Robotic Process Automation (RPA) has rapidly evolved into a widely recognized and influential software technology. Its growing relevance has sparked diverse research efforts across various disciplines. This study aims to map the scientific landscape of RPA by identifying key thematic areas, tracking their development over time, and assessing their academic impact. To achieve this, we apply an unsupervised machine learning technique Latent Dirichlet Allocation (LDA) to analyze the abstracts of over 2,000 scholarly articles. Our analysis reveals 100 distinct research topics, with 15 of the most prominent themes featured in a science map designed to support future exploration and understanding of RPA's expanding research frontier.

翻译：机器人流程自动化（RPA）已迅速发展成为一种广受认可且极具影响力的软件技术。其日益增长的重要性激发了跨多个学科的不同研究努力。本研究旨在通过识别关键主题领域、追踪其随时间的发展并评估其学术影响力，来描绘RPA的科学图景。为此，我们应用一种无监督机器学习技术——潜在狄利克雷分布（LDA），分析了超过2000篇学术论文的摘要。我们的分析揭示了100个不同的研究主题，其中15个最突出的主题被呈现在一幅科学地图中，该地图旨在支持未来对RPA不断扩展的研究前沿的探索和理解。