

Negar Arabzadeh
Postdoctoral Fellow at UC Berkeley
This is Negar Arabzadeh. I am a postdoctoral researcher at UC Berkeley, working with Professor Matei Zaharia. I completed my PhD at the University of Waterloo, where I was advised by Dr. Charles Clarke. My research lies at the intersection of information retrieval and the evaluation of—and with—large language models (LLMs) in information access systems.
News
April 20245
📚Four papers got accepted in SIGIR:
-
Benchmarking LLM-based Relevance Judgment Methods
-
w/ Charles Clarke
-
-
VAP3: Variation-Aware Prompt Performance Prediction
-
w/ Ebrahim Bagheri
-
-
A Human-AI Comparative Analysis of Prompt Sensitivity in LLM-Based Relevance Judgment
-
w/ Charles Clarke
-
-
IDAT: A Multi-Modal Dataset and Toolkit for Building and Evaluating Interactive Task-Solving Agents
-
w/ Shrestha Mohanty, Andrea Tupini, Yuxuan Sun, Alexey Skrynnik, Artem Zholus, Marc-Alexandre Cote and Julia Kiseleva
-
-
December 2024
✈️I attended SIGIR-AP in Tokyo,Japan.
-
We presented two papers and a tutorial.
✈️I attended ACML in Hanoi,Vietnam.
-
We presented two papers.
📚Two papers got accepted in ECIR:
-
exHarmony: Authorship and Citations for Benchmarking the Reviewer Assignment Problem
-
w/Sajad Ebrahimi, Sara Salamat, Mahdi Bashari and Ebrahim Bagheri
-
-
Benchmarking Prompt Sensitivity in Large Language Models
-
w/ Amirhosein Razavi, Mina Soltangheis, Sara Salamat, Morteza Zihayat and Ebrahim Bagheri
-
November 2024
✈️I attended my first EMNLP in Miami,USA.
-
WOW it was huge!
October 2024
✈️I attended CIKM in Boise, Idaho.
-
I gave my first industrial talk about Reviewer.ly.
📚 Our paper "Assessing and Verifying Task Utility in LLM-Powered Applications" was accepted at EMNLP 2024.
W/ Siqing Huo, Nikhil Mehta, Qingyun Wu, Chi Wang, Ahmed Hassan Awadallah, Charles L. A. Clarke and Julia Kiseleva
September 2024
📚 Our paper "Offline Evaluation of Set-Based Text-to-Image Generation " was accepted at SIGIR-AP 2024.
W/ Fernando Diaz and Junfeng He
📚 Our paper " Evaluating Relative Retrieval Effectiveness with Normalized Residual Gain" was accepted at SIGIR-AP 2024.
W/ Amin Bigdeli, Charles L.A. Clarke and Ebrahim Bagheri.
📚 Our tutorial "Query Performance Prediction: Techniques and Applications in Modern Information Retrieval" was accepted at SIGIR-AP 2024.
W/ Chuan Meng, Mohammad Aliannejadi nd Ebrahim Bagheri.