Federated Learning
Privacy-preserving model training across distributed clients — designing protocols, aggregation strategies, and threat models for collaborative ML without centralizing sensitive data.
Rochester Institute of Technology · Department of Computer Science
Trustworthy data.
Intelligent security.
The Laboratory of Data Quality and Intelligent Security at RIT builds the methods, models, and systems that make modern data infrastructure verifiable, resilient, and worthy of trust — from sensor networks at the edge to large-scale ML systems at the core.
4
Active OSS projects
25+
Researchers trained
90+
Peer-reviewed publications
20+
Years of research
Research Areas
Six frontiers we're pushing in parallel.
Our work sits at the intersection of data quality, security, and intelligent systems — with a hard focus on what's deployable, measurable, and trustworthy.
Data Quality at Scale
Provable quality metrics for streaming sensor data, with frameworks for evaluating fitness-for-purpose across heterogeneous sources.
ML System Security
Adversarial robustness, model integrity, and runtime monitoring for ML systems deployed in safety-critical environments.
Edge Intelligence
Neural network architectures and inference systems designed for the power, privacy, and reliability constraints of edge deployment.
Verifiable Mobile Security
Building on the lab's shipped Android apps — Android System Security Evaluation and Detector of Unverified Apps on Google Play — into continuous verification for app supply-chain integrity.
Open-Source Tooling
Releasing reproducible benchmarks, evaluation frameworks, and reference implementations the community can build on.
Active Projects
What we're shipping right now.
Open-source frameworks, peer-reviewed methods, and end-to-end systems — built by the current cohort and used by researchers beyond LDQIS.
Federated Learning Agentic AI
Python · Flower · PyTorch · LoRA
InteFL
Knowledge Management Framework for Federated Learning
A configuration and execution framework for federated learning simulations built on Flower. Implements novel PID controller-inspired and Trust-based aggregation strategies alongside classical robust methods (Krum, Multi-Krum, Bulyan, RFA, Trimmed-Mean), adversarial attack scheduling, federated LLM fine-tuning with LoRA, and a wide range of medical imaging datasets. Published in IEEE Intelligent Systems, 2026.
Federated Learning Adversarial Robustness
Flower · FastAPI · React · Celery
Phalanx-FL
Full-stack robustness platform, forked from InteFL
A platform for stress-testing federated learning under attack: configure, run, and compare experiments across 9 aggregation strategies, 11 adversarial attack types, and 20+ datasets, all from a React + FastAPI + Celery web app with live, streamed run monitoring. InteFL remains the lab's canonical research framework; Phalanx-FL is its robustness and tooling proving ground.
Federated Learning High-Performance
Rust · PyO3 · Python
Velocity-FL
Rust-core federated learning orchestration
A federated learning runtime with a Rust aggregation core and a Python-first interface. Eight paper-cited aggregation algorithms (FedAvg, FedProx, Krum, Bulyan, Geometric Median…), round-level and data-pipeline attack simulation, and aggregation kernels measured at ~92× the speed of pure Python.
Multi-Agent System Federated Learning
MCP + A2A · CLI · Pygame · Ren'Py
Kourai Khryseai
A multi-agent system with three interfaces
Ten specialized A2A-protocol agents — planner, coder, tester, reviewer, scribe, validators — that collaborate in a shared group chat to plan, build, and ship code. Three coordinated interfaces: terminal CLI, Pygame GUI, and a Ren'Py visual novel. Ongoing research into narrative interfaces for agent observability.
Publications
Selected publications.
A selection of the lab's work, from foundational to recent. For the complete record, see Dr. Reznik's Google Scholar profile.
2026
IEEE Intelligent Systems
InteFL Framework: Optimizing Federated Learning with Metacognition for Application Design and Deployment D. Korobeinikov, R. Zatsarenko, S. Chuprov, A. Barea, L. Reznik
2024
ASIA '24 · 19th Annual Symposium on Information Assurance · Albany, NY
Federated Learning Robustness on Real World Data in Intelligent Transportation Systems D. Korobeinikov, S. Chuprov, R. Zatsarenko, L. Reznik
2024
IEEE Xplore ↗ 2022
Wiley-IEEE Press
Intelligent Security Systems: How Artificial Intelligence, Machine Learning and Data Science Work For and Against Computer Security L. Reznik
2003
Encyclopedia of Life Support Systems, Developed under the Auspices of the UNESCO, EOLSS Publishers, Oxford, UK
General Principles and Purposes of Computational Intelligence in Systems Science and Cybernetics L. Reznik
People
The team — past and present.
LDQIS is the work of every researcher who has passed through it. Today's team is pushing on federated learning and agentic AI; previous cohorts built the foundations in mobile security, data quality evaluation, wearable sensors, and computational intelligence that made today's work possible.
Now
Current Team
Federated learning · Agentic AI
Lab Director · Professor of Computer Science · RIT
Faculty Collaborator · Professor · UTRGV
RIT Graduate Researcher · InteFL lead author
RIT Graduate Researcher
RIT Graduate Researcher · Phalanx-FL · Velocity-FL · Kourai Khryseai
Then
Past Research Cohort
Mobile security · Android attacks · Wearable sensors · Data quality · Computational intelligenceThese researchers led work that remains foundational to LDQIS today, including our Android security apps, the NSF Data Quality & Security Evaluation Framework, and the wearable-sensors and computational-intelligence research lines that preceded today's federated learning work.
Computational intelligence · FUZZ-IEEE 2018
Sensor selection optimization · 2019
Sentiment analysis · MS thesis
Fuzzy systems + CI software · 2015-18
Mobile trust evaluation · 2015
Data quality + Android security · 2017
Android security framework · 2017
Mobile trust evaluation · 2015
Alumnus · Neural network architectures
2018 NSF Webinar Co-Presenter
Multi-modal sensor selection · 2022
Android anomaly detection · 2019
RNNs for Android attacks · 2020
Android security framework · 2017
Wearable sensors + ML · 2018-19
Transfer learning + ML systems · 2022
Wearable sensor design · 2018
Get Involved
Support research that makes data trustworthy.
Funding from individuals, foundations, and industry partners directly supports graduate stipends, equipment, and open-source releases. Every contribution compounds — in students trained, papers published, and tools shipped.