Net.Create: Network Visualization to Support Collaborative Historical Knowledge Building

Students across disciplines struggle with sensemaking when they are faced with the need to understand and analyze massive amounts of information. This is particularly salient in the disciplines of both history and data science. Our approach to helping students build expertise with complex information leverages activity theory to think about the design of a classroom activity system integrated with the design of a collaborative open-source network-analysis software tool called Net.Create. Through analysis of network log data as well as video data of students’ collaborative interactions with Net.Create, we explore how our activity system helped students reconcile common contradictions that create barriers to dealing with complex datasets in large lecture classrooms. Findings show that as students draw on details in a historical text to collaboratively construct a larger network, they begin to move more readily between small detail and aggregate overview. Students at both high and low initial skill levels were able to increase the complexity of their historical analyses through their engagement with the Net.Create tool and activities. Net.Create transforms the limitation of large class sizes in history classrooms into a resource for students’ collaborative knowledge building, and through collaborative data entry it supports the historiographic practices of citation and revision and helps students embed local historical actors into a larger historical context.

     

A revised methodology for the determination of bulk material cohesion and adhesion

The determination of the wall liner properties and more importantly the flow properties of a bulk material is critical for the design of any bulk materials handling system. The design of such materials handling systems will be most effective when handling bulk materials at the physical properties they were designed to handle. Due to the fast-paced nature of expansion in the mining industry and demand of mineral resources, it is quite common for materials handling systems to handle bulk materials that were not intended for the system. Wet and Sticky Materials (WSM) within the materials handling stream can cause significant downtime, due to events such as blockages of bins, hoppers and transfer chutes, remains left in train wagons and dump trucks as well as conveyor belt carry back (Roberts, 2005; Connelly, 2011 [2]).WSM are problematic within the materials handling stream due to the inter-particle and boundary cohesion and adhesion forces. The current measurement techniques for WSM have limitations and new methods must be considered. The development of new testers that can measure the wall adhesion and inter-particle adhesion of a bulk material can give a quantitative value for the adhesion present in a bulk material sample. The following paper will present a revised methodology for the estimation of the adhesion of bulk materials determined from the extrapolation of the Instantaneous Yield Locus (IYL). The predicted adhesion values from this methodology will be compared to experimental measurements using an inter-particle adhesion tester.     

Adaptive sequential machine learning

Abstract

A framework previously introduced in Wilson et al. (2018 Wilson, C., Veeravalli, V. V., and Nedi?, A. (2018). Adaptive Sequential Stochastic Optimization, IEEE Transactions on Automatic Control 64: 496–509. doi:10.1109/TAC.2018.2816168[Crossref], [Web of Science ®] , [Google Scholar]) for solving a sequence of stochastic optimization problems with bounded changes in the minimizers is extended and applied to machine learning problems such as regression and classification. The stochastic optimization problems arising in these machine learning problems are solved using algorithms such as stochastic gradient descent (SGD). A method based on estimates of the change in the minimizers and properties of the optimization algorithm is introduced for adaptively selecting the number of samples at each time step to ensure that the excess risk—that is, the expected gap between the loss achieved by the approximate minimizer produced by the optimization algorithm and the exact minimizer—does not exceed a target level. A bound is developed to show that the estimate of the change in the minimizers is non trivial provided that the excess risk is small enough. Extensions relevant to the machine learning setting are considered, including a cost-based approach to select the number of samples with a cost budget over a fixed horizon, and an approach to applying cross-validation for model selection. Finally, experiments with synthetic and real data are used to validate the algorithms.     

Lead-cooled system design and challenges in the frame of Generation IV International Forum

The Generation IV International Forum (GIF) Technology Roadmap identified the Lead-cooled Fast Reactor (LFR) as a technology well suited for electricity generation, hydrogen production and actinide management in a closed fuel cycle. One of the most important features of the LFR is the fact that lead is a relatively inert coolant, a feature that conveys significant advantages in terms of safety, system simplification, and the consequent potential for economic performance.In 2004, the GIF LFR Provisional System Steering Committee was organized and began to develop the LFR System Research Plan. The committee selected two pool-type reactor concepts as candidates for international cooperation and joint development in the GIF framework: these are the Small Secure Transportable Autonomous Reactor (SSTAR); and the European Lead-cooled System (ELSY).The high boiling point (1745 °C) of lead has a beneficial impact to the safety of the system, whereas its high melting point (327.4 °C) requires new engineering strategies, especially for In-Service-Inspection and refuelling. Lead, especially at high temperatures, is also relatively corrosive towards structural materials. This necessitates that coolant purity and the level of dissolved oxygen be carefully controlled, in addition to the proper selection of structural materials.For the GIF LFR concepts, lead has been chosen as the coolant rather than Lead-Bismuth Eutectic primarily because of its greatly reduced generation of the alpha-emitting ²¹⁰Po isotope formed in the coolant. This results in significantly reduced levels of radioactive contamination of the coolant while minimizing the effect of decay power in the coolant from such contaminants; an additional consideration is the desire to eliminate dependence on bismuth which might be a limited resource.This paper provides an overview of the historical development of the LFR, a summary of the advantages and challenges associated with heavy liquid metal coolants, and an update of the current status of development of LFR concepts under consideration. The main characteristics of the SSTAR and ELSY systems are summarized, and the current status of design of each system is presented. Because of the significant recent efforts in the ELSY system design, greater emphasis is placed on the ELSY plant, with focus on the technological development and design provisions intended to overcome or alleviate recognized drawbacks to the use of heavy liquid metal coolants. In the case of the SSTAR system for which development has proceeded more slowly, a more limited summary is provided. It is noted that both systems share many of the same research needs and objectives thus providing a strong basis for international collaboration.     

Rethinking random Hough Forests for video database indexing and pattern search

Computational Visual Media - Hough Forests have demonstrated effective performance in object detection tasks, which has potential to translate to exciting opportunities in pattern search. However,...     

The Simulacrum System as a Construct for Mass Spectrometry of Triacylglycerols and Others

A construct called a simulacrum is defined that provides all possible solutions to a sum of two mass spectral abundances, based on values (abundances) or ratios of those values. The defined construct is applied to atmospheric pressure chemical ionization mass spectrometry (MS) of triacylglycerols (TAG). A simulacrum has precisely defined components, specifically a simulacrum sum, four Possibilities to Observe, two Cases, and eight solutions. A simulacrum with no restrictions is the First General Form of a Simulacrum. When one value is specified to be 1 (as in MS), the construct is called a Unit Simulacrum, also called the First Specified Form of a Simulacrum. When one value is 1 and no value can be greater than 1 (the two specifications dictated by mass spectrometry), the construct is called the Second Specified Form of a Simulacrum, or the Mass Spectrometry Simulacrum. Simulacra are used with three Critical Ratios calculated from raw abundances in mass spectra of TAG to provide structural information about the degree of unsaturation in TAG, the identity and quantity of regioisomers, and other structural characteristics. Three‐level‐deep nested simulacrum solutions yield the recently reported Updated Bottom Up Solution, from which the protonated molecule, [MH]⁺, and all diacylglycerol‐like fragments, [DAG]⁺, of TAG can be reproduced from the Critical Ratios. Thus, the simulacrum solutions constitute a reduced data set in which more information is provided in fewer values than raw abundances, such that the Critical Ratios constitute a compact library of mass spectra.