I am a master student of artificial intelligence in Barcelona and currently focused on exploring (inverse) reinforcement learning and generative models such as GANs. With a background in software engineering, I have developed a passion for research in machine learning and artificial intelligence. I am leading the open source project IRL benchmark.
I believe that AI has huge potentials for the future of humanity, but that a lot of directed work and research will be necessary to align advanced AI systems with our values and objectives. To this end, I am exploring methods of applying machine learning to the problem of value learning: observing human behavior and deducing what their motivations were.
IRL benchmark: A open source framework for testing different inverse reinforcement learning algorithms on a variety of different problems to compare their performance and robustness IRL benchmark repository .
AI Safety Camp: I was main organizer of the second AI Safety Camp in Prague. We created a retreat for almost 30 aspiring researchers to work on topics related to AI alignment AISC website .
Generative Adversarial Networks: Experimenting on the capabilities of using GANs to generate additional data for supervised learning with few available training instances Report and Code on GitHub .
Automated Planning: Implementation of the Stanford Research Institute Problem Solver (STRIPS) Code on GitHub .
Bachelor Thesis: Anomaly Detection on Time Series Data for Predictive Maintenance. Using the restoration error of neural autoencoders to detect deviations from normal behavior of industrial time series data. Available upon request.
Have a look at my blog about topics related to AI alignment: thinking wires