Prof. Dr. Alberto Ferreira De Souza

Prof Alberto Ferreira De Souza, PhD (University College London)

Professor Emeritus of Computer Science

Universidade Federal do Espírito Santo (UFES)

6,846

Total Citations

27

h-index

77

i10-index

Last updated: March 2026

About

My research career began in computer architecture and high-performance processing during my undergraduate studies in Electronic Engineering at UFRJ, through undergraduate research at the Systems Laboratory of the Graduate Program in Systems and Computer Engineering (PESC) at COPPE/UFRJ [1]. My undergraduate project, developed within PESC, was the first parallel machine developed in Brazil [2].

During my master's degree at COPPE, I conducted research in computer architecture and high-performance processing, being one of the designers of the NCP-1 supercomputer - Brazil's first supercomputer [3]. In the same period, I had my first contact with artificial intelligence in general and neural networks in particular, topics to which I would later dedicate myself.

In 1993, I joined UFES as a professor. From 1996 to 1999, I completed my PhD at University College London, UK, where I proposed the processor architecture called Dynamically Trace Scheduled VLIW (DTSVLIW) [4]. Upon returning, in 2002, I created the High-Performance Computing Laboratory (LCAD), which I coordinated until 2019, when I retired and began serving officially as a volunteer professor at UFES.

From the end of my doctorate to the present, I have contributed to the university in various capacities, including Coordinator of the Graduate Program in Informatics (PPGI), Coordinator of the Doctoral Program Creation Commission (created in 2010), Director of the Institute of Technology, Vice-Director of the Technology Center, Dean of Planning and Institutional Development, and Research Director of the university. In 2023, I was awarded the title of Professor Emeritus of UFES.

From the late 2000s, I began to dedicate myself more intensely to AI, neural networks, and their applications in autonomous robotics and other areas. I served as Paper Submission Chair of the IEEE World Congress on Computational Intelligence 2018 (WCCI'2018) and Director of I2CA. My highest-impact scientific articles are from these areas.

Currently, I serve as a faculty member of the Graduate Program in Informatics (PPGI) at the Federal University of Espírito Santo (UFES). In addition to my academic activities, I am the Chief Technology Officer (CTO) and co-founder of Lume Robotics S.A., the Chief Scientific Officer (CSO) and co-founder of AUMO S.A., and an advisor to Motora Technologies S.A., in which I also hold equity participation. While my earlier work concentrated on artificial intelligence for autonomous systems and large-scale computational infrastructures, my current research focus has increasingly shifted toward Digital Biology, particularly the integration of genomic foundation models, large neural architectures, and neurosymbolic reasoning frameworks for interpretable scientific discovery. Across academia and industry, my work aims to advance artificial intelligence and applied computational intelligence, fostering technology transfer and innovation between research laboratories and high-tech startups. You can learn more about me and my early academic trajectory here.

Current Affiliations

International Collaborations

  • City St George's, University of London (Prof. Artur Garcez)
  • University of Twente (Prof. Giancarlo Guizzardi)
  • Rutgers University (Prof. Kostas Bekris)

Education

PhD in Computer Science

University College London, UK

1996 – 1999

MSc in Computer Engineering

COPPE/UFRJ, Brazil

1989 – 1993

BSc in Electronic Engineering (Cum Laude)

UFRJ, Brazil

1984 – 1988

Generative AI

GPT, Large Language Models, Transformers Architecture

Deep Learning

Neural Networks, CNNs, Computer Vision Applications

Visual Cognition

Computational Models of Human Visual Perception

Autonomous Robots

Mobile Robotics, Path Planning, Autonomous Vehicles

Teaching Topics

My teaching focuses on cutting-edge topics in Artificial Intelligence and Robotics at both Master's and PhD levels. Recent courses cover Generative Pre-Trained Transformers and large language models (2023-2025), while foundational courses address Deep Learning architectures and their applications to computer vision and autonomous systems.

