RESEARCH and PROJECTS

My current research focuses on robust, reliable, and resource-efficient embedded AI solutions for real-time signal and image processing in resource-constrained settings, ranging from underwater acoustics to medical imaging. The constraints apply to both training (data collection and labeling) and inference (time, bandwidth, power, and memory), motivating research in physics-augmented AI, intelligent data collection and labeling, robust AI, and energy-efficient AI.  I use tools from statistical signal processing, machine learning, information theory, and acoustics.

My past research interests include adaptive filtering and underwater acoustic communications.

Since this is an ongoing project, we continue publishing the results in relevant conferences and journals. 

1. Sida Liang, Alex M. Condon, Adam A. Markowicz, Kayvan F. Tehrani, Hamza El-Kebir, Guillermo L. Monroy, Dariush Kari, Trung-Hieu Hoang, Edita Aksamitiene, Jessica J. Saw, Eugene Povrozin, Adrian R. Liversage, Arif Mohd Kamal, Jindou Shi, Mosbah Aouad, Anirudh Choudhary, Mark A. Anastasio, Minh Do, Ravishankar Iyer, Stephen A. Boppart,, ``Electrocautery-Driven Alterations in Tissue Optical Properties Assessed by Simultaneous Label-free Autofluorescence Multi-harmonic (SLAM) microscopy and Optical Coherence Tomography (OCT),'' In Label-free Biomedical Imaging and Sensing (LBIS) 2026, San Fransisco, CA, USA

2. Guillermo L. Monroy,  Hamza El-Kebir, Arif Mohd Kamal,  Kayvan F. Tehrani, Adrian R. Liversage, Alex Condon, Adam A. Markowicz, Sida Liang, Dariush Kari, Trung-Hieu Hoang, Jindou Shi, Mosbah Aouad, Anirudh Choudhary, Benjamin D. Havens, Mark A. Anastasio, Minh N. Do, Ravishankar Iyer, Miglena Komforti, Janani S. Reisenauer, James W. Jakub, Nikhil Muralidhar, Yongseok Lee, Dongshen Ye, Girish Krishnan, Joseph Bentsman, Chihab Nadri, Mojgan Zoaktafi, Kaitlin M. Skurnak, Gavin P. Bourgeois, Abigail R. Woodridge, M. Susan Hallbeck, Colleen B. Bannon Bushell, Kenton McHenry, Sarah N. Chronister, Christopher Pond, Chris Stephens, Mikolaj Kowalik, Lisa P. Gatzke, Jessica Saw, Lijiang Fu, Katherine Arneson, Matthew J. Berry, Bingji Guo, Christopher Havlin, Chad Olson, Rob Kooper, Chen Wang, Minu Mathew, Yong Wook Kim, Todd C. Nicholson, Rashmil Panchani, Thihan M. Kyaw, Eugene Povrozin, Edita Aksamitiene, Darold R. Spillman, Amy Abrahamson, Daniele L. Frerichs, Katherine M. Koch, Deborah S. Miller, Randall M. Cullen, Stephen A. Boppart, ``Development of label-free multi-modal multi-contrast optical imaging platform for in-vivo intraoperative surgical margin assessment,'' In Advanced Biomedical and Clinical Diagnostic and Surgical Guidance Systems XXIV, San Fransisco, CA, USA

Data-Driven Underwater Acoustic Localization

Underwater acoustics is known for challenges such as limited bandwidth, severe signal attenuation, multipath propagation, and variable channel conditions. Moreover, collecting underwater data from diverse environments is prohibitively expensive. This project focuses on enhancing the generalization performance of data-driven methods for underwater acoustic localization in the presence of mismatches between the training and test environments. The proposed methods combine model-based approaches with data-driven techniques to address challenges posed by environmental uncertainty, computational complexity, and data scarcity. 

