Abstract Decision tree ensembles (DTE) are a popular model for a wide range of AI classification tasks, used in multiple safety critical domains, and hence verifying properties on these models has been an active topic of study over the last decade. One such verification question is the problem of sensitivity, which asks, given a DTE, whether a small change in subset of features can lead to misclassification of the input. In this work, our focus is to build a quantitative notion of sensitivity, tailored to DTEs, by discretizing the input space of the model and enumerating the regions which are susceptible to sensitivity. We propose a novel algorithmic technique that can perform this computation efficiently, within a certified error and confidence bound. Our approach is based on encoding the problem as an algebraic decision diagram (ADD), and further splitting it into subproblems that can be solved efficiently and make the computation compositional and scalable. We evaluate the performance of our technique over benchmarks of varying size in terms of number of trees and depth, comparing it against the performance of model counters over the same problem encoding. Experimental results show that our tool $$\textsf{EnSensCount}$$ EnSensCount achieves significant speedup over other approaches and can scale well with the increasing sizes of the ensembles.

Advances in virtual reality (VR) are transforming interaction paradigms by shifting towards gesture-based control driven by physiological sensing, enabling more intuitive and embodied experiences. Surface electromyography (sEMG) is emerging as a reliable modality for this hands-free and expressive gesture recognition in VR. However, prolonged mid-air gestures can lead to muscle fatigue and physiological changes that degrade overall recognition performance. Further, this degradation is not uniform across gestures which can impact user performance and experience in VR applications. While existing literature has shown that fatigue alters sEMG signals, its effects during extended immersive interaction and across various gestures remain underexplored. We conducted a 35-participant study in which each participant continuously performed five gesture in VR for 20 minutes each, while we collected high-resolution sEMG data from eight forearm sensors and real-time subjective fatigue ratings using the Borg CR10 scale. Further, we evaluate how gesture recognition models behave under fatigue and explore the impact of incorporating both objective (signal-derived) and subjective (user-reported) fatigue features into classification models. Our results show that integrating fatigue signals enhances model robustness and improves recognition accuracy during extended use.