eSLIM++ - An Efficient Subjective Logic Implementation in C++ Providing Easy-to-Use Python Interfaces
Thomas Wodtko, Thomas Griebel, Michael Buchholz · 2025
In recent years, subjective logic has gained significant attention for applications in reasoning under uncertainties. By explicitly considering the evidence in each statement or trust, it provides a second-order probability framework that also relates closely to the Dempster-Shafer theory of evidence. Despite theoretical richness, its practical adoption remains limited due to the lack of efficient yet accessible implementations. However, lowering the barrier to the reproducibility of current research findings is key to boosting progress in applying subjective logic to solve contemporary questions. With this work, we present our novel eSLIM++ library, which aims to close this gap between state-of-the-art contributions to subjective logic and their application in real-world scenarios. The library's efficient C++ and CUDA-compatible implementation makes it suitable for evaluating large-scale and complex decision processes. Additionally, Python interfaces provide easy-to-use, low-barrier access that allows for rapid development and evaluation and provides comprehensive examples on a smaller scale. eSLIM++ is publicly available and open-source. It enables researchers and the community to extend its functionality and deploy subjective logic in modern applications such as object tracking and autonomous systems.