12023 research outputs found
Sort by
The Developmental Organization of Regulatory States in the Sea Urchin Larva
Get PDFDevelopment is an inherently dynamic process where cell fate specification occurs continuously and in a progressive manner. Thus, a major focus in developmental biology is solving the gene regulatory networks (GRNs) that underlie specification of cell fates. GRNs specify new spatial domains of cells by controlling the expression of their changing regulatory states throughout development. Regulatory states are composed of combinations of expressed regulatory genes which encode transcription factors that form regulatory circuits which function to carry out the specific developmental tasks involved in cell fate specification.
To investigate the differences of GRNs operating in the embryo and their change over development, we sought to identify and characterize the regulatory states present in multiple developmental stages of sea urchin embryogenesis. We performed a genome-wide survey and embryo-wide annotation of regulatory gene expression by whole mount in situ hybridization at five consecutive developmental time-points in order to determine regulatory states and their developmental trajectory. We determined at least 74 distinct regulatory states expressed in discrete developmental domains which coincide with larval morphological structures and show that their progenitor domains foreshadow the ensuing larval morphology. Among these domains, we identified bilateral ciliary photoreceptors in the larva which express a distinct regulatory state that include factors known in ciliary photoreceptor specification. We show that this photoreceptor regulatory state does not express the genes of the retinal determination network that specify eyes in both flies and vertebrates. In addition, we show that though the sizes of regulatory states are comparable over developmental time, no two regulatory states are equal, even those expressed in a given domain at previous or subsequent developmental time-points. Lastly, we found that similarities among regulatory states reflect a common developmental function but not necessarily a common developmental history. The results suggest that the combinations of transcription factors defining regulatory states are both spatially and temporally dynamic in their progressive specification of cell fates during development and that regulatory state expression is tightly associated with the developing morphology of the larva.</p
The Discovery of Novel Materials for the Electrocatalytic Reduction of Carbon Dioxide
Get PDFThe removal of atmospheric carbon dioxide is likely the only route to mitigating the effects of decades of increased fossil fuel combustion. Artificial photosynthesis presents one method for removal and conversion of problematic carbon dioxide into chemically useful products. By coupling electrochemical CO2 reduction (CO2R) to a renewable energy source atmospheric CO2 could be converted back into a fuel such as ethanol, or a commodity chemical such as ethylene. These products could then be consumed for energy or used to generate plastics effectively removing CO2 from the atmosphere. Significant advances in current electrocatalysts are needed in order for large scale CO2R to become a reality. Most known catalysts are only capable of transferring 2 electrons with needed protons to CO2 producing either carbon monoxide or formic acid. Copper is the only known metal capable of reducing CO2 to hydrocarbons at appreciable rates and low overpotentials. This work aims to find new materials that produce similar hydrocarbons, but at lower overpotentials with higher rates and greater selectivity than current copper catalysts. By implementing a cyclic process referred to as the Catalyst Discovery Cycle (CDC) iterations between predications, catalyst testing, and active site characterization allow for the rational design and discovery of new and improved catalysts. This methodology led to the discovery of nickel-gallium bimetallics as low overpotential catalysts for CO2R to methane, ethylene, and ethane. In addition, theoretical and experimental observations have determined a proposed active site and side reactions detrimental to their activity
Optimizing Resource Management in Cloud Analytics Services
Get PDFThe fundamental challenge in the cloud today is how to build and optimize machine learning and data analytical services. Machine learning and data analytical platforms are changing computing infrastructure from expensive private data centers to easily accessible online services. These services pack user requests as jobs and run them on thousands of machines in parallel in geo-distributed clusters. The scale and the complexity of emerging jobs lead to increasing challenges for the clusters at all levels, from power infrastructure to system architecture and corresponding software framework design.
