An Overview and Discussion on Using Large Language Models for Implementation Generation of Solutions to Open-Ended Problems
Hashmath Shaik, Alex Doboli
TL;DR
The paper analyzes the gap between traditional automated implementation generation and the capabilities of large language models. It surveys LLM foundations, prompting techniques, retrieval-augmented generation, and reinforcement learning as enabling tools, arguing that LLMs can support problem framing, exploration, and assessment for open-ended problems. It highlights methods to ground reasoning (RAG), guide generation (prompt engineering), and refine behavior (RL), while noting core limitations in hierarchical reasoning and memory. Together, these insights indicate a path toward integrated, multi-modal, memory-aware implementation generation with substantial practical impact across domains requiring flexible problem solving.
Abstract
Large Language Models offer new opportunities to devise automated implementation generation methods that can tackle problem solving activities beyond traditional methods, which require algorithmic specifications and can use only static domain knowledge, like performance metrics and libraries of basic building blocks. Large Language Models could support creating new methods to support problem solving activities for open-ended problems, like problem framing, exploring possible solving approaches, feature elaboration and combination, more advanced implementation assessment, and handling unexpected situations. This report summarized the current work on Large Language Models, including model prompting, Reinforcement Learning, and Retrieval-Augmented Generation. Future research requirements were also discussed.