Machine Learning

Much of the progress in the task of extracting knowledge from data is obtained through research in machine learning. Two sub-areas of AM are of particular interest in this research project: (i) Deep Learning and (ii) Reinforced Learning.

Deep Learning encompasses a set of techniques designed to simulate the behavior of the human brain in tasks such as visual recognition, speech recognition and natural language processing. These techniques try to produce high-level hierarchical representations from input data, through layers of sequential processing.

Reinforcement learning studies how software agents can learn to perform actions rationally in an environment to maximize their reward. Learning in this context happens by providing reinforcement signals (which can be negative or positive) associated to the actions selected by the agent. RL can be applied to various tasks, such as planning, task scheduling, games and more.

Researchers:

• Diego Haddad
• Eduardo Bezerra (Coordinator)
• Pedro Gonzalez

Financial Information:

1. Announcement CNPq-SETEC / MEC Nº 17/2014 – Line 1: PD&I (Support for Cooperative Projects for Applied Research and Technological Extension), project “FXCode: data compression for digital files”, in the 2015-2016 period, with coordination of the professor Eduardo Bezerra. The project had 04 ADC-A grants and 01 DTI-B grants during its term. Financed amount R$ 128,868.00.
2. Notice of Research Group of CEFET / RJ, project “Research Group in Machine Learning”, in the 2016-current period, with the coordination of professor Eduardo Bezerra. Financed amounts: R$ 21,319.18 (2017) and R$ 120,000.00 (2019).
3. PIBIC Scholarships

These projects have been under development by group members since 2015 and have a total financing value of approximately R$ 270,187.18.