Filippo Maria Bianchi Homepage

I am a post-doctoral fellow at the Machine Learning Group at UiT the Arctic University of Norway, Tromsø.

My research activities in machine learning are focused on deep learning in recurrent neural networks and autoencoders, time series analysis and reservoir computing. I also worked with clustering and graph matching for pattern recognition and data analysis. The main applications I worked with are the prediction of load (electricity, traffic, telephonic activity), time series classification, analysis of call data records and health records.

Main Research Interests

  • Echo State Networks - theoretical studies on internal dynamics; design of novel training methods; applications in prediction and system identification.
  • Recurrent Neural Networks - design of new architectures; application to load forecast and time series classification.
  • Autoencoders - learning representations with kernel methods; handling missing data; multivariate time series encoding.
  • Time Series Analysis - identification of underlying system dynamics; state space reconstruction; prediction.
  • Graph Matching - representation of unstructured data with graphs; application to granular computing techniques; image classification.
  • Clustering - density-based clustering on unstructured data; agent-based clustering; multi-metric learning to cluster characterization.

You can follow my research papers on ResearchGate and Google Scholar.


Call for Papers: Special Issue on "Non-Iterative Learning Approaches and Their Applications" - Cognitive Computation, Springer. [CfP]