Research

A collection of my publications exploring the frontiers of AI system design, safety and explainability.

Filter Research

Exploring autonomous agents through the lens of large language models: A review

Saikat Barua (2024)

Autonomous Agents

Large Language Models (LLMs)

Agentic AI

Literature Review

This review explores LLM-powered autonomous agents, their architecture, capabilities, challenges, and future research directions.

View Paper

Guardians of the agentic system: Preventing many shots jailbreak with agentic system

Saikat Barua, Mostafizur Rahman, Md Jafor Sadek, Rafiul Islam, Shehenaz Khaled, Ahmedul Kabir (2025)

LLM Security

Jailbreak Prevention

Agentic Systems

Adversarial Attacks

Proposes systems to detect and prevent advanced jailbreaking attacks on LLM-based autonomous AI agents.

View Paper

KAXAI: An Integrated Environment for Knowledge Analysis and Explainable AI

Saikat Barua, Sifat Momen (2024)

Explainable AI (XAI)

Automated Machine Learning (AutoML)

Synthetic Data Generation

Human-Computer Interaction

KAXAI integrates AutoML, XAI, and synthetic data generation for accessible machine learning and interpretability.

View Paper

RESCUED: Robust Quantum Error Correction with Surface Code in Noisy Channels using Ensemble Decoder

Saikat Barua, Syed Emad Uddin Shubha, Monika Rahman, Apurba Jalal Uchash, MRC Mahdy (2024)

Quantum Error Correction

Surface Codes

Ensemble Methods

Noisy Quantum Channels

RESCUED proposes an ensemble decoder for robust quantum error correction with surface codes in noisy channels.

View Paper

ELMAGIC: Energy-Efficient Lean Model for Reliable Medical Image Generation and Classification using Forward Forward Algorithm

Saikat Barua, Mostafizur Rahman, Mezbah Uddin Saad, Rafiul Islam, Md Jafor Sadek (2024)

Medical Image Generation

Energy-Efficient Models

Forward-Forward Algorithm

Image Classification

ELMAGIC offers an energy-efficient lean model for medical image generation and classification using Forward Forward.

View Paper

Pygen: A collaborative human-ai approach to python package creation

Saikat Barua, Mostafizur Rahman, Md Jafor Sadek, Rafiul Islam, Shehenaz Khaled, Md Shohrab Hossain (2024)

Human-AI Collaboration

Code Generation

Python

Software Development

PyGen is a human-AI collaborative platform for automating Python package creation from abstract ideas.

View Paper

QuXAI: Explainers for Hybrid Quantum Machine Learning Models

Saikat Barua, Mostafizur Rahman, Shehenaz Khaled, Md Jafor Sadek, Rafiul Islam, Shahnewaz Siddique (2025)

Quantum Machine Learning (QML)

Explainable AI (XAI)

Hybrid Quantum-Classical

Model Interpretability

QuXAI introduces a framework and explainer (Q-MEDLEY) for hybrid quantum-classical machine learning model transparency.

View Paper