| ||
News
|
||
ResearchMy research centers on the intersection of multi-agent networks, machine learning, and optimization. Specifically, I'm interested in multi-agent networks, where agents have different views/modalities of data. These agents, which could be robots, sensors, or machine learning models, are designed to share information amongst themselves to fulfill their individual objectives. My projects aim to harness these agents' distinct modalities or characteristics to enhance decision-making processes, such as bandwidth allocation and data sharing for machine learning models. Data Transmission and Information Extraction across Diverse ModalitiesThis research thread explores the optimization of data transmission in scenarios where data is derived from various modalities, each offering a unique view or sensory input. The aim is to refine the information extraction and exchange processes by first identifying the importance of each data modality and then developing transmission strategies that are tailored to the unique objective of each agent. Our methods are tailored to meet the specific requirements of agents for each data type, thereby boosting the efficiency and reliability of systems.The projects concentrate on three key areas: (1) enhanced multi-modal data sharing, (2) implementing differential privacy mechanisms for data sharing, and (3) mitigating the effects of adversarial attacks on false information sharing. PublicationsPo-han Li, Sravan Kumar Ankireddy, Ruihan Zhao, Hossein Nourkhiz Mahjoub, Ehsan Moradi Pari, Ufuk Topcu, Sandeep P. Chinchali, Hyeji Kim Advances in Neural Information Processing Systems (Neurips), 2023 Blog / Code Po-han Li, Sandeep P. Chinchali, Ufuk Topcu Under Review, 2024 Po-han Li, Sandeep P. Chinchali, Ufuk Topcu Under Review, 2024 Po-han Li, Sandeep P Chinchali, Ufuk Topcu American Control Conference (ACC), 2023 Blog / Code Po-han Li, Ufuk Topcu, Sandeep P Chinchali 58th Allerton Conference on Communication, Control, and Computing, 2022 Large Language Model SelectionLanguage models like ChatGPT and Bard have become integral to our daily lives, yet their performance hinges on user context, training data, and model architecture, which users typically have limited knowledge of. Interactions with these models occur through internet APIs, treating them as black boxes. My project tackles a key question: How can we select the right large language model on the Internet for a specific task? How do we even define tasks of our chatbot conversation? To answer this, I seek to uncover the correlation between model performance and user context of conversation. Finally, I want create a decision-making algorithm for choosing the best model.PublicationsPo-han Li, Oyku Selin Toprak, Aditya Narayanan, Ufuk Topcu, Sandeep Chinchali Under review, 2024 Blog / Code Decentralized Data SharingThese projects aim to improve federated learning (FL), which typically requires all devices to have identical neural network structures. My projects explore sharing raw data, not gradients, among devices to overcome this limitation. Additionally, I focus on challenges related to limited network bandwidth and privacy preservation when sharing valuable data. My work covers topics such as out-of-distribution detection, data valuation, active learning, differential privacy, and distributed optimization.PublicationsOguzhan Akcin, Po-han Li, Shubhankar Agarwal, Sandeep P. Chinchali Conference on Robot Learning (CoRL), 2022 Yuchong Geng, Dongyue Zhang, Po-han Li, Oguzhan Akcin, Ao Tang, Sandeep P Chinchali Conference on Robot Learning (CoRL), 2021 MiscellaneousI enjoy working on projects with brilliant minds across labs, even if they are not directly related to my research. The followings are some of my side, but cool projects.Efficient Image RetrievalMohammad Omama, Po-han Li, Sandeep P. Chinchali Under Review, 2024 OthersAAAI Conference on Artificial Intelligence, 2023 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2024 Conference on Decision and Control (CDC), 2023 |
This website is based on Jon Barron's source code.
His website can be found here.
|