The COVID-19 pandemic has disrupted teaching over the past year, pushing many instructors and students into remote learning. These changes have forced new discussions about serious issues with the digital divide and an array of intersectional inequities, and they have prompted conversations about the physical and mental health of everyone involved. While initial transitions to remote learning were treated as distinct from previous in-person or online learning, increasingly we are seeing a push to “return to normal.” This essay argues that pandemic recoveries take many forms, and risk and uncertainty must continue to shape our teaching. We must continue to engage with critical issues related to inequity, intersectionality, and broad discussions of health if we are to ensure a safe return.

Omnidirectional cameras that give a 360degrees panoramic view of the surroundings have recently been used in many applications such as robotics, navigation, and surveillance. This paper describes the application of parametric ego-motion estimation for vehicle detection to perform surround analysis using an automobile-mounted camera. For this purpose, the parametric planar motion model is integrated with the transformations to compensate distortion in omnidirectional images. The framework is used to detect objects with independent motion or height above the road. Camera calibration as well as the approximate vehicle speed obtained from a CAN bus are integrated with the motion information from spatial and temporal gradients using a Bayesian approach. The approach is tested for various configurations of an automobile-mounted omni camera as well as a rectilinear camera. Successful detection and tracking of moving vehicles and generation of a surround map are demonstrated for application to intelligent driver support.

Citrus huanglongbing (HLB) associated with the pathogen Candidatus Liberibacter asiaticus (Las) is a devastating disease for the US citrus industry (1, 2). To gain knowledge on the mechanism(s) by which Las evades host defense responses, we first expressed salicylate hydroxylase (sahA) of Las in Escherichia coli. Our data indicate that Las encodes a functional salicylate hydroxylase, which converts salicylic acid (SA) into catechol, a product that does not induce resistance. The sahA gene was highly induced in planta compared to psyllid vector suggesting its important role in disease progression. To determine expression level of defense related genes after Las infection, Xanthomonas axonopodis pv. citri strain A^W (Xac A^W) was used to induce PR gene expression. The PR-1 gene expression in Xac A^w-challenged plants, which were previously infected with Las, was lower than Xac A^w-challenged healthy plants. Using SA biosensor strain (Acinetobacter sp. ADPWH_lux), 4 fold reduction in SA accumulation was observed in the Las-infected as compared to healthy plants. To understanda possible synergistic effect of the presence of Las on the citrus canker [caused by X. citri subsp. citri (Xcc)] we inoculated Xcc in Las-infected and healthy leaves of grapefruit. The population levels of Xcc were significantly higher during all the observation time points (up to 14 days) in Las-infected as compared to healthy citrus indicating that modulation of SA production and subsequent regulation of defense related genes such as PR-1 gene could be one of the mechanisms deployed by Las to evade plant defense responses. The Las-infected plants compromised with defense responses could further succumb to the infection by other pathogens. We also conducted experiments to restore the SA level in Las infected plants using SA hydroxylase inhibitors and test their effect in controlling HLB.

Microorganisms found in the environment around us exhibit a multitude of desirable characteristics that make them suitable hosts for the industrial production of biochemicals and biofuels. A common strategy to engineer microbes is using rational design approaches that entail manipulation of genes involved in one or more native metabolic pathways to improve biochemical synthesis or other relevant phenotypes. The advent of synthetic biology tools such as CRISPR-Cas systems for gene editing, and next-generation sequencing technologies has, however, made it possible to elucidate pangenome-wide targets for strain engineering by performing experiments at a genomic scale, such as genome-wide pooled CRISPR knockout screens, and analyzing the resulting high-throughput data using bioinformatic methods. A crucial challenge in evaluating the outcomes of pooled CRISPR screens is accounting for the variability in sgRNA knockout efficiency, as low-activity guides can potentially mask screening hits to result in false negatives. Towards that end, we developed an analysis method, acCRISPR, that processes NGS data from pooled CRISPR knockout screens, and provides an activity correction to accurately call statistically significant genes for the phenotype under study. We applied acCRISPR to CRISPR-Cas9 and CRISPR-Cas12a screening datasets from the oleaginous yeast Yarrowia lipolytica to identify a high-confidence set of essential genes for growth on glucose, as well as genes important for providing tolerance to high salt stress conditions. We further used the experimental sgRNA activity profiles from these screens to determine in silico sgRNA activity prediction accuracy of deep learning-based models trained on balanced and imbalanced experimental datasets, and improve prediction power with imbalanced training datasets by augmenting them with synthetic sgRNA. In another study, we sought to identify genetic targets responsible for phenazine biosynthesis in the bacterium Pseudomonas chlororaphis by employing a population genomics approach. We sequenced 34 Pseudomonas isolates using short- and long-read sequencing technologies, characterized them for phenazine production, and performed a microbial genome-wide association study (mGWAS) on the genomic-phenotypic data to elucidate the most influential phenazine biosynthesis targets across the pangenome. Overall, this work demonstrates the utility of high-throughput experimental-computational frameworks for identifying microbial strain engineering targets at a genomic scale and establishing novel genotype-phenotype relationships.

Modern vehicular automation technology is enabled by complex interactive systems composed of discrete blocks that accomplish specific tasks. These blocks are responsible for perceiving the environment surrounding the vehicle, predicting the future states and intents of other agents in its vicinity, planning the vehicle's path with an eventual goal in mind, and finally, providing control signals and actuations that achieve this goal. While such systems are widely adopted, they fail to consider a key component of the driving process - the human occupants inside the vehicle. This dissertation proposes that to ensure a safe, smooth, and enjoyable travel experience, it is essential for the human driver and vehicular automation to be aware of each other's states and impending failures, i.e., a relationship built on earned trust rather than blind faith.

To codify this requirement, we employ the Looking-In & Looking-Out (LILO) approach, where systems simultaneously model the inside and outside of vehicles. In addition to the set of "looking-out" tasks that are used in conventional automation, we propose a parallel set of "looking-in" tasks that model and analyze the interior of the vehicle. The outputs generated by both these parallel systems can then be integrated to provide useful controls based on a more complete understanding of the circumstances and the environment.

This dissertation presents the models and algorithms we have developed to accomplish the individual tasks that make up the system described above. This includes our research on real-time models for tasks such as object detection, multi-object tracking, and maneuver/trajectory prediction. We also introduce new approaches for the relatively under-explored "looking-in" tasks - which we hope will inspire many such studies in the future. Some of our contributions in this space include real-time models to analyze different aspects of the driver's state, data augmentation and automatic labelling schemes for tasks with limited data, and solutions to other common hindrances. Finally, we provide a concrete example of how the LILO framework can be used in practice - where the driver’s state estimated by "looking-in" is used in conjunction with metrics computed by "looking-out" to initiate safer and smoother control transitions.

Behavior analysis of vehicles surrounding the ego-vehicle is an essential component in safe and pleasant autonomous driving. This study develops a framework for activity classification of observed on-road vehicles using 3D trajectory cues and a Long Short Term Memory (LSTM) model. As a case study, we aim to classify maneuvers of surrounding vehicles at four way intersections. A variety of sensros including stereo cameras, LIDAR, GPS, and IMU measurements are used to extract ego-motion compensated surround trajectories from five different datasets. The impact of different prediction label space choices, feature space input, noisy/missing trajectory data, and LSTM model architectures are analyzed, presenting the strengths and limitations of the proposed approach.