MD Azizul Hakim

MD Azizul Hakim

He/Him

Dip. Eng. (CST) @BSPI | AI Researcher

Publications

You can also find my articles on my Google Scholar profile.

Journal Articles

Biologically inspired neural network layer with homeostatic regulation and adaptive repair mechanisms

Published in Scientific Reports, 2025

This paper introduces a novel neural network layer incorporating homeostatic regulation and adaptive repair mechanisms inspired by biological nervous systems, demonstrating superior performance on biomedical datasets.

Authors: Hakim, MD Azizul, & Alam, Mohammad Ifazul. (2025). "Biologically inspired neural network layer with homeostatic regulation and adaptive repair mechanisms." Scientific Reports, 15, 9114.
Download Paper

Conference Papers

NutriGuard: LLM-Driven Nutritional Assessment for Chronic Disease Prevention

Published in IEEE Conference Proceedings, 2025

This conference paper presents NutriGuard, an innovative multimodal AI framework integrating optical character recognition (OCR), deep learning, and fine-tuned large language models (LLMs) for personalized nutrition management. The system delivers real-time, context-aware dietary recommendations tailored to users’ health profiles, advancing AI-driven preventive healthcare through multilingual, clinically validated frameworks.

Authors: Hakim, M.A., Ifty, R.A., Delowar, K.E., Chowdhury, S.H., Rashid, I., & Shakib, M. (2025). "NutriGuard: LLM-Driven Nutritional Assessment for Chronic Disease Prevention." IEEE Conference Proceedings. DOI: 10.1109/QPAIN66474.2025.11171750
Download Paper

Neuroadaptive Intelligence: A Biologically-Inspired, Self-Repairing Neural Layer with Adaptive Learning for Molecular Graph Classification

Published in IEEE Conference Proceedings, 2025

This conference paper introduces Neuroadaptive Intelligence, a novel approach inspired by biological systems. It details the development of a self-repairing neural layer incorporating adaptive learning mechanisms specifically designed for the complex task of molecular graph classification. This work represents a significant contribution to biologically-inspired AI and its application in scientific domains, particularly drug discovery and molecular property prediction.

Authors: Hakim, M.A. (2025). "Neuroadaptive Intelligence: A Biologically-Inspired, Self-Repairing Neural Layer with Adaptive Learning for Molecular Graph Classification." IEEE Conference Proceedings. DOI: 10.1109/QPAIN66474.2025.11171860
Download Paper

Preprints

Brevity Constraints Reverse Performance Hierarchies in Language Models

Published in arXiv preprint, 2026

This preprint investigates how imposing brevity constraints on language model outputs can unexpectedly reverse established performance hierarchies across model families and sizes. The findings have significant implications for understanding test-time compute efficiency and the deployment behaviour of large language models under constrained generation settings.

Authors: Hakim, MD Azizul (2025). "Brevity Constraints Reverse Performance Hierarchies in Language Models." arXiv preprint. arXiv:2604.00025
Download Paper

Multi-Scale Temporal Homeostasis Enables Efficient and Robust Neural Networks

Published in arXiv preprint, 2025

This preprint introduces multi-scale temporal homeostasis as a mechanism for enhancing the efficiency and robustness of artificial neural networks. Inspired by biological nervous systems that coordinate activity across multiple temporal scales, the proposed framework enables more stable training dynamics and improved resilience under adverse conditions.

Authors: Hakim, MD Azizul (2025). "Multi-Scale Temporal Homeostasis Enables Efficient and Robust Neural Networks." arXiv preprint. arXiv:2602.07009
Download Paper