These challenges come in many forms. Today's clusters are built on commodity hardware and hardware failures are unavoidable. Resource competition, network congestion, and mixed generations of hardware make the hardware environment complex and hard to model and predict. Such heterogeneity becomes a crucial roadblock for efficient parallelization on both the task level and job level. Another challenge comes from the increasing complexity of the applications. For example, machine learning services run jobs made up of multiple tasks with complex dependency structures. This complexity leads to difficulties in framework designs. The scale, especially when services span geo-distributed clusters, leads to another important hurdle for cluster design. Challenges also come from the power infrastructure. Power infrastructure is very expensive and accounts for more than 20% of the total costs to build a cluster. Power sharing optimization to maximize the facility utilization and smooth peak hour usages is another roadblock for cluster design.
In this thesis, we focus on solutions for these challenges at the task level, on the job level, with respect to the geo-distributed data cloud design and for power management in colocation data centers.
At the task level, a crucial hurdle to achieving predictable performance is stragglers, i.e., tasks that take significantly longer than expected to run. At this point, speculative execution has been widely adopted to mitigate the impact of stragglers in simple workloads. We apply straggler mitigation for approximation jobs for the first time. We present GRASS, which carefully uses speculation to mitigate the impact of stragglers in approximation jobs. GRASS's design is based on the analysis of a model we develop to capture the optimal speculation levels for approximation jobs. Evaluations with production workloads from Facebook and Microsoft Bing in an EC2 cluster of 200 nodes show that GRASS increases accuracy of deadline-bound jobs by 47% and speeds up error-bound jobs by 38%.
Moving from task level to job level, task level speculation mechanisms are designed and operated independently of job scheduling when, in fact, scheduling a speculative copy of a task has a direct impact on the resources available for other jobs. Thus, we present Hopper, a job-level speculation-aware scheduler that integrates the tradeoffs associated with speculation into job scheduling decisions based on a model generalized from the task-level speculation model. We implement both centralized and decentralized prototypes of the Hopper scheduler and show that 50% (66%) improvements over state-of-the-art centralized (decentralized) schedulers and speculation strategies can be achieved through the coordination of scheduling and speculation.
As computing resources move from local clusters to geo-distributed cloud services, we are expecting the same transformation for data storage. We study two crucial pieces of a geo-distributed data cloud system: data acquisition and data placement. Starting from developing the optimal algorithm for the case of a data cloud made up of a single data center, we propose a near-optimal, polynomial-time algorithm for a geo-distributed data cloud in general. We show, via a case study, that the resulting design, Datum, is near-optimal (within 1.6%) in practical settings.
Efficient power management is a fundamental challenge for data centers when providing reliable services. Power oversubscription in data centers is very common and may occasionally trigger an emergency when the aggregate power demand exceeds the capacity. We study power capping solutions for handling such emergencies in a colocation data center, where the operator supplies power to multiple tenants. We propose a novel market mechanism based on supply function bidding, called COOP, to financially incentivize and coordinate tenants' power reduction for minimizing total performance loss while satisfying multiple power capping constraints. We demonstrate that COOP is "win-win", increasing the operator's profit (through oversubscription) and reducing tenants' costs (through financial compensation for their power reduction during emergencies).</p
Definable Combinatorics of Graphs and Equivalence Relations
Get PDFLet D = (X, D) be a Borel directed graph on a standard Borel space X and let χB(D) be its Borel chromatic number. If F0, …, Fn-1: X → X are Borel functions, let DF0, …, Fn-1 be the directed graph that they generate. It is an open problem if χB(DF0, …, Fn-1) ∈ {1, …, 2n + 1, ℵ0}. Palamourdas verified the foregoing for commuting functions with no fixed points. We show here that for commuting functions with the property that there is a path from each x ∈ X to a fixed point of some Fj, there exists an increasing filtration X = ⋃m < ω Xm such that χB(DF0, …, Fn-1↾ Xm) ≤ 2n for each m. We also prove that if n = 2 in the previous case, then χB(DF0, F1) ≤ 4. It follows that the approximate measure chromatic number χapM(D) ≤ 2n + 1 when the functions commute.
