Readspeaker menu

PDEng Wasim Ahmad

Research Associate
Wasim Ahmad


Address: Computer Vision Group
Department of Mathematics and Computer Science
Friedrich Schiller University of Jena
Ernst-Abbe-Platz 2
07743 Jena
Phone: +49 (0) 3641 9 46427
Room: 1221

Curriculum Vitæ

 since May 2020
  PhD Student
Computer Vision Group Jena, Friedrich Schiller University Jena, Germany
Research Topic: Causal Inference using Deep Learning
 2017 - 2019
  Professional Doctorate in Engineering (PDEng)
Dynamics based Maintenance (DBM) Group, University of Twente, Enschede Netherlands
STRUKTON RAIL, Utrecht Netherlands
PDEng Project: Artificial Intelligence based Condition Monitoring of Rail Infrastructure
 2015 - 2017
  M.Sc. Computer Engineering
Ulsan Industrial Artificial Intelligence (UIAI) Laboratory, University of Ulsan, South Korea
Master Thesis: Prognosis of Induction Motors using Machine Learning & Digital Signal Processing Techniques
 2008 - 2012
  B.Sc. Information and Communication Systems Engineering
2008 to 2012 School of Electrical Engineering and Computer Science (SEECS), National University of Sciences and Technology (NUST), Pakistan
Bachelor Thesis: Next Generation Resource Tracking (XRT) Using RFID Technology

Research Interests

  • Applied Deep Learning / Machine Learning
  • Causal Reasoning
  • Data-driven Predictive Maintenance


Causal Inference using Deep Learning

The aim of the project is to investigate to what extent causal relationships in multivariate nonlinear dynamical systems can be derived from learned, deep networks. Estimating causal relations is vital in understanding the complex interactions in multivariate time series. Non-linear coupling of variables is one of the major challenges in accurate estimation of cause-effect relations. We propose to use deep autoregressive networks (DeepAR) in tandem with counterfactual analysis to infer nonlinear causal relations in multivariate time series. In this work, we extend the concept of Granger causality using probabilistic forecasting with DeepAR. We achieved promising results using our method in detecting nonlinear causal dependency in multivariate time series from synthetic as well as real-world datasets.