Arash Akbari

I am a second-year PhD student in Computer Engineering at Northeastern University, under the supervision of Professor Yanzhi Wang. My research focuses on efficient deep learning and foundation models. I hold a Bachelor's degree in Computer Science from the University of Tehran.

Previously, I interned at CNRS with Timothée Masquelier and at CISPA with Xiao Zhang.

Email  |  GitHub  |  Google Scholar  |  LinkedIn

Research Focus: I develop efficient machine learning techniques to make state-of-the-art models more practical for real-world deployment. My work focuses on optimizing foundation models such as VLA models for scalability and deployment efficiency across diverse hardware platforms.

headshot
News
Research

Preprints
VOTE

VOTE: Vision-Language-Action Optimization with Trajectory Ensemble Voting

Juyi Lin, Amir Taherin, Arash Akbari*, Arman Akbari, Lei Lu, Guangyu Chen, Taskin Padir, Xiaomeng Yang, Weiwei Chen, Yiqian Li, Xue Lin, David Kaeli, Pu Zhao, Yanzhi Wang

Preprint 📄 PDF 💻 Code

We propose VOTE, an efficient and general framework for the optimization and acceleration of VLA models. VOTE is a novel tokenizer-free fine-tuning approach for parallel accurate action prediction, which reduces computational overhead and accelerates inference speed. Our method achieves state-of-the-art performance with 35 times faster inference and 145 Hz throughput.

Published Papers
Fairness Analysis

Beyond Overall Accuracy: Pose- and Occlusion-driven Fairness Analysis in Pedestrian Detection for Autonomous Driving

Mohammad Khoshkdahan, Arman Akbari, Arash Akbari*, Xuan Zhang

IEEE ITSC 2025 📄 PDF

In this work, we systematically investigate how variations in the pedestrian pose—including leg status, elbow status, and body orientation—as well as individual joint occlusions, affect detection performance. We evaluate five pedestrian-specific detectors alongside three general-purpose models on the EuroCity Persons Dense Pose dataset.

Grasping Algorithm

A 2D Geometry Based Grasping Pose Generation Algorithm for a Two-Finger Robot Hand

Arash Akbari*, Arman Akbari, Mehdi Tale Masouleh

ICEE 2023 📄 PDF

A geometry-based algorithm is presented which can find grasp poses based on the geometry of the unknown object and propose the ones which may lead to successful grasping. The algorithm produces key points based on the 2D shape of the object and outputs successful grasp poses based on three grasp quality metrics.

Academic Services
  • 2026 Spring: Fundamentals of Computer Engineering, Teaching Assistant.

template adapted from this awesome website