Publications

A Deep Reinforcement Learning Approach with Visual Semantic Navigation with Memory for Mobile Robots in Indoor Home Context

Published in Journal of Intelligent and Robotic Systems (JINT) , 2022

Navigate is a crucial task to be performed by social robot agents in complex and uncertain environments like homes and offices. Search for specific target objects is usually a required activity. This work aimed to propose a visual semantic navigation with memory architecture model. Based on recent advances in convolutional neural networks and graph neural networks, a visual semantic navigation architecture model (GCN-MLP) was extended with recurrent neural networks for memory mechanisms (GCN-GRU and GCN-LSTM) while exposing a robot agent in navigation experiences to learn navigation policies. The models were evaluated quantitatively and qualitatively, where memory enhanced models demonstrated early convergence, better performance in evaluation metrics, increased successfully terminated episodes, more efficient path trajectories, and lower decrease in performance when exposed to challenger test scheme, presenting a more exploratory behavior. Finally, were analyzed differences between GRU and LSTM, where GRU performed similarly to LSTM in some cases, being a viable option.

Recommended citation: Santos, I.B.d.A., Romero, R.A.F. A Deep Reinforcement Learning Approach with Visual Semantic Navigation with Memory for Mobile Robots in Indoor Home Context. J Intell Robot Syst 104, 40 (2022). https://doi.org/10.1007/s10846-021-01566-0 https://link.springer.com/article/10.1007/s10846-021-01566-0

Deep Reinforcement Learning for Visual Semantic Navigation with Memory

Published in Dissertação (Mestrado em Ciências de Computação e Matemática Computacional) - Instituto de Ciências Matemáticas e de Computação, Universidade de São Paulo, São Carlos, 2020, 2020

The navigation of mobile robots is a subject vastly studied in the last decades, being a crucial task for the insertion of robots in diverse scenarios. However, complex and changeable environments, as indoors of houses, still shows challengers to be transpassed, being an object of study in several works that adopts approaches as computer vision without topological or metric maps. This work proposes an architecture for the navigation of mobile robots aiming target-object search in indoor ambiances of houses, using computer vision methods and semantic information with memory. The proposed architecture can generalize through a priori acknowledgment of detect objects in scenes and reinforce relationships over experiences of the past, in a learning-based navigation approach. Therefore, the following models of machine learning will be adopted: neural convolutional netwoks, graph neural networks, recorrent neural networks and deep reinforcement learning, in a targetobject approach. This architecture has trained in several domestic ambiances, adopting a photo-realistic simulated environment. The architecture was evaluated through qualitative analysis, executing episodes of the agent in the simulated environment with visual insight, and quantitative analysis, adopting metrics like success rate and success rate weighted by path length. Policies learn by the proposed architecture were compared with agents using random policies, agents using only reinforcement learning, and, finally, agents with navigation semantic policies without memory. The experiments performed showed a more exploratory behavior of the proposed architecture when compared with the nonmemory approaches. reaching better success rates in the tasks for both metrics. When exposed to restrict scenarios, consequently being of greater difficulty, the policies learn by such models demonstrated better results, with a lower decrease in its performance when compared with less restrictive executions and other models. Thus, the proposed model presented consistent results with better policies learn by the agents, resulting in behaviors more successful in the task of target-object search in indoor-home environments.

Recommended citation: SANTOS, Iury Batista de Andrade. &quote Aprendizado por reforço profundo para navegação visual semântica com memória."e; Dissertação (Mestrado em Ciências de Computação e Matemática Computacional) - Instituto de Ciências Matemáticas e de Computação, Universidade de São Paulo, São Carlos, 2020. 1(1). https://www.teses.usp.br/teses/disponiveis/55/55134/tde-16122020-164714/pt-br.php

Explainable Machine Learning for Breast Cancer Diagnosis

Published in 2019 8th Brazilian Conference on Intelligent Systems (BRACIS), 2019

Cancer is already the leading cause of death in most Brazilian cities and in the world. The understanding of its internal mechanisms and the design of computational models capable of improving its diagnosis will have strong benefits for humanity. New technologies have made available a wealth of data, which can be used to improve the diagnosis of cancer. As a manual analysis of this data is impracticable, many black-box machine learning algorithms have been employed successfully for cancer diagnosis. Despite their high accuracy prediction abilities, black-box models sacrifice transparency and accountability. In contrast, interpretable machine learning algorithms are powerful tools for understanding the underlying mechanism present within a large corpus of data. In this work, Linear Projections and Radviz were used as visualization techniques for data exploration and feature selection. Further, Decision Tree induction algorithms were used to create models that are able to differentiate between Malignant and Benign breast tumors from breast mass images. These models can be considered white-box models which means their inner workings are easier to explain and interpret. The result shows Classification and Regression Trees achieved an accuracy of 96% in predicting breast cancer.

Recommended citation: T. Brito-Sarracino, M. Rocha dos Santos, E. Freire Antunes, I. Batista de Andrade Santos, J. Coelho Kasmanas and A. C. Ponce de Leon Ferreira de Carvalho, "Explainable Machine Learning for Breast Cancer Diagnosis," 2019 8th Brazilian Conference on Intelligent Systems (BRACIS), Salvador, Brazil, 2019, pp. 681-686, doi: 10.1109/BRACIS.2019.00124. https://ieeexplore.ieee.org/document/8923961

Human-House Interaction Model Based on Artificial Intelligence for Residential Functions

Published in International Conference on Human-Computer Interaction, 2017

The scenarios as smart homes and its devices requires novel ways to perform interactive actions. In this work we explore and develop a model to interact, in a natural, easy learning and intuitive manner, with a smart home, without use special sensors or another controllers, based on interpretation of complex context images captured with a trivial camera. We use artificial intelligence and computer vision techniques to recognize action icons in a uncontrolled environment and identify user interact actions gestures.

Recommended citation: Santos B.F.L., de Andrade Santos I.B., Guimarães M.J.M., Benicasa A.X. (2017) "Human-House Interaction Model Based on Artificial Intelligence for Residential Functions." Stephanidis C. (eds) HCI International 2017 – Posters Extended Abstracts. HCI 2017. Communications in Computer and Information Science, vol 714. Springer, Cham. 1(2). https://link.springer.com/chapter/10.1007/978-3-319-58753-0_51

Dynamic Natural Interaction with Projected Images

Published in XIII Encontro Nacional de Inteligência Artificial e Computacional, 2016

In the last years computers have been turned in common day-by-day devices. Trying to facilitate and approximate the interaction between man and the machine, new forms to interact has been researched and developed, as the use of gestures. This work has as objective the development of a alternative model to real sensors, taking advantage of techniques for scenes interpretation and computational vision, through capture of images, pre-processing, visual attention, segmentation and classification. The model was tested in several environments and, according to the accuracy values obtained from a confusion matrix, has been possible conclude for its applicability and efficiency.

Recommended citation: I. B. A. Santos, B. F. L. Santos, L. A. Fonseca Sobrinho, Alcides Xavier Benicasa (2016). "Dynamic Natural Interaction with Projected Images." XIII Encontro Nacional de Inteligência Artificial e Computacional. 1(1). http://www.lbd.dcc.ufmg.br/colecoes/eniac/2016/062.pdf

Iury Andrade