Report

Kjonmukisa

July 24, 2024 - July 28, 2024

Papers

Paper 1: Byzantine-Resilient Federated Learning: An Adaptive Architecture Integrating Trimmed Mean Averaging and Dynamic Learning Rates (tentative)
Paper 2: AdaptTrim: An Adaptive and Efficient Architecture for Federated Learning Using Trimmed Averaging (tentative)
Paper 3: Efficient Federated Learning with Intrusion Detection For Enhancing Monitoring of COPD patients and ensured security (tentative)

65% Progress
Last Week's Progress

1. Modeling new federated learning strategy that is robust

This Week Todo's

1. I will begin developing a new federated learning strategy that is resilient to Byzantine faults. This involves integrating trimmed mean averaging techniques to mitigate the impact of outliers and adversarial attacks.

2. I plan includes starting to incorporate adaptive mechanisms, such as dynamic learning rates, to address the identified limitations and improve system efficiency

Project Progress

PurePet:
Digital Twins:

50% Progress
Last Week's Project

PurePet: This week, I successfully coded the initial API endpoints, enabling the application to interact with the server. This is a significant step towards creating a seamless user experience.
Digital Twins: I analyzed case studies focusing on the deployment of digital twins in tunnel infrastructure. This helped identify best practices and common challenges, which will inform the project's direction.

This Week's Project

PurePet: The focus will be on testing the API endpoints to ensure they can accurately retrieve data from the database. This involves debugging and optimizing the code for efficiency.
Digital Twins: I will continue analyzing case studies, now shifting focus to digital twins in bridge infrastructure. Additionally, I plan to begin learning the basics of 3D modeling, which is essential for creating accurate digital twins.

Monthly Goals

The primary goal for this month is to optimize the Federated Average algorithm to enhance its robustness and efficiency

Annual Goals

1.Submit at least 2 articles to scientific journals.
2.Submit and attend at least one NSL-approved domestic conference.
3.Submit and attend at least one NSL-approved international conference.