Ehsan Lari profile picture

Ehsan Lari, Ph.D.

About Me

Hi! My name is Ehsan, but you can also call me Ethan!

I got my PhD from the Department of Electronic Systems (IES) at the Norwegian University of Science and Technology (NTNU) in Trondheim, Norway.

My research is centered on the application of signal processing techniques to enhance the privacy, security, robustness, and efficiency of machine learning algorithms.

Interests

My main interests are:

• Distributed and Federated Machine Learning

• Applications of Signal Processing in Machine Learning

• Privacy-Preserving Techniques

• Reliable Uncertainty Quantification using Conformal Prediction Methods

• Bayesian Approaches to Learning and Inference

• Causal Discovery and Inference in Machine Learning

Projects

Norway Travel Guide

Cloud-deployed FastAPI and React application with automated validation and scoring

A travel decision platform that integrates real-time weather observations and forecast streams from MET Norway REST APIs with transit routing from Entur. Features automated schema validation, missing-data imputation, multi-criteria scoring algorithms, and AI-generated summaries using Gemini.

PSFed: Partial Model Sharing for Federated Learning

Open-source Python library published on PyPI with CI/CD automation

A modular Python library designed for distributed machine learning with selective parameter sharing. Reduces per-cycle network data transfer by 50% to 90% while guaranteeing convergence. Integrates with PyTorch and Flower. Maintained with automated GitHub Actions CI/CD workflows for testing (pytest), static typing (mypy), and linting (ruff).

Interactive Portfolio AI Assistant

Conversational assistant powered by Google Gemini and Vercel serverless functions

An interactive AI assistant embedded directly into this website to answer questions about academic background, research publications, and technical skills. Built with a responsive JavaScript interface, a Vercel serverless backend, and Google's Gemini API grounded on a curated CV and portfolio knowledge base.

PhD Thesis

Distributed Learning with Enhanced Efficiency, Robustness and Privacy PDF

E. Lari

Publications

Communication-Efficient Byzantine-Robust Federated Conformal Prediction via Partial Sharing

E. Lari, R. Arablouei, and S. Werner

IEEE Transactions on Signal Processing

Partial Model Sharing Improves Byzantine Resilience in Federated Conformal Prediction

E. Lari, R. Arablouei, and S. Werner

EUSIPCO 2026

Noise-Robust and Resource-Efficient ADMM-Based Federated Learning For WLS Regression

E. Lari, R. Arablouei, V. C. Gogineni, and S. Werner

Elsevier Signal Processing

Resilience In Online Federated Learning: Mitigating Model-Poisoning Attacks Via Partial Sharing

E. Lari, R. Arablouei, V. C. Gogineni, and S. Werner

IEEE Transactions on Signal and Information Processing over Networks
Show more publications

Privacy-Preserving Distributed Nonnegative Matrix Factorization

E. Lari, R. Arablouei, and S. Werner

EUSIPCO 2024

Distributed Maximum Consensus Over Noisy Links

E. Lari, R. Arablouei, N. K. D. Venkategowda, and S. Werner

EUSIPCO 2024

On The Resilience Of Online Federated Learning To Model Poisoning Attacks Through Partial Sharing

E. Lari, V. C. Gogineni, R. Arablouei, and S. Werner

ICASSP 2024

Continual Local Updates For Federated Learning With Enhanced Robustness To Link Noise

E. Lari, V. C. Gogineni, R. Arablouei, and S. Werner

APSIPA 2023

Resource-Efficient Federated Learning Robust To Communication Errors

E. Lari, V. C. Gogineni, R. Arablouei, and S. Werner

SSP 2023

Teaching