If X is a set, E is an equivalence relation on X, and n ∈ ω, then define [X]nE = {(x0, ..., xn - 1) ∈ nX: (∀i,j)(i ≠ j → ¬(xi E xj))}. For n ∈ ω, a set X has the n-Jónsson property if and only if for every function f: [X]n= → X, there exists some Y ⊆ X with X and Y in bijection so that f[[Y]n=] ≠ X. A set X has the Jónsson property if and only for every function f : (⋃n ∈ ω [X]n=) → X, there exists some Y ⊆ X with X and Y in bijection so that f[⋃n ∈ ω [Y]n=] ≠ X. Let n ∈ ω, X be a Polish space, and E be an equivalence relation on X. E has the n-Mycielski property if and only if for all comeager C ⊆ nX, there is some Borel A ⊆ X so that E ≤B E ↾ A and [A]nE ⊆ C. The following equivalence relations will be considered: E0 is defined on ω2 by x E0 y if and only if (∃n)(∀k > n)(x(k) = y(k)). E1 is defined on ω(ω2) by x E1 y if and only if (∃n)(∀k > n)(x(k) = y(k)). E2 is defined on ω2 by x E2 y if and only if ∑{1⁄(n + 1): x(n) ≠ y(n)} < ∞. E3 is defined on ω(ω2) by x E3 y if and only if (∀n)(x(n) E0 y(n)). Holshouser and Jackson have shown that ℝ is Jónsson under AD. The present research will show that E0 does not have the 3-Mycielski property and that E1, E2, and E3 do not have the 2-Mycielski property. Under ZF + AD, ω2/E0 does not have the 3-Jónsson property.
Let G = (X, G) be a graph and define for b ≥ 1 its b-fold chromatic number χ(b)(G) as the minimum size of Y such that there is a function c from X into b-sets of Y with c(x) ∩ c(y) = ∅ if x G y. Then its fractional chromatic number is χf(G) = infb χ(b)(G)⁄b if the quotients are finite. If X is Polish and G is a Borel graph, we can also define its fractional Borel chromatic number χfB(G) by restricting to only Borel functions. We similarly define this for Baire measurable and μ-measurable functions for a Borel measure μ. We show that for each countable graph G, one may construct an acyclic Borel graph G' on a Polish space such that χfBM(G') = χf(G) and χBM(G') = χ(G), and similarly for χfμ and χμ. We also prove that the implication χf(G) = 2 ⇒ χ(G) = 2 is false in the Borel setting.</p
Interfacial and Stability Studies of Photocathodes for Hydrogen Evolution
Get PDFPhotoelectrochemical (PEC) water splitting is a promising way to generate clean hydrogen fuel from water and sunlight. The ideal photocathodes for hydrogen evolution reaction (HER) should have good electrical contact and mechanical adhesion on the interface between the semiconductor and the catalyst, and be stable during operation. However, the interfacial properties and the stability have not been intensively studied. We investigated the electrical and mechanical properties on the nanoscale of the interface of commonly used Si/Pt nanoparticles (Pt-NPs) electrodes with Pt-NPs as a catalyst, and showed that the Pt-NPs have a weaker adhesion in electrolyte than in air, and less than half of the Pt-NPs carry high currents, limiting the performance of the common Si/Pt-NPs electrodes. Furthermore, we explored the interfacial engineering of using TiO2 deposited by atomic layer deposition (ALD), and showed that annealed TiO2 led to higher open circuit voltages than the as grown ones by the possible formation of an interfacial Si-O-Ti mixture layer. Besides, the stability and corrosion behavior of CdTe electrodes for HER in the dark was studied in 1.0 M H2SO4(aq) and 1.0 M KOH(aq). The conditions studied herein include the electrochemical corrosion when biased at -100 mV vs. the reversible hydrogen electrode (RHE), the chemical corrosion when left at open circuit voltage (OCV), and the electrochemical corrosion with an active HER Pt catalyst overlayer when biased at -100 mV vs. RHE. The corrosion comes mostly from chemical corrosion and is reduced at negative bias in electrochemical condition. With a Pt catalyst overlayer at -100 mV vs. RHE, the corrosion rate is further reduced, indicating the promising utilization of CdTe for HER in PEC cells.</p
The Role of Human Emotion and Character in Shakespeare's Othello vs. God and Heaven in Cinthio's "A Moorish Captain"
No full textShakespeare's Othello is based upon "A Moorish Captain," the seventh novella of the third decade of Giraldi Cinthio's Hecatommithi. Shakespeare's play is clearly inspired by Cinthio's short story as both works are tragedies with the same underlying plot - Othello, a Moorish general in the Venetian army is tricked into believing lies from his deceitful ensign, Iago, that implicate his wife, Desdemona, for being unfaithful. Although both works revolve around the same main characters, fundamental storyline, and similar themes of love, jealousy, betrayal, and revenge, Shakespeare introduces a number of variations to Cinthio's short story. The changes Shakespeare brings about consist of modifications such as an alternative route to the ending and also significant omissions and additions including new characters. Shakespeare implements alterations within his play Othello to downplay the role of inhuman forces such as God and Heaven and instead, emphasize that the action and events are a result of the characters' deep-rooted emotions and natures
On the Evolutionary Pathways of Stars and Extrasolar Planets
Get PDFIn this thesis, I present several studies aimed at determining the evolutionary states of stars and the planets that orbit them. Multiple approaches are undertaken to determine the physical parameters of stars over a wide range of masses, and in the process I evaluate current theoretical models which are commonly used to indirectly determine the properties of stars and planets.
