Research

Research Interests:

• NLP

• Metaverse

• Machine Learning

• Artificial intelligence

• X Reality in Education

• Federated Learning

Research Experience:

1. Trustworthiness of Metaverse:  A Complete Survey [2022]

• Supervisor: Dr. Firoz Mridha

• Authors: Asif Zaman (AIUB), Mohammad Tanzib(AIUB), Jamiul Jim

• Publication: Review ongoing

• The paper is all about the parameters of Metaverse and the base of trustworthiness. We made the parameters reliable and trustworthy which enhanced the Metaverse system security and trustworthiness for the users.

2. An Analysis of Personalized Federated Learning Algorithms with an Intuitive Protected Model Training [2022]

• Supervisor: Dr. Firoz Mridha

• Authors: Asif Zaman (AIUB), Mohammad Tanzib(AIUB), Mushfique, Jamiul Jim

• Publication: Review ongoing

• The paper focuses on assessing the effectiveness of four personalized Federated Learning (FL) algorithms, namely FedAvg, APPLE, FedBABU, and FedProto, in a decentralized setting using the Fashion MNIST dataset. The dataset has a non-IID (non-identically distributed) data distribution, meaning that the data on individual devices differs significantly.

3. Enhancing Trustworthiness of Mixed Reality in Education: A Comprehensive Framework [2023]

• Authors: Asif Zaman (AIUB), Mohammad Tanzib(AIUB)

• Publication: Review ongoing

• The research paper proposes a comprehensive framework for assessing and enhancing trustworthiness in the use of Mixed Reality (MR) technology in the education sector. MR technology offers immersive and interactive learning experiences but requires trust to be established to ensure reliability, effectiveness, and ethical use. The framework addresses three key dimensions of trustworthiness. Firstly, it evaluates the technical aspects of MR systems, including system stability, tracking accuracy, and latency. These factors are essential for establishing trust between users and the technology, as any technical issues can undermine the learning experience.

4. Privacy Preserving Machine Learning Model Personalization through Federated Personalized Learning

• Authors: Asif Zaman (AIUB), Mohammad Tanzib(AIUB), MD. Shahriar Sajid (Ruet), Shadman Sakeeb Khan (NSU)

• Status: Proceedings

• Publication-link: IEEE Explore Data'23 Conference

• The research paper focuses on addressing the challenges posed by data privacy concerns in the context of advancing artificial intelligence and machine learning technologies. It introduces a novel approach called "Privacy-Preserving Machine Learning combined with Federated Personalized Learning (PPML-FPL)" and presents a comprehensive performance analysis of this approach.