1. Dariush Kari, Yongjie Zhuang, Andrew C. Singer, ``Mismatch-Robust Underwater Acoustic Localization Using A Differentiable Modular Forward Model'' In 2025 59th Annual Conference on Information Science and Systems (CISS), Baltimore, MD, USA, 2025, pp. 1-6

2. Dariush Kari, Hari Vishnu, Andrew C. Singer, ``Joint Source-Environment Adaptation of Data-Driven Underwater Acoustic Source Ranging Based on Model Uncertainty'', [Under Review], 2025

3. Dariush Kari, Lav R. Varshney, Andrew C. Singer, ``Invariant Feature Extraction for Robust Underwater Acoustic Localization'', [In Progress], 2025

4. Dariush Kari, Andrew C. Singer, ``Joint Source-Environment Adaptation for Deep Learning-Based Underwater Acoustic Source Ranging,'' In 2024 IEEE 58th Asilomar Conference on Signals, Systems and Computers, Pacific Grove, CA, USA, 2024, pp. 470-474

5. Dariush Kari, Andrew C. Singer, ``Wall Localization in A Water Tank Using A Cooperative Source of Opportunity,'' In 2024 IEEE Conference on Computational Imaging Using Synthetic Apertures (CISA), Boulder, CO, 2024, pp. 01-05

6. Dariush Kari, Andrew C. Singer, Hari Vishnu, Amir Weiss,``A gradient-based optimization approach for underwater acoustic source localization,'' In Proceedings of Meetings on Acoustics (POMA), AIP Publishing (2023), Vol. 51. [Best Student Paper Award]

Underwater Acoustic Communications

Underwater acoustic (UWA) communication faces significant challenges due to the dynamic underwater environment, impulsive noise, and limited data availability. This research addresses these issues through innovative approaches, including diffusion models for generating realistic synthetic datasets to support data-driven systems, adaptive robust channel estimators that enhance stability in impulsive noise, and hierarchical nonlinear equalization techniques. These methods collectively advance the efficiency, reliability, and adaptability of UWA communication systems in challenging environments. 

1. Yongjie Zhuang, Dariush Kari, Zhengnan Li, Milica Stojanovic, Andrew C. Singer, ``Generating Underwater Acoustic Communication Channel Impulse Responses Using A Diffusion Model," In Seventh Underwater Communications and Networking Conference (UComms), Sestri Levante, Italy, 2024, pp. 1--5

2. Dariush Kari, Nuri Denizcan Vanli, and Suleyman Serdar Kozat, ``Adaptive and Efficient Nonlinear Channel Equalization for Underwater Acoustic Communication,'' Physical Communication, vol. 24, pp. 83--93, September 2017

3. Dariush Kari, Iman Marivani, Farhan Khan, Muhammed Omer Sayin, and Suleyman Serdar Kozat, ``Robust Adaptive Algorithms for Underwater Acoustic Channel Estimation and Their Performance Analysis,'' Digital Signal Processing, vol. 68, pp. 57--68, September 2017

Adaptive Filtering and Online Learning

This research focuses on adaptive filtering and online learning in dynamic and high-dimensional environments. One study tackles sequential regression with nonlinearity and time-varying statistics by tracking the underlying manifold of high-dimensional data, reducing computational complexity and memory requirements while maintaining adaptability. Another work introduces the concept of boosting to adaptive filtering, combining multiple algorithms to achieve performance improvements in tasks like equalization and regression. Another contribution develops a sequential anomaly detection algorithm that accurately models nominal data distributions and dynamically adjusts thresholds to prevent contamination by anomalies. 

1. Dariush Kari, Ali H. Mirza, Farhan Khan, Huseyin Ozkan, and Suleyman Serdar Kozat,``Boosted Adaptive Filters,'' Digital Signal Processing, vol. 81, pp. 61--78, October 2018

2. Iman Marivani, Dariush Kari, Ali Emirhan Kurt, Eren Manis, ``Online anomaly detection in case of limited feedback with accurate distribution learning,'' In 25th Signal Processing and Communications Applications Conference (SIU), pp. 1--4, 2017

3. Dariush Kari, Iman Marivani, Ibrahim Delibalta, and Suleyman Serdar Kozat, ``Boosted LMS-based Piecewise Linear Adaptive Filters,'' In 24th European Signal Processing Conference (EUSIPCO), pp. 1593--1597, 2016

4. Burak C. Civek, Dariush Kari, Ibrahim Delibalta, and Suleyman Serdar Kozat, ``Big Data Signal Processing Using Boosted RLS Algorithm,'' In 24th Signal Processing and Communication Application Conference (SIU), pp. 1089--1092, May 2016

5. Farhan Khan, Dariush Kari, Ilyas Alper Karatepe, and Suleyman Serdar Kozat, ``Universal Nonlinear Regression on High Dimensional Data Using Adaptive Hierarchical Trees,'' IEEE Transactions on Big Data, vol. 2, no. 2, pp. 175--188, June 2016