Chapter II concerns the ages of nearby stars more massive than the Sun. These stars, because they are bright and young (less than 1 Gyr) on the average, constitute attractive targets for surveys aiming to directly capture light from planets in wide orbits. The precise masses of directly imaged companions are important for constraining star and planet formation theories, but rely critically on the host star ages. I show that sky-projected rotational velocity is a vital parameter in age-determination for intermediate-mass stars. Rapid rotation induces large pole-to-equator gradients in the photospheric temperature and surface gravity, such that a star seen pole-on appears hotter and higher gravity (and thus younger) than a star with identical properties seen edge-on. I use intermediate-band photometry centered on the hydrogen Balmer series and projected rotational velocities to determine atmospheric parameters and ages for approximately 3500 nearby stars with masses in the range of 1-10 solar masses. I validate the method using four open clusters, in the process finding ages for α Persei and the Hyades that are younger at ~70 Myr and older at ~830 Myr, respectively, than canonical values.
In Chapters III through V, I present orbital solutions and fundamental parameters for eclipsing binaries (EBs), newly discovered from the K2 mission, in the Pleiades open cluster (125 Myr) and the Upper Scorpius OB association (5-10 Myr). EBs, particularly those in coeval stellar populations, are valuable benchmarks for evaluating evolutionary models. Such benchmarks are particularly rare at low masses and young ages, and my work has increased the sample by 80% for stars less massive than the Sun and younger than 150 Myr. By jointly fitting eclipse observations and radial velocity measurements, one can directly determine the masses and radii of stars in an EB with percent-level precision. I use newly determined masses and radii to demonstrate a systematic temperature offset of approximately 200 K between empirical relations and model predictions. This result is in agreement with literature on the temperature suppression observed in low-mass stars, believed to be related to magnetic activity and inhibited convection. I show that stellar ages determined from the mass-radius diagram appear systematically older than those determined for the same stars in a Hertzsprung-Russell (H-R) diagram. A precise distance determination for the Pleiades EB HCG 76 is in agreement with the literature consensus and formally excludes the now discredited trigonometric measurement from Hipparcos. The orbital periods, eccentricities, and stellar spin periods determined from K2 photometry are compared with theoretical expectations from tidal dissipation. In Chapter VI, I present preliminary results on more recently discovered EBs in Upper Scorpius, and updated interpretations of previously published systems. I use the combined data to pave the way for an empirical pre-main-sequence mass-radius relation over a broad range of masses and determine an age of Upper Scorpius which is intermediate to the canonical age (3-5 Myr, as derived from low-mass pre-main-sequence stars) and a more recent estimate (9-13 Myr, as derived from intermediate-mass stars) from H-R diagram analyses.
