a simple and understandable visualization of the geographic areas with their corresponding relevance score against a fixed set of categories. We trained a product document classifier with 72% F1-score, and conducted the validation of our approach by applying the process to several cities and obtaining their characterization. Our approach provides practitioners with a mechanism to combine geographic and natural language features to derive insight from geospatial and text data, and researchers with several opportunities to continue advancing the field from our contributions.

Contribution Summary
As a summary, our main contributions are as follows:

  • We demonstrate the usefulness of a trained knowledge corpus of topics, based on natural language text messages that users post on the internet, as well as their location context, to characterize dense and partitioned geographic areas. Furthermore, we present the analysis results in an effective visualization that helps researchers and practitioners to gain insight from the characterization.
  • We propose a new approach to perform spatial marketing segmentation exploiting user generated content in the form text messages, and we validate the workflow using data sets of considerable size. Furthermore, we argue that our approach can be adapted and applied to other scenarios such as customer engagement, or even geographical analysis of human thinking.
  • We created and made publicly available the product reviews knowledge corpus data set, which consists of eight million document vectors and ten product categories. Furthermore, we made available the document vector model that we used to obtain the product document embeddings, which is useful to generate vectors for new documents, and our product document classifier.

Project’s code
Docker image

Project advisors:Javier Díaz-Cely, Norha M. Villegas
Principal researcher:Luis Ferro (Master Student)
Research Line:(Big Data) Analytics: Techniques and Processing Architectures

2017

Monitoring of Patients in Home Hospitalization Programs through a Telemedicine Platform

This project developed a software solution for the assistance and monitoring of patients admitted to home hospitalization programs, contributing to improve access, efficacy, efficiency, quality, and safety of health-care service provision. The main innovation of this project was the design and implementation of a software solution with characteristics of context-driven self-adaptation, linking current health information systems with mobile technologies and contextual geolocation in time. The project results improved the scheduling efficiency of routes that health personnel must follow to comply with health care programs and home hospitalization routines, resulting in a better use of human and logistic resources in the provision of the service, being able to schedule 163 appointments per unit of time for every 100 allowed by the base system.

The main result of the project, in context, allows the institutions providing health services to increase their capacity to offer health care services for patients both hospitalized in the institution, and those with home hospitalization, with more efficiency and better controls.

Project Leader:Gabriel Tamura
Participants:Juan Manuel Reyes, Norha M. Villegas
Research Line:Situation-Aware, Self-Adaptive Software

Re-Design of a Software Family of Products’ Reference Architecture

In this project, inspired in DevOps principles, we developed a basic framework for performing experiments with the reference architecture of a family of software systems.

In summary, we applied several domain-specific design patterns to a concrete software system based on the reference architecture, producing different variations of the software system. These variations were deployed and executed in a close-to-production infrastructure, running a series of specified experiments, and gathering different measurements. Once the experiments finished, the results were compared and the best alternative’s architecture design patterns were chosen for the reference architecture.

As a result, the re-designed reference architecture processed 5 times the number of requests  that the original one processed per unit of time.

Project Leader:Gabriel Tamura
Participants:Norha M. Villegas
Research Line:Software Architecture, Self-Adaptive Software, Continuous Experimentation

Personalized Web-Tasking with Situation-Aware Smart Applications

The long-term goal of this project is the development of a framework and prototype to guide the design and implementation of situation-aware smart software systems for personalized web-tasking, which includes user-driven web task automation—in the smart internet.
Project Leader:Hausi A. Müller
Participants:Lorena Castañeda, Norha M. Villegas
Research Line:Situation-Aware, Self-Adaptive Software

Application of Context Management Techniques to Big Data Predictive Analytics

This project focuses on the investigation of approaches to improve the accuracy and user-relevance of recommendation systems by applying dynamic context management techniques.
Project Leader:Norha M. Villegas
Participants:Gabriel Tamura, Lorena Castañeda, Cristian Sánchez
Research Line:Situation-Aware, Self-Adaptive Software

2016

SHIFT

An open-source software product line engineering framework and engine.
Project Leader:Hugo F. Arboleda
Participants:Gabriel Tamura, Andrés Paz, Miguel A. Jiménez, Julián Cifuentes
Research Line:Situation-Aware Self-Adaptive Software, Model-Driven Software Development

2013

Integrated System of Indicators and Metrics for Monitoring and Control of Software Factories Projects

To unify data collection of systems that support the sub-processes of technical solution (i.e., solution architecture, implementation and transition to operation) at Carvajal Tecnología y Servicios software factory into a software system that automatically calculates metrics that allow to analyze the performance of those sub-processes.
Project Leader:Gabriel Tamura
Participants:Juan Sebastian Cortes, Sebastian Santamaría, Luis Miguel Muñoz Piedrahita
Research Line:Software Engineering

PaSCAni: A Domain Specific Language for Automated and Composable Testing of Software Components