I also teach Visual Cognition, exploring computational models of human visual perception, and supervise numerous Directed Studies on autonomous vehicles, path planning, and AI applications in healthcare and industry. All courses are offered at the graduate level (PINF-6*** for Master's and PINF-7*** for PhD programs).

Research Interests

Digital Biology

Genomic Foundation Models, Multi-Species Learning, AI-Driven Biomarker Discovery

Neurosymbolic AI

Neural-Symbolic Integration, Knowledge Graphs, Ontology-Guided Reasoning

Deep Learning

Transformers, Large Language Models, Generative AI Systems

Autonomous Vehicles

Self-Driving Cars, Computer Vision, Sensor Fusion

Current Research Focus

My current research focuses on developing AI systems for digital biology that integrate genomic foundation models with neurosymbolic reasoning frameworks. The goal is to enable interpretable discovery from large-scale genomic and multi-species data, identifying biologically meaningful patterns associated with healthy longevity and age-related diseases.

Methodologically, this work combines long-context genomic transformers and large language model–style architectures with ontology-guided reasoning and biological knowledge graphs. By coupling neural prediction with structured validation, the approach aims to move beyond black-box inference toward hypothesis generation and mechanistic insight, contributing to AI-driven discovery in complex biological systems.

Selected Publications

View all publications

Self-driving cars: A survey

C Badue, R Guidolini, RV Carneiro, P Azevedo, VB Cardoso, A Forechi, ...

Expert systems with applications 165, 113816, 2021, 2021

1,748 citations

Facial expression recognition with convolutional neural networks: coping with few data and the training sample order

AT Lopes, E De Aguiar, AF De Souza, T Oliveira-Santos

Pattern recognition 61, 610-628, 2017, 2017

1,015 citations

Keep your eyes on the lane: Real-time attention-guided lane detection

L Tabelini, R Berriel, TM Paixao, C Badue, AF De Souza, ...

Proceedings of the IEEE/CVF conference on computer vision and pattern …, 2021, 2021

622 citations

Polylanenet: Lane estimation via deep polynomial regression

L Tabelini, R Berriel, TM Paixao, C Badue, AF De Souza, ...

2020 25th international conference on pattern recognition (ICPR), 6150-6156, 2021, 2021

479 citations

Prediction-based portfolio optimization model using neural networks

FD Freitas, AF De Souza, AR De Almeida

Neurocomputing 72 (10-12), 2155-2170, 2009, 2009

262 citations

Copycat cnn: Stealing knowledge by persuading confession with random non-labeled data

JR Correia-Silva, RF Berriel, C Badue, AF De Souza, T Oliveira-Santos

2018 International joint conference on neural networks (IJCNN), 1-8, 2018, 2018

259 citations

A dataset for improved rgbd-based object detection and pose estimation for warehouse pick-and-place

C Rennie, R Shome, KE Bekris, AF De Souza

IEEE Robotics and Automation Letters 1 (2), 1179-1185, 2016, 2016

240 citations

Cross-domain car detection using unsupervised image-to-image translation: From day to night

VF Arruda, TM Paixao, RF Berriel, AF De Souza, C Badue, N Sebe, ...

2019 International joint conference on neural networks (IJCNN), 1-8, 2019, 2019

155 citations

Large-scale mapping in complex field scenarios using an autonomous car

F Mutz, LP Veronese, T Oliveira-Santos, E De Aguiar, FAA Cheein, ...

Expert Systems with Applications 46, 439-462, 2016, 2016

121 citations

Traffic light recognition using deep learning and prior maps for autonomous cars

LC Possatti, R Guidolini, VB Cardoso, RF Berriel, TM Paixão, C Badue, ...

2019 international joint conference on neural networks (IJCNN), 1-8, 2019, 2019

108 citations

Neurosymbolic Genomics

Neurosymbolic AI for Genomics

Developing a neurosymbolic AI platform for identifying genetic and epigenetic biomarkers associated with healthy longevity and age-related diseases. Combines genomic transformers with biological knowledge graphs.

