Our study aims to determine how physiological measures of affect such as heart rate variability and galvanic skin response may be linked to other representations of emotion that arise naturally in text-based communication during online collaboration. Research in this area is important because people increasingly collaborate in a distributed manner across greater distances, and text-based communication is one of the primary means of facilitating these collaborations. The inspiration for this work comes from Aragon and Williams' (2011) paper, which introduces a complex systems model of collaborative creativity and describes ways that affect contributes to this model. In this study, we test and validate the model by using physiological measures to detect and explore the relationship between affect and creativity. Possible outcomes include correlations of creative events with physiological measures of affect and/or the occurrence of affect in chat messages.

 The performance of machine learning (ML) classification algorithms in an open-ended problem with manual labels is difficult to assess, because errors can exist both in classification and manually labeled datasets. An additional dataset of manually re-annotated data can help to differentiate classifier and manually labeled data errors; however, there is currently no effective way to visualize all three datasets simultaneously. To address this problem, we introduce a new type of visualization called a confusion diamond. The confusion diamond consists of 8 triangles, with each comparing different dimensions of all three datasets. The arrangement of the triangles complement traditional confusion matrices and utilize many of their visual properties. The confusion diamond layout allows users to pre-attentively perceive data errors as distinct from classification errors, which makes the visualization easy to use. Multiple confusion diamonds can be analyzed at once, making it useful for datasets containing multiple labels, and giving researchers the ability to quickly identify which labels have errors and how to solve them. This work presents the key design elements of the confusion diamond visualization, relevant usage scenarios, and findings from semi-structured interviews with members of a research team interested in using the confusion diamond to help them visualize ML classification performance on large social media datasets.

Methylotrophic bacteria have long been recognized for their metabolism of reduced single carbon (C₁) compounds. Their potential for biotechnology and the occurrence of C₁ metabolism in all organisms signify their importance on an industrial and fundamental scale. Currently, mechanisms used by methylotrophs to regulate and reset their metabolic network in response to perturbations are not well understood. Previous studies suggest abrupt shifts in nucleotide levels drastically affect growth. This project will examine how nucleotide ratios correlate with carbon flux in Methylobacterium extorquens AM1. Once C₁ compounds undergo serial oxidation to formate, carbon flux is partitioned for further oxidation or assimilation. Two reactions involved in C₁ oxidation are known to generate reducing power: (1) the dehydrogenation of methylene-dH₄MPT into methenyl-dH₄MPT catalyzed by MtdA or MtdB and (2) the oxidation of formate into CO₂ catalyzed by Fdh1 or Fdh2. Genetic studies demonstrated differential phenotypes in which mtdB fdh1 grew similarly to wild type while mtdB fdh2 grew worse relative to the mtdB mutant strain. One possible explanation is a global decrease in NADH production offsets the delicate balance of intracellular nucleotides required for an already reducing power-limited mode of growth. To explore this hypothesis, I will conduct experiments to: (1) quantify the levels of NADP⁺, NADPH, NAD⁺, and NADH in WT, mtdB, mtdB fdh1, and mtdB fdh2 and (2) manipulate nucleotide ratios with overexpression and deletion of the pntAB and udhA genes. These genes encode for membrane bound (PntAB) and soluble (UdhA) transhydrogenases capable of interconverting between NADPH and NADH. Since mechanisms exist in other bacteria to alter nucleotides pools to meet metabolic demands, a similar mechanism can be predicted to operate in M. extorquens AM1. Success in this project will lead to better understanding of how cells balance reducing power and energy for oxidation and growth, ultimately leading to improved biotechnological utilization.

Prior to about 650 million years ago, the Earth's surface is believed to have completely frozen and thawed multiple times, each frozen period lasting for millions of years. The term "Snowball Earth" describes the runaway effect of ice expanding equatorward, with its brightly reflecting surface causing reduced absorption of sunlight, increased cooling, and further ice expansion. We used Global Climates Models to better understand the conditions necessary for other planets to be trapped in a Snowball state. We modeled planets that are Earth-like but with ocean-only surfaces. Our object is to investigate and further define the inner edge of the habitable zone - the theoretical region around a star where liquid water can exist on the surface. Our research has big implications for habitability, as liquid water is required for life to exist as we know it. The obliquity of a planet is also very important, as it controls the relative proportion of sunlight received at each latitude. Even within our own solar system there is a wide range. The Earth is tilted 23 degrees, Uranus 97, and Venus 177. We present results for an ensemble of planets with differing obliquities to demonstrate its effects. Furthermore, we also present results for planets around G-type stars, like our Sun, and M-type stars, also known as red dwarfs. Red dwarfs are by far the most common type of star in the Milky Way Galaxy. It is easier to detect orbiting extrasolar planets around M-stars. M-stars tend to have a relatively high abundance of Earth-sized and smaller planets, which are more likely to be rocky - a necessary condition for habitability. The effect of an M star’s solar spectrum on a planet’s energy balance is significantly different than a G star’s, which changes the location of the habitable zone. Our model study provides exciting new insights into the differing dynamics.

Everyday, people's lives are affected by the weather; whether sunshine or rain, weather is a part of daily life. Extreme weather events can bring normal activities to a halt. Weather events known as "blocking" involve a pause in the of the typical passage of storms in the jet stream, causing weather systems to stay in place for lengthened periods of time. Wintertime blocking events often cause extreme cold winter temperatures over land, and data show that blocking frequency is increasing. A potential cause for this is due to sea ice loss and the warming of the Atlantic Ocean. Our investigation is of the extraordinary January 2012 blocking event over Europe. The nature of this extreme event was tested using the Weather and Research Model applied to the domain including the Western Arctic Ocean, the North Atlantic Ocean, and Europe, with a resolution of 40 kilometers. The event was simulated over the period from January 15, 2012 to February 6, 2012, with horizontal boundaries and initial conditions constrained by atmospheric observations. In a series of model integrations, we manipulated the thickness of sea ice and sea surface temperature in high altitudes and found that there is a link between thinner sea ice and warmer ocean temperatures and an increased intensity of the blocking event over Europe. This discovery allows for a better understanding of the potential impacts of melting sea ice as the climate changes.