Nirupam Gupta

E-mail: nirupam [dot] gupta [at] georgetown [dot] edu
Google Scholar Profile

Nirupam Gupta

Currently, I am a "post-doctoral research fellow" in the department of computer science at Georgetown University, where I work with Prof. Nitin H. Vaidya on Byzantine fault tolerance in distributed optimization and learning.

Education: I got my PhD in mechanical engineering from the University of Maryland - College Park in 2018, under the guidance of Prof. Nikhil Chopra. During my PhD, I worked on privacy and resilience in distributed multi-agent collaboration algorithms, such as distributed consensus and optimization.

I got my bachelors in electrical engineering from the Indian Institute of Technology Delhi in 2013.

Ph.D. dissertation: Privacy in distributed multi-agent collaboration: consensus and optimization

Area of Interest: I work on distributed optimization (+ consensus) algorithms that are resilient to Byzantine faults and ill-conditioning.

Publications & Ongoing Projects:

  1. Byzantine Fault Tolerant Distributed Linear Regression (arXiv , pdf)
    with Nitin Vaidya
    [In Preparation]
  2. Privacy of Agents' Costs in Peer-to-Peer Distributed Optimization (arXiv)
    with Nikhil Chopra
    [To be submitted].
  3. Model-Based Encryption in Networked Control Systems
    with Nikhil Chopra
    [Under Review] Transactions on Automatic Control (IEEE) .
  4. Information-Theoretic Privacy in Distributed Average Consensus: Bounded Integral Inputs (arXiv)
    with Jonathan Katz, Nikhil Chopra
    [Under review] Automatica .
  5. Statistical Privacy in Distributed Average Consensus on Bounded Real Inputs (arxiv)
    with Jonathan Katz, Nikhil Chopra
    American Control Conference (2019, IEEE).
  6. False Data Injection Attacks in Bilateral Teleoperation Systems
    with Yimeng Dong, Nikhil Chopra
    Transactions on Control Systems Technology (2019, IEEE).
  7. Model-Based Encryption: Privacy of States in Networked Control Systems (pdf)
    with Nikhil Chopra
    Allerton Conference at University of Illinois at Urbana-Champaign (2018, IEEE).
  8. Privacy in Distributed Average Consensus (pdf)
    with Jonathan Katz, Nikhil Chopra
    IFAC World Congress (2017, Elsevier).
  9. Robustness of distributive double-integrator consensus to loss of graph connectivity
    with Yimeng Dong, Nikhil Chopra
    American Control Conference (2017, IEEE).
  10. Content modification attacks on consensus seeking multi-agent system with double-integrator dynamics
    with Yimeng Dong, Nikhil Chopra
    Chaos: An Interdisciplinary Journal of Nonlinear Science (2016, AIP).
  11. Confidentiality in distributed average information consensus
    with Nikhil Chopra
    Conference on Decision and Control (2016, IEEE).
  12. On content modification attacks in bilateral teleoperation systems
    with Yimeng Dong, Nikhil Chopra
    American Control Conference (2016, IEEE) .
  13. Stability analysis of a two-channel feedback networked control system
    with Nikhil Chopra
    Indian Control Conference (2016, IEEE).

Other Technical Reports:

  1. One-Time Pad For Real Values
    Description: An equivalent of the well-known one-time pad encryption scheme for (bounded) real values using the concept of fractional part.
  2. Distributed Computation of Continuous Functions Using Distributed Average Consensus
    Description: It is well-known that the arguments of sigmoidal functions -- superimposed for universal approximation of functions -- in neural networks are weighted linear combination of the inputs. Thus, this gives a natural extension for distributed average consensus techniques for computing any function asynchronously over a jointly-connected graph(network).
  3. Solving System of Linear Equations With Uncertain Coefficients
    Description: Proposed a convex relaxation (relies on solution of convex optimization problem) for determining the solution set of system of linear equations with coefficients belong to a bounded real-valued intervals.
  4. Nearest Neighbor Based Heuristic Solution to Optimization Problemes with Boolean Constraints
    Description: Proposed a heuristic algorithm based on the nearest neighbor search approach for solving a non-convex robust linear programming with boolean constraints, and compared our proposed algorithm against the branch and bound algorithm.
  5. Switched Autonomous Dynamic Systems
    Description: Studying design of switching controllers for system with switching dynamics and guarantee stability using multiple Lyapunov functions (MLFs).