Antonio Roberto
roberto.antonio@outlook.it
As a Applied Scientist at Amazon, I bring extensive experience in the field of Artificial
Intelligence and a passion for solving complex problems. With a PhD in AI applied to social robotics, I
specialize in Deep and Machine Learning algorithms applied to Audio and Language analysis, and have
successfully deployed these algorithms in real applications.
My approach to problem-solving is characterized by a deep level of organization and precision, and a
willingness to dive deeply into the challenge at hand to find the best possible solution within the
appropriate timeframe. In addition to my technical expertise, I also possess strong leadership skills, honed
through experience guiding small groups of students.
My career objective is to continue to grow as a scientist in the field of Artificial Intelligence, with a
particular focus on real-world applications. I am passionate about leveraging the latest AI technologies to
drive innovation and make a positive impact in society.
Experience
Applied Scientist
Working as Applied Scientist at Amazon Alexe Devices.
Machine Learning Researcher
Design and implementation of Anomaly Detection and Root Cause Analysis algorithms aimed at the development
of Autonomous Networks.
Collaboration with customers to develop Networking Monitoring Standards (IETF).
Applied Scientist Internship
Working as applied scientist for the development of Speech-to-Text deep learning algorithms to deploy on board of the Alexa's devices. Data-driven model design based on large-scale databases.
Visiting Researcher
Research project: "Speech analysis for Speaker Identification and Soft-Biometrics recognition based on Deep Learning methods". Collaboration with the IMAGE team of the GREYC laboratory.
Research grant
Research grant for developing deep learning algorithms for Sound Event Detection.
Erasmus research experience
Erasmus period in collaboration with the Intelligent Systems research group on the topic "Financial time series forecasting".
Full-stack Developer
Development of the front-end and the back-end of cross-platform mobile applications for Android and iOS.
Publications
Identity, Gender, Age, and Emotion Recognition from Speaker Voice with Multi-task Deep Networks for Cognitive Robotics
Cognitive Computation, Springer US
A social robot architecture for personalized real-time human-robot interaction
IEEE Internet of Things Journal, IEEE
DEGramNet: Effective audio analysis based on a fully learnable time-frequency representation
Neural Computing and Applications, Springer
Few-shot re-identification of the speaker by social robots
Autonomous Robot, Springer
Efficient Transformers for On-Robot Natural Language Understanding
2022 IEEE-RAS 21st International Conference on Humanoid Robots (Humanoids)
Predicting Polypharmacy Side Effects Through a Relation-Wise Graph Attention Network
Joint IAPR International Workshops on Statistical Techniques in Pattern Recognition (SPR) and Structural and Syntactic Pattern Recognition (SSPR)
Which are the factors affecting the performance of audio surveillance systems?
25th International Conference on Pattern Recognition (ICPR)
DENet: a deep architecture for audio surveillance applications
Journal of Neural Computing and Applications - Springer
A deep convolutionary network for automatic detection of audio events
3rd International Conference on Applications of Intelligent Systems
Emotion analysis from faces for social robotics
IEEE International Conference on Systems, Man and Cybernetics (SMC)
A Challenging Voice Dataset for Robotic Applications in Noisy Environments
International Conference on Computer Analysis of Images and Patterns
Projects
ANTAGONIST - Time Series Tagging
AnTagOnIst (Anomaly Tagging On hIstorical data) is a tool that supports the visual analysis and the tagging of anomalies on telemetry data. This is done by providing a user-friendly interface to "Tag" anomalous data on multiple telemetry metrics and produce some metadata reflecting the semantic of those anomalies.
Social Robotics application
Design and development of a Social Robotic application to be used in a National fair. Spoken Language Understanding, Dialogue Management, Soft-Biometrics Recognition, People Tracking at edge on a NVIDIA Jetson Xavier NX embedded system.
Open-Set One-Shot Face-Recognition
Python implementation of a Face Recognition systems working with just ONE image for each face to recognize. The system works in an open-set configuration, it means that it is able to reject not known people.
Real Time Sound Event Detection
Python implementation of a Sound Event Detection systems working in real time.