In Chapter VII, I present observations of the low-mass pre-main-sequence star RIK-210. This star shows a variable eclipse-like signature, which persisted through 78 days of K2 photometry but was apparently absent in archival photometry. Follow-up observations demonstrate that RIK-210 is a single star with no massive companion orbiting at the period of the eclipse-like signature. The flux diminutions are in phase with the stellar rotation, behavior seen in some young stars that are periodically obscured by an accretion disk, but RIK-210 lacks such a protoplanetary disk. I consider various explanations for the observations and favor a model in which charged dust is trapped in a rigidly-rotating magnetosphere. The source of such dust could be from one or more close-in planets or residual planet-forming material drifting in towards the star.
In Chapter VIII I present the discovery and characterization of K2-33 b, a Neptune-sized planet closely orbiting a low-mass star in Upper Scorpius and one of the youngest exoplanets currently known. K2-33 b provides evidence that planets with substantial gaseous envelopes can be found close to their stars shortly after dispersal of the primordial protoplanetary disk. Given the planet's age (5-10 Myr), in situ formation or migration through the protoplanetary disk are the only plausible formation scenarios. K2-33 b is unusually large in size when compared to planets with similar orbital periods around low-mass field stars. I interpret this as tentative evidence that the planet is still contracting, experiencing photoevaporative atmospheric mass-loss, or both. In Chapter IX I present preliminary results from a study aimed at determining the prevalence of close-in planets at the epoch of primordial disk dispersal. Using K2 photometry for hundreds of pre-main-sequence stars in Upper Scorpius I search for transiting planets, assess survey completeness, and determine the occurrence rates or upper limits to such rates for large planets in close orbits around low-mass stars. With the singular detection of K2-33 b, I determine a rate of close-in Neptune- to Jupiter-sized planets higher than that for low-mass field stars but in closer agreement with the rate for sub-Neptune planets. Given the extreme youth of K2-33 b, I tentatively interpret these results as an indication that the planet is a progenitor of the abundant class of close-in sub-Neptunes.
Considered collectively, my results highlight the importance of benchmark systems. Nearly all branches of astrophysics are reliant in some regard on stellar evolutionary models, but the magnitude and sense of any particular systematic offset contained in these models is often poorly understood. Eclipsing binaries and planets located in stellar clusters provide firm anchors to theoretical models aiming to predict the evolution of such objects. The need for benchmarks and model calibration is particularly acute at young ages, when stars and planets are rapidly evolving and where theoretical evolutionary models are highly uncertain. While statistical studies of large populations can yield far-reaching results that move fields forward, it is the detailed characterization of individual systems that often reveal salient clues about the inaccuracies of assumptions underlying larger scale studies. Benchmark systems, whether they are clusters, eclipsing binaries, or well-characterized planets, offer the best path forward in refining our understanding of the evolution of stars and planets.</p
Reducing Latencies in Earthquake Early Warning
Get PDFExisting Earthquake Early Warning (EEW) algorithms use waveform analysis for earthquake detections, estimation of source parameters (i.e., magnitude and hypocenter location), and prediction of peak ground motions at sites near the source. The latency of warning delivery due to data collection significantly restricts the usefulness of the system, especially for users in the vicinity of the earthquake source, as the warning may not arrive before the strong shaking. This presentation discusses several methods to reduce the warning latency, while maintaining reliability and robustness, so that the warning time can be maximized for users to take appropriate actions to reduce causalities and economic losses.
Firstly, we incorporated the seismicity forecast information from Epidemic-Type Aftershock Sequence (ETAS) model into EEW as prior information, under the Bayesian probabilistic inference framework. Similar to human’s decision-making process, the Bayesian approach updates the probability of the estimations as more information becomes available. This allows us to reduce the required time for reliable earthquake signal detection from at least 3 seconds to 0.5 second. Furthermore, the initial error of hypocenter location estimation is reduced by 58%. The performance of the algorithm is further improved during aftershock sequences and swarm earthquakes.
Secondly, we introduce the use of multidimensional (KD tree) data structure to organize seismic database, so that the querying time can be reduced for the nearest neighbor search during earthquake source parameter estimation. The processing time of KD tree is approximately 15% of the processing time of linear exhaustive search, which allows the potential use of large seismic databases in real-time.
