Martin G. Frasch

Martin G. Frasch

Modern machine learning optimizes for accuracy without explicitly accounting for internal computational cost, even though physical and biological systems operate under intrinsic energy constraints. We evaluate energy-aware learning across 2,203 experiments spanning vision, text, neuromorphic, and physiological datasets, using 10 seeds per configuration and performing a factorial statistical analysis. Three findings emerge. First, architecture alone explains negligible variance in accuracy (partial eta^2 = 0.001). In contrast, the architecture x dataset interaction is large (partial eta^2 = 0.44, p < 0.001), demonstrating that optimal architecture depends critically on task modality and rejecting the assumption of a universal best architecture. Second, a controlled lambda-sweep over four orders of magnitude validates a single-parameter energy-regularized objective L = L_CE + lambda * E(theta, x): internal activation energy decreases to 6% of baseline at moderate lambda with no accuracy degradation on MNIST. Third, energy-first architectures inspired by an action-principle framework yield 5-33% within-modality training-efficiency gains over conventional baselines. These results emerge from a research program that interprets learning through a structural correspondence between the action functional in classical mechanics, free energy in statistical physics, and KL-regularized objectives in variational inference. We frame this correspondence as a design hypothesis rather than a derivation.

Martin G. Frasch

Identifying physical laws from noisy observational data is a central challenge in scientific machine learning. We present Minimum-Action Learning (MAL), a framework that selects symbolic force laws from a pre-specified basis library by minimizing a Triple-Action functional combining trajectory reconstruction, architectural sparsity, and energy-conservation enforcement. A wide-stencil acceleration-matching technique reduces noise variance by 10,000x, transforming an intractable problem (SNR ~0.02) into a learnable one (SNR ~1.6); this preprocessing is the critical enabler shared by all methods tested, including SINDy variants. On two benchmarks -- Kepler gravity and Hooke's law -- MAL recovers the correct force law with Kepler exponent p = 3.01 +/- 0.01 at ~0.07 kWh (40% reduction vs. prediction-error-only baselines). The raw correct-basis rate is 40% for Kepler and 90% for Hooke; an energy-conservation-based criterion discriminates the true force law in all cases, yielding 100% pipeline-level identification. Basis library sensitivity experiments show that near-confounders degrade selection (20% with added r^{-2.5} and r^{-1.5}), while distant additions are harmless, and the conservation diagnostic remains informative even when the correct basis is absent. Direct comparison with noise-robust SINDy variants, Hamiltonian Neural Networks, and Lagrangian Neural Networks confirms MAL's distinct niche: interpretable, energy-constrained model selection that combines symbolic basis identification with dynamical rollout validation.

Martin G. Frasch, Marlene J. E. Mayer, Clara Becker et al.

Prenatal psychological stress affects 15-25% of pregnancies and increases risks of preterm birth, low birth weight, and adverse neurodevelopmental outcomes. Current screening relies on subjective questionnaires (PSS-10), limiting continuous monitoring. We developed deep learning models for stress detection from electrocardiography (ECG) using the FELICITy 1 cohort (151 pregnant women, 32-38 weeks gestation). A ResNet-34 encoder was pretrained via SimCLR contrastive learning on 40,692 ECG segments per subject. Multi-layer feature extraction enabled binary classification and continuous PSS prediction across maternal (mECG), fetal (fECG), and abdominal ECG (aECG). External validation used the FELICITy 2 RCT (28 subjects, different ECG device, yoga intervention vs. control). On FELICITy 1 (5-fold CV): mECG 98.6% accuracy (R2=0.88, MAE=1.90), fECG 99.8% (R2=0.95, MAE=1.19), aECG 95.5% (R2=0.75, MAE=2.80). External validation on FELICITy 2: mECG 77.3% accuracy (R2=0.62, MAE=3.54, AUC=0.826), aECG 63.6% (R2=0.29, AUC=0.705). Signal quality-based channel selection outperformed all-channel averaging (+12% R2 improvement). Mixed-effects models detected a significant intervention response (p=0.041). Self-supervised deep learning on pregnancy ECG enables accurate, objective stress assessment, with multi-layer feature extraction substantially outperforming single embedding approaches.

Martin G. Frasch, Shadrian B. Strong, David Nilosek et al.

Despite broad application during labor and delivery, there remains considerable debate about the value of electronic fetal monitoring (EFM). EFM includes the surveillance of the fetal heart rate (FHR) patterns in conjunction with the maternal uterine contractions providing a wealth of data about fetal behavior and the threat of diminished oxygenation and perfusion. Adverse outcomes universally associate a fetal injury with the failure to timely respond to FHR pattern information. Historically, the EFM data, stored digitally, are available only as rasterized pdf images for contemporary or historical discussion and examination. In reality, however, they are rarely reviewed systematically. Using a unique archive of EFM collected over 50 years of practice in conjunction with adverse outcomes, we present a deep learning framework for training and detection of incipient or past fetal injury. We report 94% accuracy in identifying early, preventable fetal injury intrapartum. This framework is suited for automating an early warning and decision support system for maintaining fetal well-being during the stresses of labor. Ultimately, such a system could enable a physician to timely respond during labor and prevent adverse outcomes. When adverse outcomes cannot be avoided, they can provide guidance to the early neuroprotective treatment of the newborn.

Yu-Ting Lin, Yu-Lun Lo, Chen-Yun Lin et al.

Object: It is increasingly popular to collect as much data as possible in the hospital setting from clinical monitors for research purposes. However, in this setup the data calibration issue is often not discussed and, rather, implicitly assumed, while the clinical monitors might not be designed for the data analysis purpose. We hypothesize that this calibration issue for a secondary analysis may become an important source of artifacts in patient monitor data. We test an off-the-shelf integrated photoplethysmography (PPG) and electrocardiogram (ECG) monitoring device for its ability to yield a reliable pulse transit time (PTT) signal. Approach: This is a retrospective clinical study using two databases: one containing 35 subjects who underwent laparoscopic cholecystectomy, another containing 22 subjects who underwent spontaneous breathing test in the intensive care unit. All data sets include recordings of PPG and ECG using a commonly deployed patient monitor. We calculated the PTT signal offline. Main Results: We report a novel constant oscillatory pattern in the PTT signal and identify this pattern as a sawtooth artifact. We apply an approach based on the de-shape method to visualize, quantify and validate this sawtooth artifact. Significance: The PPG and ECG signals not designed for the PTT evaluation may contain unwanted artifacts. The PTT signal should be calibrated before analysis to avoid erroneous interpretation of its physiological meaning.

Ruilin Li, Martin G. Frasch, Hau-tieng Wu

There is a need for affordable, widely deployable maternal-fetal ECG monitors to improve maternal and fetal health during pregnancy and delivery. Based on the diffusion-based channel selection, here we present the mathematical formalism and clinical validation of an algorithm capable of accurate separation of maternal and fetal ECG from a two channel signal acquired over maternal abdomen.