arXiv cs.CL· Jatin Gupta, Akhil Sharma, Saransh Singhania, Ali Imam Abidi·· 12 小时前AI 评分22
CA-ThinkFlow：面向印度注册会计师考试的检索增强推理框架
AI-Enabled Question-Answering for Chartered Accountancy: A Retrieval-Augmented Reasoning Framework
AI 导读
CA-ThinkFlow 是一个参数高效的 RAG 框架，基于 14B、4-bit 量化的 14B-DeepSeek-R1 推理模型，并搭配保持文档结构的 Docling 版面感知抽取系统。
正文 · 原文
Abstract:The inception of Large Language Models (LLMs) has catalyzed AI adoption in the finance sector, yet their reliability in complex, jurisdiction-specific tasks like Indian Chartered Accountancy (CA) remains limited. The models display difficulty in executing numerical tasks which require multiple steps while also needing advanced knowledge about legal regulations and the method of scaling their operations is not feasible in settings which have limited access to resources. We present CA-ThinkFlow as a parameter-efficient Retrieval-Augmented Generation (RAG) framework which operates with a 14B, 4-bit-quantized reasoning model, 14B-DeepSeek-R1, and a layout-aware Docling extraction system which maintains document structure during extraction. CA-ThinkFlow uses a basic RAG method which automatically adds retrieved information into the prompt, while it depends on the model's built-in Chain-of-Thought (CoT) functions to create context and produce correct answers. The system we developed system operates at performance levels which match large proprietary models when we tested it on the multi-level CA-Ben benchmark, achieving Scholastic Reliability Coefficient (SRC) results which equal 68.75\% of GPT-4o and Claude 3.5 Sonnet. The framework shows high efficiency and strength in handling parameters, but essential reasoning abilities fail to process complex regulatory texts which exist in fields such as Taxation.
| Comments: | Accepted for publication at ICITI-2026: 3rd International Conference on Information Technology and Intelligence (ICITI), NIT Agartala, India, 2026 |
| Subjects: | Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Information Retrieval (cs.IR) |
| Cite as: | arXiv:2605.00257 [cs.CL] |
| (or arXiv:2605.00257v2 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2605.00257 arXiv-issued DOI via DataCite |
Submission history
From: Jatin Gupta [view email]
[v1]
Thu, 30 Apr 2026 21:50:49 UTC (67 KB)
[v2]
Sun, 1 Nov 2026 10:00:34 UTC (67 KB)
来源：arXiv cs.CL · arxiv.org