Inverse Mapping: Employing Interactive Manipulation to Transform Computational Models

• Grant number: 15/08118-6
• Support Opportunities: Scholarships in Brazil - Doctorate
• Start date: January 01, 2016
• End date: June 30, 2020
• Field of knowledge: Physical Sciences and Mathematics - Computer Science - Computing Methodologies and Techniques
• Principal Investigator: Fernando Vieira Paulovich
• Grantee: Gabriel Dias Cantareira
• Host Institution: Instituto de Ciências Matemáticas e de Computação (ICMC). Universidade de São Paulo (USP). São Carlos , SP, Brazil
• Associated research grant: 11/22749-8 - Challenges in exploratory visualization of multidimensional data: paradigms, scalability and applications, AP.TEM
• Associated scholarship(s): 17/08817-7 - Distance Learning and Inverse Mapping of Visualizations Applied to Text Mining, BE.EP.DR

Abstract

With the rising quantity and complexity of data stored by computational systems of all kinds, exploring and extracting knowledge from such data has become an increasingly hard task. In order to solve this problem, techniques based on data mining, machine learning and information visualization are in high demand, appearing in many recent scientific developments. The field known as Visual Analytics corresponds to the utilization of visual models to allow users to apply their knowledge in the process of mining or learning, thus resulting in better efficiency in the execution of certain tasks. However, even though there are many visualization techniques that carry information from the data set to the visual space, the inverse process is not always done in a simple way, making transmitting ideas and knowledge from the visual model back to the computational model a flawed and sometimes inefficient procedure. This doctorate project's proposal is to study and develop new visualization approaches that support the concept of inverse mapping, especially regarding multidimensional projection techniques, focusing on their usability in visual analytics and offering users the ability to understand how the data shown differs from the original model and how interactions in the visualization affect such model.

Scientific publications

(References retrieved automatically from Web of Science and SciELO through information on FAPESP grants and their corresponding numbers as mentioned in the publications by the authors)

CANTAREIRA, GABRIEL D.; PAULOVICH, FERNANDO, V; ETEMAD, ELHAM; KERREN, A; HURTER, C; BRAZ, J. Visualizing Learning Space in Neural Network Hidden Layers. VISAPP: PROCEEDINGS OF THE 15TH INTERNATIONAL JOINT CONFERENCE ON COMPUTER VISION, IMAGING AND COMPUTER GRAPHICS THEORY AND APPLICATIONS, VOL 4: VISAPP, v. N/A, p. 12-pg., 2020-01-01. (17/08817-7, 15/08118-6)

CANTAREIRA, GABRIEL D.; ETEMAD, ELHAM; PAULOVICH, FERNANDO V.. Exploring Neural Network Hidden Layer Activity Using Vector Fields. INFORMATION, v. 11, n. 9, p. 15-pg., 2020-09-01. (17/08817-7, 15/08118-6)

Academic Publications

(References retrieved automatically from State of São Paulo Research Institutions)

CANTAREIRA, Gabriel Dias. Enhancing Dimensionality Reduction Techniques for Deep Neural Network Visualization. 2020. Doctoral Thesis - Universidade de São Paulo (USP). Instituto de Ciências Matemáticas e de Computação (ICMC/SB) São Carlos.