Thesis Project. A Domain Specific Language for specifying and executing automated, composable, and traceable test specifications for component-based systems complaint with the SCA standard, using the OW2 FraSCAti middleware..
Project Leader:Gabriel Tamura, Angela Villota Gómez
Participants:Fabián A. Caicedo, Miguel A. Jiménez
Research Line:Situation-Aware Self-Adaptive Software

2012

Improving Context-Awareness in Situation-Aware Smart Software and Self-Adaptive Systems

This project focuses on the investigation of models and mechanisms to represent, acquire, provide, and reason about context information to improve context-awareness in user-centric situation-aware smart software (SASS) systems.
Project Leader:Norha M. Villegas
Participants:Gabriel Tamura
Research Line:Situation-Aware Self-Adaptive Software, Software Architectures and Emerging Approaches

Self-Healing: Exploration of the State of the Art, Approaches and Specifications

The main goal of this project is to characterize requirements specification and corresponding verification and validation (V&V) methods. Based on this characterize, to propose models and incorporate them and corresponding suitable V&V methods for them in a controlling feedback-loop that dynamically monitors the satisfaction of the system requirements, as proposed by Tamura et al.
Project Leader:Gabriel Tamura
Participants:Norha Villegas, Angela Villota, Miguel Jimenez, Andrés Villegas
Research Line:Situation-Aware Self-Adaptive Software

SmarterContext: Ensuring Security and Confidentiality in the Management of Context Information for Smarter Web Applications

This project focuses on enabling SmarterContext, the dynamic context management infrastructure proposed in Norha Villegas’ dissertation, with efficient mechanisms to guarantee security, privacy and confidentiality in the management of context information.
Project Leader:Norha M. Villegas
Participants: Juan C. Muñoz, Gabriel Tamura, Hausi A. Müller
Research Line:Situation-Aware, Self-Adaptive Software

Application and Preliminary Evaluation of the Feedback-Control Reference Model for Building Self-Adaptive Software.

The goal of this project is to evaluate this reference model by applying it to a case study. The analysis of the obtained results will give us important insights about the practical applicability of our reference model and to further improve it. The objective of the evaluation will be focused on the support that our model offers to (i) maintain the separation of concerns between the adaptation mechanism and the managed application software system; and (ii) the achievement of the continuous fulfillment of non-functional requirements under changing execution context.
Project Leader:Gabriel Tamura
Participants:Juan David Gómez, Lorena Castañeda
Research Line:Situation-Aware Self-Adaptive Software

2011

Adaptive Context Management to Smarten Up User-Centric Web Applications in the Personal Web

This project focuses on the implementation of a prototype for a user-driven context management framework to discover personal context entities and enable the user in the dynamic integration of these entities into an e-commerce personal web sphere.
Project Leader:Norha M. Villegas
Participants:Gabriel Tamura, Juan Carlos Muñoz, Hausi Müller
Research Line:Situation-Aware Self-Adaptive Software, Software Architectures and Emerging Approaches

Feedback-Control Reference Architecture and Implementation for Building Self-Adaptive Applications.

This project was focused on to design and implement a reference architecture and corresponding functional proof-of-concept, inspired by control engineering, that makes explicit the crucial elements of feedback-loops, and with them, the feedback-loops themselves that are required for context-based self-adaptive software systems.
Project Leader:Gabriel Tamura
Participants:Cristophe Demarey, Lorena Castañeda
Research Line:Situation-Aware Self-Adaptive Software

2010

Exploration of Dynamic Self-Adaptive Components in Model-Driven Software Product Lines for Service Oriented Applications

This project aims to contribute the search for solutions for these problems, focusing on the realization of automatic derivation of dynamic self-adaptive (context-aware) software products, based on Model Driven and Dynamic Service-Oriented Product Lines (MD-DSOPL).
The project will be developed about a particular case of study based on service oriented applications for mobile devices.
Project Leader:Gabriel Tamura
Participants:
Research Line:Situation-Aware Self-Adaptive Software

2008

Integration and Interoperability of Orders and Results of Clinical Laboratory Tests Using HL7 and Smart Cell Phones

The main purpose of this project is to increase the efficacy, efficiency and cost-opportunity, in human-health terms, of the healthcare and tracking of patients’ clinical state through the development of a comprehensive implementation of the proof-of-concept software system described in section 1, initially in the context of the Imbanaco Medical Center, and as a basis for the Colombian HL7 Laboratory Use Case Technical Committee. The proposed system should be able of sending alert notifications to medical doctors through smart cell phones via SMS, together with a web system for ordering these kinds of tests and for publishing their detailed results, as well as the corresponding patients’ state, accessible from smart cell phones through WAP or other HCIs (Human Computer Interfaces).
Project Leader:Gabriel Tamura
Participants:Norha Villegas, Mauricio Rodríguez, Diana Lorena Gómez, Juan Sebastián Botero, Harold Candelo, Carlos Guarín, Andrés Giraldo, Cristhian Cuenca, Andrés Espada, Fernando Portilla
Research Line:   Software for Medical Applications (inactive)