🤖 AI Summary
This work addresses critical limitations of large language models (LLMs) in complex question answering (CQA)—including low accuracy, poor interpretability, uncontrolled knowledge integration, and frequent hallucinations—particularly in high-stakes domains such as multi-objective energy policy decision-making. To overcome these challenges, we propose the first systematic hybrid CQA architecture, integrating five synergistic techniques: domain adaptation, multi-step task decomposition, neuro-symbolic fusion, human-in-the-loop reinforcement supervision, and program synthesis. Our framework employs structured knowledge anchoring, iterative decomposition, and multimodal retrieval augmentation to enhance cross-cultural reasoning and multi-objective decision-making. We further introduce a rigorous evaluation benchmark emphasizing fairness, robustness, and anti-hallucination capabilities. The study establishes a paradigm shift toward trustworthy, auditable, and human-intervenable CQA systems driven by controllable hybrid architectures, providing both theoretical foundations and practical guidelines for next-generation intelligent QA.
📝 Abstract
This paper reviews the state-of-the-art of language models architectures and strategies for"complex"question-answering (QA, CQA, CPS) with a focus on hybridization. Large Language Models (LLM) are good at leveraging public data on standard problems but once you want to tackle more specific complex questions or problems (e.g. How does the concept of personal freedom vary between different cultures ? What is the best mix of power generation methods to reduce climate change ?) you may need specific architecture, knowledge, skills, methods, sensitive data protection, explainability, human approval and versatile feedback... Recent projects like ChatGPT and GALACTICA have allowed non-specialists to grasp the great potential as well as the equally strong limitations of LLM in complex QA. In this paper, we start by reviewing required skills and evaluation techniques. We integrate findings from the robust community edited research papers BIG, BLOOM and HELM which open source, benchmark and analyze limits and challenges of LLM in terms of tasks complexity and strict evaluation on accuracy (e.g. fairness, robustness, toxicity, ...) as a baseline. We discuss some challenges associated with complex QA, including domain adaptation, decomposition and efficient multi-step QA, long form and non-factoid QA, safety and multi-sensitivity data protection, multimodal search, hallucinations, explainability and truthfulness, temporal reasoning. We analyze current solutions and promising research trends, using elements such as: hybrid LLM architectural patterns, training and prompting strategies, active human reinforcement learning supervised with AI, neuro-symbolic and structured knowledge grounding, program synthesis, iterated decomposition and others.