View on GitHub

Deep Tech Startups

Spin-off Companies

Co-founder of three deep tech startups born at LCAD, collectively employing over 150 people:

IARA - Intelligent Autonomous Robotic Automobile

The most advanced autonomous vehicle in the Southern Hemisphere. IARA made an unprecedented autonomous journey from Vitória to Guarapari (74 km). The project includes over 13 million lines of open-source code.

View on GitHub

Embraer Autonomous Taxiing System

Development of the world's first autonomous taxiing system for passenger jet aircraft, in partnership with Embraer. This groundbreaking project is now part of aviation history.

World First Achievement

Honors & Awards

🏅 2023

Professor Emeritus

Universidade Federal do Espírito Santo (UFES)

🏆 2021

CONFAP Science, Technology & Innovation Award

3rd place - Innovative Researcher Category

🌟 2017

Researcher of the Year

State Council of Science and Technology - ES

🤖 2015

7th Place - Amazon Picking Challenge

Amazon Inc.

🎯 2013

12th Place - DARPA Virtual Robotics Challenge

DARPA - Defense Advanced Research Projects Agency

🎖️ 2009

IEEE Senior Member

Institute of Electrical and Electronics Engineers

Media Gallery

Lume Robotics: Autonomous Vehicle Tests in the Steel Industry

Autonomous tri-articulated truck transporting 120 tons of steel coils at ArcelorMittal

Lume Robotics: Mercedes-Benz Partnership for Autonomous Trucks

Mercedes-Benz announces partnership with Lume Robotics for autonomous trucks

Lume Robotics: Autonomous Minibus

Lume autonomous minibus demonstration

Lume Robotics: Autonomous Sample Transport at CENPES Petrobras

100% electric and autonomous vehicle for sample and cargo transport

IARA on Globo TV (Jornal Nacional)

National television coverage - 74km autonomous journey

First Brazilian Electric Autonomous Car

Partnership LCAD/UFES - Lume Robotics

DARPA Virtual Robotics Challenge

12th place worldwide

In The Press

Selected media coverage of research activities and achievements.

TV January 2026

Industry 4.0: Autonomous Trucks and Robots Revolutionize Logistics

VTV News report on how autonomous trucks and robots are transforming logistics operations in Amparo.

Professors Ricardo Carminati de Mello and Alberto Ferreira with the Prometheus humanoid robot
Newspaper October 2025

How About a Coffee Made by a Robot?

The Prometheus humanoid robot is being trained at UFES to prepare coffee autonomously — boiling water, positioning the filter, measuring coffee, pouring water, and serving. The project, coordinated by Prof. Alberto Ferreira, aims to develop Autonomous Computational Intelligences at the I2CA Institute, with completion expected by mid-2026.

Read on A Tribuna
Lume Robotics autonomous three-axle truck operating at ArcelorMittal
News July 2025

Where's the Driver? Three-Axle Truck Drives Itself in Espírito Santo

A tri-articulated truck capable of carrying up to 120 tons traveled without human intervention through ArcelorMittal facilities. Lume Robotics, an exponent of the autonomous vehicle industry flourishing in Vitória, achieved a world-first: the first three-axle truck operating autonomously.

Read on UOL Tilt
Lume Robotics autonomous truck transporting steel coils at ArcelorMittal
International July 2025

Autonomous Truck Succeeds in Steel Transport

International coverage of Lume Robotics' successful autonomous truck operation transporting steel coils at ArcelorMittal facilities in Brazil.

Read on SelfDriveNews
TV July 2017

UFES Autonomous Car Featured on Jornal Nacional

IARA (Intelligent Autonomous Robotic Automobile), the autonomous car from UFES, was featured on Jornal Nacional, Brazil's main news program on Rede Globo.

TV July 2017

Autonomous Car Completes Historic Trip to Guarapari

For the first time, a Brazilian autonomous vehicle traveled 74 km on urban and highway roads without human intervention, from UFES campus to Meaípe beach in Guarapari.