EEW is an interdisciplinary subject that involves collaboration among different scientific and engineering communities. Only by optimizing the warning time, such a unified system could be successful in taking protective actions before, during, and after earthquake natural disasters.</p
Computation and Comparison of Value Signals in Simple Perceptual and Economic Choices
Get PDFHow do we choose between different foods from a restaurant menu, or between a vacation overseas and more money in our savings account? Certain mechanisms in our brains allow us to make these and many other kinds of decisions effectively and efficiently. In this dissertation, I describe three projects which aim to advance our understanding of the systems and algorithms involved in the process of human decision making.
Chapter 2 investigates the application of the attentional Drift-Diffusion Model to a perceptual decision making task. Perceptual decisions requiring the comparison of spatially distributed stimuli that are fixated sequentially might be influenced by fluctuations in visual attention. We used two psychophysical tasks with human subjects to investigate the extent to which visual attention influences simple percep- tual choices, and to test the extent to which the attentional Drift-Diffusion Model provides a good computational description of how attention affects the underlying decision processes. We found that this model provides a reasonable quantitative description of the relationship between fluctuations in visual attention, choices, and response times. We also found evidence for the sizable attentional choice biases predicted by the model, and that exogenous manipulations of attention induce choice biases consistent with these predictions.
Chapter 3 compares two methods for fitting the parameters of the Drift-Diffusion Model using experimental data. A large number of studies have proposed that sequential integrator models of decision making, such as the Drift-Diffusion Model and its variants, provide a simple computational description of the algorithms used to make a large number of simple decisions. This is based on the fact that this class of models has been able to produce reasonably accurate descriptions of how choices, response times, and fixations are related to each other and to exogenous trial parameters, in a wide range of tasks. A difficult step in those studies is the estimation of a small number of free parameters to find the ones that explain the observed data best. The estimation method used in most studies is computationally very expensive since it approximates the likelihood of the observed data by simulating the model thousands of times and then counting the frequency with which the outcomes match the observed data. This problem is exacerbated with more complex models, such as the attentional Drift-Diffusion Model, or models with collapsing bounds, which contain a larger number of free parameters. We propose an alternative method for estimating the free parameters which relies on computing only the probability of the actual observed data, bypassing the need for the additional simulations. We present the results of simulation tests which show that our approach provides two key advantages over the alternative widely used method: a smaller number of experimental trials is needed in order to obtain comparable estimation accuracy, and the execution time of the estimation algorithms is substantially reduced.
Finally, Chapter 4 studies simple economic choices involving two distinct classes of valuation systems: an experiential system, which assigns value based on the history of previous reward experiences with similar options, and a descriptive system, which computes values using information about the options and environment available at the time of decision. Although these two systems often assign similar relative desirability to the different options, they do not always do so. When conflict arises with the experiential system favoring one option and the descriptive system favoring another, the brain needs to resolve the conflict to select a single option. We present the results of a psychometric study designed to characterize the basic interactions of these two valuation systems, with and without conflict.</p
Windowed Integral Equation Methods for Problems of Scattering by Defects and Obstacles in Layered Media
Get PDFThis thesis concerns development of efficient high-order boundary integral equation methods for the numerical solution of problems of acoustic and electromagnetic scattering in the presence of planar layered media in two and three spatial dimensions. The interest in such problems arises from application areas that benefit from accurate numerical modeling of the layered media scattering phenomena, such as electronics, near-field optics, plasmonics and photonics as well as communications, radar and remote sensing.
A number of efficient algorithms applicable to various problems in these areas are pre- sented in this thesis, including (i) A Sommerfeld integral based high-order integral equation method for problems of scattering by defects in presence of infinite ground and other layered media, (ii) Studies of resonances and near resonances and their impact on the absorptive properties of rough surfaces, and (iii) A novel Window Green Function Method (WGF) for problems of scattering by obstacles and defects in the presence of layered media. The WGF approach makes it possible to completely avoid use of expensive Sommerfeld integrals that are typically utilized in layer-media simulations. In fact, the methods and studies referred in points (i) and (ii) above motivated the development of the markedly more efficient WGF alternative.</p