Abstract:This paper proposes a methodology for discovering meaningful properties in data by exploring the latent space of unsupervised deep generative models. We combine manipulation of individual latent variables to extreme values with methods inspired by causal inference into an approach we call causal disentanglement with extreme values (CDEV) and show that this method yields insights for model interpretability. With this, we can test for what properties of unknown data the model encodes as meaningful, using it to glean insight into the communication system of sperm whales (Physeter macrocephalus), one of the most intriguing and understudied animal communication systems. The network architecture used has been shown to learn meaningful representations of speech; here, it is used as a learning mechanism to decipher the properties of another vocal communication system in which case we have no ground truth. The proposed methodology suggests that sperm whales encode information using the number of clicks in a sequence, the regularity of their timing, and audio properties such as the spectral mean and the acoustic regularity of the sequences. Some of these findings are consistent with existing hypotheses, while others are proposed for the first time. We also argue that our models uncover rules that govern the structure of units in the communication system and apply them while generating innovative data not shown during training. This paper suggests that an interpretation of the outputs of deep neural networks with causal inference methodology can be a viable strategy for approaching data about which little is known and presents another case of how deep learning can limit the hypothesis space. Finally, the proposed approach can be extended to other architectures and datasets.

Machine-Guided Discovery of a Real-World Rogue Wave Model

An Intelligent Model with Chaos and Causality

Early Detection of the Draupner Wave Using Deep Learning

Robust data driven discovery of a seismic wave equation

Rogue waves and large deviations in deep sea

Prediction of freak waves from buoy measurements

A Novel Hybrid Machine Learning Model for Rapid Assessment of Wave and Storm Surge Responses Over an Extended Coastal Region

Predicting beach profiles with machine learning from offshore wave reflection spectra

A Machine Learning Framework to Forecast Wave Conditions

Data-driven modeling of nonlinear traveling waves

AI Descartes: Combining Data and Theory for Derivable Scientific Discovery

Study of Nonlinear Interaction between Waves and Ocean Currents Using High-Fidelity Simulation and Machine Learning

Deterministic wave prediction model for irregular long-crested waves with Recurrent Neural Network

Random walks across the sea: the origin of rogue waves?

Mining Causality from Continuous-time Dynamics Models: An Application to Tsunami Forecasting

Data-driven rogue waves solutions for the focusing and variable coefficient nonlinear Schrödinger equations via deep learning

Phase-resolved wave prediction with varying buoy positions in the field using machine learning-based methods

Data-Driven system identification of 6-DoF ship motion in waves with neural networks

Linear and Nonlinear Rogue Wave Statistics in the Presence of Random Currents

Machine learning for naval architecture, ocean and marine engineering

Approaching an unknown communication system by latent space exploration and causal inference