arXiv subjects

Aroosa Hameed

Publications and source records attributed to Aroosa Hameed.

3 recordsLinked to original sources

arXiv2026

Sustainable Code Generation Using Large Language Models: A Systematic Literature Review

Large Language Models (LLMs) are widely used in software engineering to generate, complete, translate, and fix code, improving developer productivity. While most research focuses on the energy consumption and carbon emissions of model training and inference, far less attention has been given to the sustainability of the code these models produce. The efficiency of generated code affects the long-term environmental impact of software systems. Inefficient code can increase CPU usage, memory consumption, execution time, and overall energy use during deployment and operation. As LLM-generated code becomes more common in real-world projects, even small inefficiencies can lead to high environmental costs over time. This paper examines existing research on the sustainability of code generated by LLMs. We conduct a systematic literature review to analyze selected primary studies and investigate the extent to which LLMs are capable of producing sustainable code. In addition, we examine how sustainability is defined and measured in this context, including the metrics and evaluation strategies used to assess energy efficiency and resource usage. We also explore whether techniques such as fine-tuning and prompt engineering influence the sustainability of generated code. Through a structured analysis of the selected studies, we categorize research efforts based on their methodological approaches, evaluation practices, and experimental settings. The findings indicate that research in this area remains relatively limited and fragmented, with no widely accepted framework for measuring or benchmarking the sustainability of LLM-generated code. These observations highlight the need for clearer definitions, standardized evaluation methods, and systematic research to support environmentally friendly AI-assisted software engineering.

cs.SE↗

arXiv2025

RefactorCoderQA: Benchmarking LLMs for Multi-Domain Coding Question Solutions in Cloud and Edge Deployment

To optimize the reasoning and problem-solving capabilities of Large Language Models (LLMs), we propose a novel cloud-edge collaborative architecture that enables a structured multi-agent prompting framework. This framework comprises three specialized components: GuideLLM, a lightweight model deployed at the edge to provide methodological guidance; SolverLLM, a more powerful model hosted in the cloud and responsible for generating code solutions; and JudgeLLM, an automated evaluator for assessing solution correctness and quality. To evaluate and demonstrate the effectiveness of this architecture in realistic settings, we introduce RefactorCoderQA, a comprehensive benchmark designed to evaluate and enhance the performance of LLMs across multi-domain coding tasks. Motivated by the limitations of existing benchmarks, RefactorCoderQA systematically covers multiple technical domains, including Software Engineering, Data Science, Machine Learning, and Natural Language Processing, using authentic coding challenges sourced from Stack Overflow. We propose RefactorCoder-MoE, a fine-tuned mixture-of-experts (MoE) code language model based on DeepSeek-Coder-7B-Instruct, adapted to the RefactorCoderQA benchmark using QLoRA for domain-specific coding question answering. Extensive experiments demonstrate that RefactorCoder-MoE achieves strong and competitive performance, significantly outperforming all evaluated open-source and commercial baselines, with an overall accuracy of 76.84%.

cs.CL↗

arXiv2025

Sequencing to Mitigate Catastrophic Forgetting in Continual Learning

To cope with real-world dynamics, an intelligent system needs to incrementally acquire, update, and exploit knowledge throughout its lifetime. This ability, known as Continual learning, provides a foundation for AI systems to develop themselves adaptively. Catastrophic forgetting is a major challenge to the progress of Continual Learning approaches, where learning a new task usually results in a dramatic performance drop on previously learned ones. Many approaches have emerged to counteract the impact of CF. Most of the proposed approaches can be categorized into five classes: replay-based, regularization-based, optimization-based, representation-based, and architecture-based. In this work, we approach the problem from a different angle, specifically by considering the optimal sequencing of tasks as they are presented to the model. We investigate the role of task sequencing in mitigating CF and propose a method for determining the optimal task order. The proposed method leverages zero-shot scoring algorithms inspired by neural architecture search (NAS). Results demonstrate that intelligent task sequencing can substantially reduce CF. Moreover, when combined with traditional continual learning strategies, sequencing offers enhanced performance and robustness against forgetting. Additionally, the presented approaches can find applications in other fields, such as curriculum learning.

cs.LG↗