Our collaboration with Luis Garcia, and Amy McDonnell (U of Utah) resulted in a new accepted paper “Human Modeling Gaps: Safety, Security, and Privacy Risks in Automated Driving” ” in the 3rd International conference on security and privacy in CPS and smart vehicles. Special Track: Human-Machine Interfaces in CPS and Smart Vehicles.
Category: News
Two new papers accepted at NeurIPS’25 workshops! Congrats Tyler, Mahmoud, and Beckham
Two new papers from the Pervasive Autonomy Lab and collaborators are accepted in the NeurIPS’25 workshops.
1- Mahmoud Srewa, Tianyu Zhao, Salma Elmalaki, “A Systematic Evaluation of Preference Aggregation in Federated RLHF for Pluralistic Alignment of LLMs,” has been accepted at the NeurIPS 2025 Workshop on Evaluating the Evolving LLM Lifecycle!
Congrats Mahmoud and Tianyu (Tyler)!
2- Chung-Mou Pan, Salma Elmalaki, Yasser Shoukry, Sitao Huang, “NeuSym-HLS: Learning-Driven Symbolic Distillation in High-Level Synthesis of Hardware Accelerators,” has been accepted at the NeurIPS 2025 Workshop on Machine Learning (ML) for Systems (MLForSys2025)
This is a collaboration project between Prof. Shoukry and Prof. Huang.
Congrats Chung-Mou (Beckham)!
Salma is invited to the NSF M3X visioning workshop on human factors in computation models!
Salma has been part of the recent NSF visioning workshop on the Formalization of Human Factors into Computational Models.
This workshop was a powerful example of what happens when we bring together a truly multidisciplinary group. Researchers from fields as varied as computer science, HCI, psychology, cognitive science, and engineering came together to tackle a crucial challenge: how to better integrate human factors into our computational models. It was inspiring to collaborate on this shared goal of designing more effective and human-centric systems.
More about the workshop: https://lnkd.in/g8-mt5bY
More about the M3X program: https://lnkd.in/gAtP5qqH
Engineering AI for Society Institute has launched!
The Engineering+ Society Institute: Engineering AI for Society will use a socio-technical approach and bring together a core engineering and computer science team with strategic partners from other disciplines. The core team includes engineering faculty members Yasser Shoukry, Yanning Shen, Hyoukjun Kwon, Salma Elmalaki, Alexandra Voloshina, Zhou Li and Aparna Chandramowlishwaran, as well as computer science faculty Sameer Singh. Strategic partners include UCI School of Social Sciences Dean Bill Maurer, Beall Center for Art & Technology Director Jesse Colin Jackson, Beall Applied Innovation’s Errol Arkilic (chief executive) and Stuart Matthews (AI strategist and innovation), and Logic and Philosophy Department Chair James Owen Weatherall.
https://engineering.uci.edu/news/2025/6/samueli-schools-engineering-society-institute-focus-ai
Salma wins a new NSF award on Neurosymbolic High-level Synthesis!
NeuHLS: A Neursoymbolic Framework for High-level Synthesis of Multi-Task Learning (NSF:SHF 2025-2029)
The growing demand for smart and autonomous systems has driven a surge in the deployment of edge devices. However, their limited computational resources and energy constraints often hinder the deployment of complex deep neural networks (DNNs). Optimizing DNNs for edge devices is crucial to unlock their full potential and enable a wider range of innovative applications. Our project aims to address the challenges of deploying DNNs on edge devices by developing a flexible and efficient framework. This project aims to exploit recent advances in multi-task learning, neurosymbolic AI, and high-level synthesis to develop new generation of tools that can automatically generate hardware accelerators for edge devices while satisfying latency and hardware platform constraints. In addition, the synthesized hardware must maximize the number of DNN weights implemented using software tunable parameters to allow for flexible fine-tuning at runtime. The proposed toolchain consists of three phases. In the first one, it aims to merge a set of single-task DNNs into one multi-task DNN by sharing representations from different single-task DNNs and hence reduces the model size. Next, the tool chain will exploit symbolic knowledge distillation (SKD) to compress the multi-task DNN into a neurosymbolic model which can be then processed by novel neurosymbolic high-level synthesis (HLS) techniques that will optimize the deployment while balancing accuracy, hardware utilization, and latency. The proposed toolchain will be rigorously evaluated using existing benchmarks and real-world application deployment in the domain of autonomous drones.
Mahmoud received the 2025 DTEI Summer Graduate Scholarship!
Congrats Mahmoud!
More details on the scholarship (https://dtei.uci.edu/opportunities/grads-postdocs/dtei-graduate-fellowship/)
Tianyu Zhao wins first place in the GEECS Annual Technology Showcase 2025 (GEECSANTS)
GEECSANTS is a celebration of graduate student research, innovation, and community within the Electrical Engineering and Computer Science (EECS) department at UCI.
Tyler’s research spans the general area of ethical AI, focusing on fairness in human-in-the-loop CPS and fairness alignment of LLM.
Diana Romero and Mahmoud Srewa showcased their research and celebrated Tyler’s award.
Salma serves as a panelist in HumanSys’25
HumanSys was held at UCI as part of the CPS-IoT’25 week. The panel session focused on the challenges and future of human-centered sensing, cyber-physical systems, and human-in-the-loop AI for health and well-being.
Two new papers accepted in HumanSys’25
Congratulations to Diana, Tyler, and Mahmoud for accepting their papers at HumanSys’25.
Papers are available:
- Diana Romero, Fatima Anwar, Salma Elmalaki, “MoCoMR: A Collaborative MR Simulator with Individual Behavior Modeling,” arXiv:2503.09874
- Mahmoud Srewa, Tianyu Zhao, Salma Elmalaki, “PluralLLM: Pluralistic Alignment in LLMs via Federated Learning,” arXiv:2503.09925
Salma organized a workshop during the 2025 NSF CPS PI meeting on Human-in-the-Loop CPS in the Era of Human-AI Alignment: Impact and Future Direction
This workshop addresses critical challenges in Human-in-the-Loop Cyber-Physical Systems (CPS) by focusing on human-AI alignment. We will explore three key gaps: (1) Formalizing Alignment Concepts: Developing rigorous frameworks for defining and measuring crucial aspects like trustworthiness, transparency, explainability, fairness, and sustainability. (2) Limitations in Sociotechnical Approaches: Evaluating current sociotechnical methods and identifying their shortcomings in achieving effective human-AI collaboration. (3) Assessing Societal Impact: Analyzing the broad impact of advancements in human-AI collaboration on large-scale CPS. The workshop features invited talks from leading experts and a panel followed by interactive roundtable discussions designed to foster brainstorming and collaborative solutions. Join us in contributing to the future of aligned and beneficial human-AI integration in CPS.
https://cps-vo.org/group/cps-pimtg25/miniworkshops
