Albert Huizing
|
RESEARCH AREA
In recent years, Artificial Intelligence (AI) based on deep learning has achieved tremendous success in specialized tasks such as speech recognition, machine translation, and the detection of tumours in medical images. Despite these successes there are also some clear signs of the limitations of the current state-of-the-art in AI. For example, accidents with self-driving cars and unwanted biases in AI-enabled face recognition have shown that AI cannot yet be trusted in safety-critical or societally sensitive applications.
My ambition is to make AI more trustworthy and useful by initiating and supervising research that enables AI to evolve from a stand-alone tool to a reliable member of a human-machine team, from applications where the governance of the AI is permissive to applications where governance needs to be strict with respect to compliance with laws, ethical norms, and societal values, from operations in a controlled environment to operations in an open world, and from special purpose tasks to more general purpose problem solving. These improvements in AI will help our society to become more healthy, sustainable, competitive, and safe.
RECENT RESULTS
In the past 3 years, I have been responsible for defining the research lines of the Early Research Program Hybrid AI that is now part of the TNO-wide program on Applied AI (Appl.AI). Hybrid AI is a key enabler of trustworthy AI by combining two different paradigms in AI: knowledge-based reasoning and optimization, and data-driven machine learning. As a member of the Appl.AI management team, I am supervising the exploratory AI research within TNO and initiate partnerships with other research organisations such as the Fraunhofer Institute in Germany.
After a long career in radar, I have also been working in the past 5 years on radar applications of AI such as the recognition of human gestures with a low-power radar and a deep convolutional neural network, and the classification of drones with a cognitive radar.
TOP PUBLICATIONS
- Huizing, M. Heiligers, B. Dekker, J. de Wit, L. Cifola, R. Harmanny, "Deep Learning for Classification of Mini-UAVs Using Micro-Doppler Spectrograms in Cognitive Radar," in IEEE Aerospace and Electronic Systems Magazine, vol. 34, no. 11, pp. 46-56, 2019.
- J. M. Elands, A. G. Huizing, L. J. M. Kester, M. M. M. Peeters, and S. Oggero, “Governing ethical and effective behaviour of intelligent systems,” Militaire Spectator, vol. 188, no. 6, pp. 303–313, 2019.
- Dekker, S. Jacobs, A. S. Kossen, M. C. Kruithof, A. G. Huizing and M. Geurts, "Gesture recognition with a low power FMCW radar and a deep convolutional neural network," 2017 European Radar Conference (EURAD), Nuremberg, pp. 163-166, 2017.