Department of Computer Science, Aalto University
I am a PhD researcher in machine learning and computational neuroscience at Aalto University, Finland, under the supervision of Stéphane Deny. My research explores how biological systems, especially the human brain, process visual information. Specifically, I focus on developing deep learning paradigms that more accurately capture the mechanisms of human visual perception. This pursuit is driven by the impetus to gain valuable insights into brain function, particularly through the lens of visual cognitive tasks such as mental rotation and the pop-out effect—whereby an outlier directly attracts attention.
Moreover, I aim to address the limitations of current deep learning models in terms of robustness and efficiency in comparison to human visual systems, while tackling the contemporary challenge of excessive data requirements for model training. To this end, I am investigating how concepts like symmetry (pose transformations) can enhance machine learning models and lead to more interpretable representations.
My hypothesis is that humans can factorize and compose, enabling compact representations by identifying structural patterns through latent symmetries.
Investigate how a machine learning model can learn visual primitives and apply symmetry transformations to form a compositional scene representation.
Investigate how neural networks can simulate human-like spatial reasoning under 3D transformations. Get deeper insights of the mental rotation process.