Arash NasrEsfahani

About Me

I am Arash Nasr Esfahani, a Master’s student in Computer Science at the University of Tehran, specializing in Artificial Intelligence.

I enjoy discovering patterns in human and machine behavior to develop insights that can meaningfully impact people's daily lives. My primary research interests include Human-Computer Interaction (HCI), Human-Robot Interaction (HRI), and Machine Learning. I am part of the Human and Robot Interaction Lab at the University of Tehran, directed by Prof. M. Tale Masouleh, where I have been conducting research for the past two years. Within our group’s work on robotic scene rearrangement, my focus has been on understanding how interaction with an environment can serve as a step toward interpreting and effectively rearranging it.

I was also part of the University of Tehran’s team for the SHL 2025 Challenge, supervised by Prof. B. Babaali. In this project, we explored the application of time-series foundation models for Human Activity Recognition, which resulted in a paper accepted to UbiComp 2025.

Publications

Ensemble of Foundation Models for Sensor-Based Locomotion and Transportation Mode Recognition

Arash Nasr Esfahani, Mohammad Foad Abdi, Yousef Alikhani, Mohammad Mahdi Azizi, Bagher BabaAli, Mehdi Mohhebi

Accepted in UbiComp 2025: ACM International Joint Conference on Pervasive and Ubiquitous Computing

AI-Driven Relocation Tracking in Dynamic Kitchen Environments

Arash Nasr Esfahani, Hamed Hosseini, Mehdi Tale Masouleh, Ahmad Kalhor, Hedieh Sajedi

Published in ICCKE 2024: 14th International Conference on Computer and Knowledge Engineering

Embodied AI for Kitchen Scene Rearrangement: A Deep Spatio-Temporal Approach

Arash Nasr Esfahani, Hamed Hosseini, Mehdi Tale Masouleh, Ahmad Kalhor

Manuscript submitted to peer-reviewed journals

Research Experience

Researcher, The Human and Robot Interaction Lab (TaarLab)

University of Tehran | Sep 2023 – Present

Led a project on robotic scene rearrangement in dynamic environments using the AI2-THOR simulator, developing methods integrating deep learning models for object detection and temporal frame analysis.

Researcher, School of Math, Statistics, and Computer Science

University of Tehran | Apr 2025 – Jul 2025

Collaborated in a five-person team under Prof. B. Babaali for the SHL 2025 Human Activity Recognition Challenge, where our team earned 1st place at the HASCA Workshop, UbiComp 2025, by applying time series foundation models and advanced data analysis techniques.

Teaching Experience

Teaching Assistant, Database Management Systems

University of Tehran | Oct 2023 – Jan 2024

Contributed to developing course materials and supervised student projects, providing guidance on database design principles.