Physical Sciences and Mathematics Commons^™

Towards Improving Accuracy And Interpretability Of Deep Learning Based On Satellite Image Classification, Yamile Patino Vargas

Dissertations and Theses

ABSTRACT

The study of satellite images provides a way to monitor changes in the surface of the Earth and the atmosphere. Convolutional Neural Networks (CNN) have shown accurate results in solving practical problems in multiple fields. Some of the more recognized fields using CNNs are satellite imagery processing, medicine, communication, transportation, and computer vision. Despite the success of CNNs, there remains a need to explain the network predictions further and understand what the network is determining as valuable information.

There are several frameworks and methodologies developed to explain how CNNs predict outputs and what their internal representations are [1, 4, …

Microfluidic Study Of Gravity-Driven Drainage And Coalescence Of Aqueous Two Dimensional Foams, Justin D. Heftel

Dissertations and Theses

Foams, a two-phase dispersion, are staples of the cosmetic, personal care, petroleum, pharmaceutical, and other industries. Central to these applications is the stability of the dispersion against separation. Foams break down by two mechanisms: the first is bubble coalescence, which is driven by the gravity drainage of the continuous phase. The drainage acts to push the bubbles against each other, and leads to the formation of thin lamellae, which break and cause the coalescence. The second is the mass transfer of the dispersed phase through the continuous phase, which is caused by the difference in pressures between the bubbles and …

Hydroclimate Drivers And Atmospheric Dynamics Of Floods, Nasser Najibi

Dissertations and Theses

Our preliminary survey showed that most of the recent flood-related studies did not formally explain the physical mechanisms of long-duration and large-peak flood events that can evoke substantial damages to properties and infrastructure systems. These studies also fell short of fully assessing the interactions of coupled ocean-atmosphere and land dynamics which are capable of forcing substantial changes to the flood attributes by governing the exceeding surface flow regimes and moisture source-sink relationships at the spatiotemporal scales important for risk management. This dissertation advances the understanding of the variability in flood duration, peak, volume, and timing at the regional to the …