Read full article (UFES)
TV July 2017

Meet IARA: UFES Autonomous Car Interview

In-depth interview about IARA, the autonomous car developed at UFES, featuring details about the technology and research behind it.

Bloomberg article: Amazon Picking Challenge
Newspaper May 2025

Amazon Seeks Robots With Human Grasp From Students (Amazon Picking Challenge — Rutgers/UFES Team, 7th Place)

Amazon.com Inc. attracted teams from 31 universities for its first "Amazon Picking Challenge" at the International Conference of Robotics and Automation in Seattle. Each team’s robot tried to pick up a shopping list of items of varying shapes and sizes off of shelves and place them in a bin. Amazon's Kelly Cheeseman and Rutgers University’s Alberto De Souza spoke at the event on Wednesday. (Source: Bloomberg)

Watch video
News July 2014

Autonomous Car Completes "Volta da UFES"

The UFES autonomous car project achieved a major milestone: completing a full lap around the Goiabeiras campus. Around midnight, on the 30th, the car traveled 3.8 km of the ring road surrounding the campus without any human interference — the longest distance ever traveled.

Read full article (UFES)
Magazine November 2013

Driverless Car - FAPESP Report

Brazilian autonomous vehicle projects bring contributions to the future of urban mobility. Featured in FAPESP research magazine (November 2013).

Read the full report (PDF)
A Tribuna newspaper article: Factory of Inventions at UFES
Newspaper May 2013

Factory of Inventions at UFES

A Tribuna newspaper full-page feature on the High-Performance Computing Laboratory (LCAD), the autonomous car IARA, facial recognition camera, robotic arm, and the DARPA Virtual Robotics Challenge.

View original (PDF)
News April 2013

UFES Group Creates Facial Recognition System

With 99.3% accuracy, the system uses neural network techniques and is already one of the best in the world. The technology could be used by police for identification purposes.

Self-Driving Car on Jornal Hoje
TV April 2013

Self-Driving Car on Jornal Hoje

Professors and students created a car that drives itself. The driver is a computer. Featured on Globo's Jornal Hoje.

Watch video
A Tribuna newspaper article: Students and Professors in ES Develop Self-Driving Car
News April 2013

Students and Professors in ES Develop Self-Driving Car

The vehicle project was developed by students and professors at UFES. The estimated time to acquire such a car commercially: 10 to 15 years.

Read full article (G1)
A Tribuna newspaper article: UFES creates a driverless car
Newspaper July 2012

UFES creates a driverless car

Researchers at the university are developing their own technology. The first tests are expected to begin next month.

A Gazeta article: Robot That Sees and Recognizes People
Newspaper February 2005

Robot That Sees and Recognizes People

A robot that can see objects and recognize people similarly to humans. The team, led by Professor Alberto Ferreira, is developing technologies to create a robot-forklift with visual capabilities. "It will have the capacity to create an internal image from its surroundings, similar to how the brain interprets the world through vision."

A Gazeta article: Enterprise - Brazil's Most Powerful University Cluster Supercomputer
Newspaper August 2004

Enterprise: Brazil's Most Powerful University Cluster Supercomputer

A supercomputer system assembled from 65 interconnected microcomputers, working together as one machine. Known as "Enterprise," it's considered the most powerful cluster operating in a Brazilian university. According to Prof. Alberto Ferreira, PhD in Computer Science, "We're working so that in 20 years we can build cars that drive themselves."

A Gazeta article: Enterprise - One of the 50 Fastest Supercomputers in the World
Newspaper June 2003

Enterprise: One of the 50 Fastest Supercomputers in the World

After seven months of work, UFES researchers created Enterprise — the fastest supercomputer in Brazil and one of the 50 fastest in the world. With 65 Athlon XP 1800+ machines, it can execute in 72 hours programs that would take 70 days on a regular PC. Maximum performance: 204 Gigaflops per second.

Get in Touch

Interested in collaboration, research opportunities, or have questions about our work? Feel free to reach out.