Farbod Siahkali

Farbod Siahkali

PhD student in Electrical Engineering at Purdue University, focusing on conformal prediction, control theory, machine learning, and optimization.

About Me

I am Farbod Siahkali, a Ph.D. student in Electrical Engineering, specializing in Automatic Control at Purdue University. I earned my Bachelor’s degree in Electrical Engineering (Control Engineering Branch) from the University of Tehran. Currently, I serve as a Graduate Research Assistant in Prof. Vijay Gupta’s Lab at Purdue, where my research focuses on applying conformal prediction to multi-view cooperative perception, trajectory optimization, and domain adaptation.

My academic interests center on the convergence of conformal prediction, control theory, machine learning, and optimization. For more details on my work, please visit my Google Scholar profile or contact me via email.

Publications

My scholarly contributions include:

  • Active Sensing and Deferred-Decision Trajectory Optimization for Robust Target Identification (2026, published in IEEE Control Systems Letters, available on IEEE Xplore)
    Develops an active-sensing extension of deferred-decision trajectory optimization for robust target identification under sensing uncertainty and resource constraints.

  • Conformal prediction-based multi-sensor classification (2026, published in Neurocomputing, article 133182, available on ScienceDirect)
    Studies conformal prediction methods for reliable multi-sensor classification with uncertainty-aware decision support.

  • Coverage Guarantees for Pseudo-Calibrated Conformal Prediction under Distribution Shift (2026, available on arXiv)
    Provides coverage guarantees for pseudo-calibrated conformal prediction under bounded distribution shift.

  • Deep Reinforcement Learning for Epidemic Control of a Networked SIS Model (September 2025, published in IEEE Conference on Control Technology and Applications (CCTA) 2025, available on IEEE Xplore)
    Proposes a novel approach to reduce infection spread in a networked susceptible-infected-susceptible (SIS) model using deep deterministic policy gradient (DDPG).

  • Towards Opinion Shaping: A Deep Reinforcement Learning Approach in Bot-User Interactions (September 2024, available on arXiv)
    Explores the impact of interference in social networks using deep reinforcement learning.

  • Image-Based and Partially Categorical Annotating Approach for Pedestrian Attribute Recognition (July 2023, available on SSRN)
    Introduces the CA-Duke dataset and a two-step learning method with a novel Separation Index metric.

  • SIVD: Dataset of Iranian Vehicles for Real-Time Multi-Camera Video Tracking and Recognition (December 2022, published in International Conference on Signal Processing and Intelligent Systems (ICSPIS) 2022, available on IEEE Xplore)
    Presents a new dataset with 29 classes and over 36,000 images for vehicle tracking and recognition.