Abolfazl Asudeh 🔊

Associate Professor, Department of Computer Science, University of Illinois Chicago

A. Asudeh is an Associate Professor of Computer Science at the University of Illinois Chicago and the director of Innovative Data Exploration Laboratory (InDeX Lab).

His research focus is on Algorithm Design for Data and AI problems. He designs efficient, accurate, and responsible solutions that leverage Approximation, Randomized, and Computational Geometry Algorithms. His research interests include the interplay of Data/Algorithms/AI, Responsible Data Science, and Ranking and Approximate Nearest Neighbor Problems, among others.

Asudeh is a senior member of ACM and a senior member of IEEE. He serves as an Associate Editor for the IEEE Transactions on Knowledge and Data Engineering (TKDE), is a regular PC member of Data Management and AI flagship venues, and served as a VLDB Ambassador and the VLDB Endowment's Liaison to NSF.

His research is supported by NSF and has received recognitions, including the Communications of the ACM's (CACM) Research Highlight, Google's Research Scholar Award, SIGMOD 2019 Research Highlight Award, “Best of VLDB” 2020 (the special issue of VLDBJ), and SIGMOD 2017 Reproducibility Award.

Active Projects

Current lab directions.

Selected Publications See Publications for the complete list.

Fair-Count-Min: Frequency Estimation under Equal Group-wise Approximation Factor.
Data-structureFairness
Nima Shahbazi, Stavros Sintos, A. Asudeh.
In SIGMOD (2026)
On Fair Epsilon Nets and Geometric Hitting Sets.
Data-structureAlgorithmsFairness
Mohsen Dehghankar, Stavros Sintos, A. Asudeh.
In VLDB (2026)
Dynamic Necklace Splitting.
[Slides]Algorithms
Rishi Advani, A. Asudeh, Mohsen Dehghankar, Stavros Sintos.
In ICDT (2026)
An Adversary-Resistant Multi-Agent LLM System via Credibility Scoring.
[Code] [Slides] [Video] LLM
Sana Ebrahimi, Mohsen Dehghankar, A. Asudeh.
In AACL, 2025
An Efficient Matrix Multiplication Algorithm for Accelerating Inference in Binary and Ternary Neural Networks.
[Code] [Video] LLMAlgorithmsMachineLearning
Mohsen Dehghankar, Mahdi Erfanian, A. Asudeh.
In ICML 2025.
Rank It, Then Ask It: Input Reranking for Maximizing the Performance of LLMs on Symmetric Tasks.
[Slides][Code] LLMMachineLearning
Mohsen Dehghankar, A. Asudeh.
In KDD 2025.
Mining the Minoria: Unknown, Under-represented, and Under-performing Minority Groups.
[Slides][Code] Responsible DataData-centric AI
Mohsen Dehghankar, A. Asudeh.
In VLDB, 2025.
Fair Set Cover.
[Slides][Code] AlgorithmsFairness
Mohsen Dehghankar, Rahul Raychaudhury, Stavros Sintos, A. Asudeh.
In KDD, 2025.
Evaluating the Feasibility of Sampling-Based Techniques for Training Multilayer Perceptrons.
[Slides] [Code] Computation at ScaleMachine Learning
Sana Ebrahimi, Rishi Advani, A. Asudeh.
In EDBT, 2025
A Survey on Reliability, Transparency, Accountability, and Fairness in LLM-based Multi-Agent Systems through the Responsibility Lens.
LLM Agents Responsible AI
Sana Ebrahimi and A. Asudeh.
Preprint (2025)
Efficient Direct-Access Ranked Retrieval.
Ranking
Mohsen Dehghankar, Raghav Mittal, Suraj Shetiya, A. Asudeh, Gautam Das.
Preprint (2025)
Unbiased Binning: Fairness-aware Attribute Representation.
FairnessResponsible Data
A. Asudeh, Mila Asoodeh, Bita Asoodeh, Omid Asudeh.
Preprint (2025)
HENN: A Hierarchical Epsilon Net Navigation Graph for Approximate Nearest Neighbor Search.
[Code]Data-structureAlgorithms
Mohsen Dehghankar and A. Asudeh.
Preprint (2025)
Needle: A Generative-AI Powered Monte Carlo Method for Answering Complex Natural Language Queries on Multi-modal Data.
[Code] Image DatabasesLLM4DataMachineLearning
Mahdi Erfanian, Mohsen Dehghankar, A. Asudeh.
Preprint (2024)

Pinned Systems and Repositories

Needle

Needle is a deployment-ready open-source image retrieval database with high accuracy for complex natural-language queries. It is fast, efficient, and precise, outperforming state-of-the-art methods while staying accessible to researchers, developers, and practitioners.

Visit Project View on GitHub Read Publication Watch Video

Detailed installation instructions: Getting Started.

RSR: Efficient Matrix Multiplication for Low-bit Neural Networks

RSR provides a fast approach to low-bit matrix multiplication for binary and ternary neural networks. The repository includes C++, NumPy, and PyTorch implementations with CPU and GPU support, plus sample experiments on 1.58-bit models and LLMs.

View on GitHub Read Publication Watch Video

Prospective Students

Every year, I look for a small number of motivated, smart, and hard-working PhD students. In particular, I look for students with a strong background in Algorithms who are interested in finding algorithmic solutions for data problems and AI (recently LLMs and Foundation Models have enabled an exceptional phenomenon for such research). You also need to have outstanding programming skills. If you are interested, check our lab page and our recent publications. If interested and qualified, please apply here and mention my name in your application.
Due to the large number of emails, I may not be able to respond to them all. If you email me, please use ``Prospective PhD Student'' as the subject and attach your CV and transcripts. Describe your skills, interests, research experience, and how those may fit InDeX Lab.