The performance of deep learning frameworks could be significantly improved through considering the particular underlying structures for each dataset. In this thesis, I summarize our three work about boosting the performance of deep learning models through leveraging structures of the data. In the first work, we theoretically justify that, for...
Polycyclic aromatic hydrocarbons (PAHs) are a group of environmental contaminants consisting of fused benzene rings. Parent-PAHs, methylated-PAHs (MPAHs), and PAHs with molecular weight of 302 a.m.u (MW302-PAHs) are considered as unsubstituted-PAHs. These unsubstituted-PAHs undergo transformation reactions resulting in the formation of PAH-transformation products (PAH-TPs), or substituted-PAHs, including nitrated-, oxygenated-, and...
The role of community colleges as open-access institutions that bring racial diversity to careers in the science, technology, engineering, and mathematics (STEM) fields is essential (Musante, 2012; Reyes, 2011). Yet, the opportunity to attend postsecondary institutions is not enough to guarantee the success of People of Color as they navigate...
River basins provide essential services for both humans and ecosystems. Understanding the connections between ecosystems and society and their function has been at the heart of resilience studies and has become an increasing important endeavor in research and practice. In this dissertation, I define basin resilience as a river basin...
This dissertation incorporates coalition formation and probabilistic planning towards a domain-independent automated planning solution scalable to multiple heterogeneous robots in complex domains. The first research direction investigates the effectiveness of Task Fusion and introduces heuristics that improve task allocation and result in better quality plans, while requiring lower computational cost...
Decision-making is a highly complex process influenced by the values, attitudes, and cognitive biases of the decision-maker. The classical model of decision-making fails to fully incorporate how decisions are made in situations of uncertainty, where having complete knowledge of all possible outcomes and alternatives is not always possible. In highly...
In this thesis, we introduce a novel Explanation Neural Network (XNN) to explain the predictions made by a deep network. The XNN works by embedding a high-dimensional activation vector of a deep network layer non-linearly into a low-dimensional explanation space while retaining faithfulness i.e., the original deep learning predictions can...
Anammox (anaerobic ammonium oxidation) bacteria are capable of providing low-cost nitrogen removal for numerous types of wastewaters. However, low growth rates cause long startup-times and inhibition by oxygen and metabolic substrates necessitate close process control to maintain performance. Incorporation of Simultaneous Anammox and Denitrification (SAD) into constructed wetlands could provide...