# A Neuro-Symbolic System for Interpretable Multimodal Physiological Signals Integration in Human Fatigue Detection

arXiv: https://arxiv.org/abs/2603.24358 · HTML: https://mlsanity.com/paper/2603.24358/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 9 of its 28 references that are in the map.
- First posted on arXiv 25 Mar 2026.

## Builds on (most cited first, 9 of 9)

- [On Calibration of Modern Neural Networks](https://mlsanity.com/paper/1706.04599/) (2017)
- [Dropout as a Bayesian Approximation: Representing Model Uncertainty in Deep Learning](https://mlsanity.com/paper/1506.02142/) (2015)
- [Simple and Scalable Predictive Uncertainty Estimation using Deep Ensembles](https://mlsanity.com/paper/1612.01474/) (2016)
- [Concept Bottleneck Models](https://mlsanity.com/paper/2007.04612/) (2020)
- [Post-hoc Concept Bottleneck Models](https://mlsanity.com/paper/2205.15480/) (2022)
- [From Statistical Relational to Neurosymbolic Artificial Intelligence: a Survey](https://mlsanity.com/paper/2108.11451/) (2021)
- [Neuro-Symbolic AI in 2024: A Systematic Review](https://mlsanity.com/paper/2501.05435/) (2025)
- [An End-to-End Review of Gaze Estimation and its Interactive Applications on Handheld Mobile Devices](https://mlsanity.com/paper/2307.00122/) (2023)
- [Quantifying the Impact of Motion on 2D Gaze Estimation in Real-World Mobile Interactions](https://mlsanity.com/paper/2502.10570/) (2025)

## Abstract (arXiv)

We propose a neuro-symbolic architecture that learns four interpretable physiological concepts, oculomotor dynamics, gaze stability, prefrontal hemodynamics, and multimodal, from eye-tracking and neural hemodynamics, functional near-infrared spectroscopy, (fNIRS) windows using attention-based encoders, and combines them with differentiable approximate reasoning rules using learned weights and soft thresholds, to address both rigid hand-crafted rules and the lack of subject-level alignment diagnostics. We apply this system to fatigue classification from multimodal physiological signals, a domain that requires models that are accurate and interpretable, with internal reasoning that can be inspected for safety-critical use. In leave-one-subject-out evaluation on 18 participants (560 samples), the method achieves 72.1% +/- 12.3% accuracy, comparable to tuned baselines while exposing concept activations and rule firing strengths. Ablations indicate gains from participant-specific calibration (+5.2 pp), a modest drop without the fNIRS concept (-1.2 pp), and slightly better performance with Lukasiewicz operators than product (+0.9 pp). We also introduce concept fidelity, an offline per-subject audit metric from held-out labels, which correlates strongly with per-subject accuracy (r=0.843, p < 0.0001).