International Journal of Smart Computing and Artificial Intelligence

Zero-Shot Text Classification Using Large Language Models for Key Audit Matters in Japanese Audit Reports

2025-03-24T13:47:36+00:00

Japanese-listed companies are required to submit audit reports to the Prime Minister of Japan. In principle, these reports must include “Key Audit Matters” (KAMs), which are matters that the auditors, as professional experts, have judged as particularly important when auditing financial statements. A previous study proposed an automatic classification method called zero-shot text classification for KAMs. We examine whether zero-shot text classification with large language models (LLMs) such as ChatGPT can automatically classify KAMs. We also examine how the following three approaches contribute to the accuracy of zero-shot text classification by LLMs: definition refinement, majority decision-making based on LLM outputs, and use of state-of-the- art models. The experimental results confirm that definition refinement and majority decision- making based on more than three results are useful to some extent. Furthermore, the latest ChatGPT model, gpt-4-1106-preview of the Generative Pre-trained Transformer 4 (GPT-4) model, achieved a classification accuracy of up to 87.2%.

Selective Potentiality and Moving Focus for Interpreting Multi-Layered Neural Network

2025-06-20T11:27:23+00:00

The present paper aims to demonstrate the existence of simplification forces in neural networks. These simplification forces can be represented by the simplest network, called ``prototype.'' To extract the prototype, we need to identify necessary and important information during learning. The structural potentiality has been proposed to reduce information, aiming to reduce unnecessary information, but one of its problems lies in excessive information reduction. To preserve important information, we need to maximize or at least weaken the excessive information reduction. To solve this problem, we introduce a new potentiality called ``selective potentiality,'' which allows us to move a focus field where a group of connection weights can be flexibly reduced. This method aims to replace the troublesome contradictory operations of potentiality reduction and augmentation with more concrete and manageable ones.

The method was applied to an artificial dataset, in which linear and non-linear relations were introduced. The results confirmed that selective potentiality could be increased to weaken structural potentiality reduction. The selective potentiality showed strong forces of simplification throughout the entire learning process. By seeking the simplest prototype, additional results were obtained, where networks tried to infer the outputs, enhancing both linear and non-linear inputs for better generalization.

Potential to Foster Relationships and Reduce Early Turnover by Assigning New Employees to Younger Teams based on Formal and Informal Relationships

2025-01-25T13:15:29+00:00

The early turnover rate of Japanese university graduates has remained at approximately 30% over the past 30 years, mainly because of poor human relations within companies. To the best of our knowledge, no previous study has examined early turnover from the perspective of both formal and informal relationships. We propose an organizational structure for companies that reduces early turnover by considering both formal and informal relationships. A simulation model was developed to evaluate the effects of such relationship networks on early turnover. To inhibit the early turnover of new employees, we recommend forming teams of younger existing employees and assigning new employees to such teams. The simulation results show that the proposed corporate organizational structure is more effective in reducing early turnover than the existing top-down corporate structures.

Advancing Arbitrary-Scale Image Super-Resolution: Introducing Residual In Residual Dense Networks with a Novel Local Implicit Image Function

2025-03-30T12:12:39+00:00

The progress in image super-resolution has seen significant advancements due to the emergence of deep convolutional neural networks. While most researchers concentrate on training models for specific scales, only a few delve into creating models adaptable to various scales. This study builds upon prior research that focuses on achieving arbitrary resolution with a single model. The model under consideration employs an auto-encoder structure. The encoder extracts feature maps from the input image, while the decoder reconstructs these feature maps to the resolution specified by the user. Referred to as RRDN-NLIIF (Residual in Residual Dense Networks with Novel Local Implicit Image Function), our experimental results demonstrate its superior performance over the benchmark model in terms of the PSNR metric.