Source author record

JaeYeon Park

JaeYeon Park appears in the imported research catalog. Authorship, coauthor and topic links are available while profile ownership is still unclaimed.

ResearcherUnclaimed source record

Artificial Intelligence eess.SY Machine Learning Systems and Control

Catalog footprint

What is connected

2works

4topics

4close collaborators

Actions

Connect this record

Open graph Browse works

Inspect adjacent papers, topics, institutions and collaborators without losing the researcher page.

Building this map preview

BZPEER is loading the nearby papers, people, topics and institutions for this page.

preprint2026arXiv

DSP-Based Sub-Switching-Period Current-Limiting Control for Grid-Tied Inverter under Grid Faults

This paper presents a sub-switching period current-limiting control for a grid-tied inverter to prevent transient overcurrents during grid faults and enable seamless fault ride-through (FRT). Sudden grid-voltage disturbances, such as voltage sags or phase jumps, can induce large transient currents within a switching period, particularly at low switching frequencies. Upon disturbance detection, the proposed method immediately modifies the pulse-width modulation carrier, enabling continuous regulation of the inverter output current within a time much shorter than a switching period without interrupting current flow. The proposed method can be implemented on commonly used digital signal processors without requiring specialized analog or digital circuits or high-speed computing devices. Experimental results from a 2-level, 3-phase inverter switching at 3.6 kHz validate the effectiveness of the proposed method under symmetric and asymmetric voltage sags and phase jumps.

preprint2025arXiv

Tazza: Shuffling Neural Network Parameters for Secure and Private Federated Learning

Federated learning enables decentralized model training without sharing raw data, preserving data privacy. However, its vulnerability towards critical security threats, such as gradient inversion and model poisoning by malicious clients, remain unresolved. Existing solutions often address these issues separately, sacrificing either system robustness or model accuracy. This work introduces Tazza, a secure and efficient federated learning framework that simultaneously addresses both challenges. By leveraging the permutation equivariance and invariance properties of neural networks via weight shuffling and shuffled model validation, Tazza enhances resilience against diverse poisoning attacks, while ensuring data confidentiality and high model accuracy. Comprehensive evaluations on various datasets and embedded platforms show that Tazza achieves robust defense with up to 6.7x improved computational efficiency compared to alternative schemes, without compromising performance.