QFedAgent: Quantum-Enhanced Personalized Federated Learning for Multi-Agent Activity Recognition

Source

• Source name: arXiv
• Source kind: paper
• Authors / organization: Quoc Bao Phan, Tuy Tan Nguyen
• Venue: arXiv: cs.LG
• Published: 2026-07-02
• Collected: 2026-07-03
• URL: http://arxiv.org/abs/2607.02426v1
• PDF: https://arxiv.org/pdf/2607.02426v1
• DOI: UNKNOWN

Plain-Language Brief

The paper proposes a new method called QFedAgent for multi-agent activity recognition in privacy-sensitive robotic sensing applications. The method uses a hybrid quantum-classical personalized Federated Learning framework, which includes a variational quantum circuit fusion module that reduces the number of parameters needed by approximately 10 times compared to classical methods.

What the Document Says

• The paper discusses federated learning (FL), a method for collaborative model training across distributed devices without sharing raw data, suitable for privacy-sensitive robotic sensing applications.
• Multi-agent systems generate heterogeneous and non-IID multimodal sensor streams that degrade conventional FL algorithms, and classical fusion modules introduce substantial parameter overhead and communication cost.
• The paper introduces QFedAgent, a hybrid quantum-classical personalized FL framework for multi-agent activity recognition.
• QFedAgent integrates a variational quantum circuit fusion module that models accelerometer--gyroscope interactions through quantum state encoding and entanglement.
• This quantum fusion module requires only 72 quantum rotation parameters versus 33K in classical multi-layer perceptron-based fusion, achieving approximately 10x total parameter reduction.
• Experiments on the OPPORTUNITY dataset under subject-based non-IID partitions demonstrate 97.7% mean test accuracy for QFedAgent, confirming that it remains competitive with conventional federated baselines.

Key Concepts Mentioned

• federated learning
• multi-agent systems
• quantum state encoding
• entanglement
• variational quantum circuit
• accelerometer--gyroscope interactions
• OPPORTUNITY dataset

Claims Made by the Document

• QFedAgent achieves approximately 10x total parameter reduction compared to classical methods.
• QFedAgent remains competitive with conventional federated baselines.

Evidence Used

• The paper states that the quantum fusion module requires only 72 quantum rotation parameters versus 33K in classical methods, implying a parameter reduction. The experiments demonstrate comparable performance to conventional federated baselines.

Methods / Approach

UNKNOWN

Limitations Stated by the Authors

UNKNOWN

Limitations Noted by the Agent

Agent interpretation — not from the source.

UNKNOWN

Open Questions

UNKNOWN

Notes for Future Q&A

May include interpretation. Labeled where it does.

UNKNOWN

Raw Abstract / Excerpt

Federated learning (FL) enables collaborative model training across distributed devices without sharing raw data, making it suitable for privacy-sensitive robotic sensing applications. However, multi-agent systems generate heterogeneous and non-independent and identically distributed (non-IID) multimodal sensor streams that degrade conventional FL algorithms, while classical fusion modules introduce substantial parameter overhead and communication cost. This paper proposes QFedAgent, a hybrid quantum-classical personalized FL framework for multi-agent activity recognition. The approach integrates a variational quantum circuit fusion module that models accelerometer--gyroscope interactions through quantum state encoding and entanglement, requiring only 72 quantum rotation parameters versus 33K in classical multi-layer perceptron-based fusion, achieving approximately 10x total parameter reduction. Experiments on the OPPORTUNITY dataset under subject-based non-IID partitions demonstrate 97.7% mean test accuracy, confirming that parameter-efficient quantum fusion remains competitive with conventional federated baselines.