Dario J. Dematties

INCIHUSA CCT CONICET. Av. Ruiz Leal s/n - Parque Gral. San Martín, M5500. (+54) 261 5244348. Mendoza. Argentina.

(+54) 261 5244348

Av. Ruiz Leal s/n - Parque Gral. San Martín

Mendoza, 5500

Dr. Dario Dematties, received his bachelor’s degree in electronic engineering from the National Technological University (UTN) in 2013 in Mendoza, Argentina. Then in 2020 he received his Ph.D. degree at the Institute of Biomedical Engineering at the University of Buenos Aires.

During his Ph.D., Dr Dematties worked to identify neuroanatomical and physiological features in mammalian cortical tissue that could leverage phonetic perception invariance and generalization as well as higher linguistic functions such as grammar and semantic sentence classification in biologically inspired computational models. Dr. Dematties used a computational approach completely unsupervised that incorporated key neurophysiological and anatomical properties that exist in the brain cortex.

He runs his models on Cooley (a visualization and analysis cluster) and on Theta GPUs (a supercomputer for training artificial intelligence (AI) datasets) both at Argonne National Laboratory. In recent years, he has focused on High-Performance Computing (HPC) and code optimizations for his computational approach.

Dr. Dematties has currently a postdoctoral position at INCIHUSA CCT CONICET where he uses Deep Learning Frameworks for the implementation of self-supervised contrastive learning of visual representations through active foveated saccades. Dr. Dematties’ current research interests are on brain-inspired Machine Learning, Experience Grounds Language, Multimodal Intelligence and on the implementation of complex optimization methods (i.e. backpropagation) on biologically plausible neural networks.

selected publications

  1. FNCir
    A Computational Theory for the Emergence of Grammatical Categories in Cortical Dynamics
    Frontiers in Neural Circuits 2020
  2. ParCo
    Towards High-End Scalability on Biologically-Inspired Computational Models
    In Advances in Parallel Computing, Parallel Computing Conference (ParCo) 2020
    Phonetic acquisition in cortical dynamics, a computational approach
    PLOS ONE 2019