Machine Learning and Big Data together offer a universal way of looking at the world phenomena, which is radically different than the classical expert based disciplinary research.
This new approach of computational modeling has inverted the classical notion of expertise from “having the answers to the known questions” to “learning to ask good questions”, where the answers can always be found with an appropriate level of modeling skills.
I finished my PhD thesis at the Chair for Computer Aided Architectural Design (CAAD), ETH Zurich by the end of April 2015. Here, you can take a look at my final PhD presentation. link
From November 2011-April 2015 I worked as a researcher in Singapore-ETH lab, Future Cities Laboratory.
Previously, I studied in Iran and received my B.S and M.S degrees in industrial and systems engineering from Isfahan University of Technology and Tehran Polytechnic respectively.
From May 2015, I am a senior researcher at the chair for Computer Aided Architectural Design (CAAD), ETH Zurich.
In general, I am fan of data driven modeling, but the main question is that what are the appropriate (new) methods for dealing with this new forms of data (usually unstructured and Big). In this regard, I am interested in the following research directions:
- Representation and Encoding in the context of system modeling.
- Unsupervised learning and specially Self Organizing Maps for dimensionality reduction and multivariate probability estimations.
- Representation Learning with Deep Neural Networks
- Invariances in random processes and Markov Chains
- Function approximation and manifold learning methods
Further, as I am originally a systems engineer I have been always dreaming about finding unity through diversity of different application domains. Therefore, I have been always eager to be engaged in as diverse as possible fields of applied problems. So far I have experienced several specific problems in the following domains:
- Spatio-Temporal Modeling:
- Study of urban forms at the global scale: More than 1 million locations across the globe
- Data driven air pollution level estimation at the global scale
- Fast and scalable urban flood risk emulation
- Real estate market dynamics
- Urban traffic modeling
- Urban air pollution estimation
- Urban economy and business activity level estimation
- Networked based economics and systemic risk
- Applied Natural Language Modeling
- Training financial time series models along with the news time series
- Developing smart personalized news papers using news API data streams
- Financial time series forecasting
Research projects for graduate students or those who are interested to collaborate with me:
If you are interested in machine learning and data driven modeling applications, you can contact me. Distance does not matter even if you are not based in Zurich. At the moment I am working on the following topics:
- Air Pollution Modeling at the global scale using data from more than 8000 locations
- Multidimensional urban exploration: combination of different modes of data including structured data such as census data in US and other unstructured data sets such as satellite images and graphical data from Open Street Map (OSM)
- Learning Physics: How to speed up slow physics-based simulation models (e.g. CFD models in urban wind or water flow models) via techniques such as Recurrent Neural Networks and Convolutional Neural Nets.
- Architectural indexing of millions of buildings from Open Street Map
- Remote sensing imagery, deep learning and informal settlements
- Data driven traffic Simulation
- Systemic risk in energy networks