To search for leverage is to use systemic design to find ways to accelerate progressive systemic change. The theory of leverage was first conceptualized by Donella Meadows with “Leverage Points: Places to Intervene in a System” in 1997. Yet while Meadows’s typology of leverage points is popular and influential, little has been done to critique or substantially advance her ideas since they were first published. As a result, we lack a modern theory of leverage. In this article, I relate systemic change to the search for leverage and outline why leverage matters. I present a brief overview of Meadows’s original work. Then, I synthesize the major contributions that have built on Meadows’s theory of leverage in the last 25 years. Next, I present a critique of Meadows’s original work, highlighting what we know about leverage and what we have yet to learn. This includes the development of a framework identifying how the degree of leverage relates to the acceleration of progressive (or retrograde) systemic change. Finally, I organize these ideas into a research agenda featuring four areas: dimensions of leverage, methods for leverage, strategy with leverage, and execution on leverage. Meadows wrote about the metaphor of “dancing with systems.” By advancing leverage theory, I believe we can better learn to “dance with systemic change.”

Extrapolation in graph classification/regression remains an underexplored area of an otherwise rapidly developing field. Our work contributes to a growing literature by providing the first systematic counterfactual modeling framework for extrapolations in graph classification/regression tasks. To show that extrapolation from a single training environment is possible, we develop a connection between certain extrapolation tasks on graph sizes and Lovasz's characterization of graph limits. For these extrapolations, standard graph neural networks (GNNs) will fail, while classifiers using induced homomorphism densities succeed, but mostly on unattributed graphs. Generalizing these density features through a GNN subgraph decomposition allows them to also succeed in more complex attributed graph extrapolation tasks. Finally, our experiments validate our theoretical results and showcase some shortcomings of common (interpolation) methods in the literature.

Data granularity is the level of direct correspondence between data in an Information System (IS) and the real-world things represented by the data. It determines the amount of detail that may be captured, stored, and used by contributors and consumers of information in an IS. We present a between-groups lab experiment in which we manipulated the granularity of a crowdsourcing project’s interface to assess the impact of granularity on data completeness, data correctness, and overall contributor participation. We found that contributors using a finer-grained data collection interface contributed more complete data, while contributors using a coarser-grained data collection interface contributed more incorrect data. Moreover, the level of granularity did not influence the degree of participation. These findings suggest that granularity is an important issue in the design of data crowdsourcing projects.

The objective of this project was to validate the efficacy of a miniature multispectral, single-sensor camera for detecting stress in deciduous juvenile tree foliage in a controlled environment. To that end, deciduous liners (one year old, nursery-grown transplants) representing five tree species (Celtis occidentalis L., Gleditsia triacanthos forma inermis Schneid, Gymnocladus dioicus (L.) K. Koch, Quercus rubra L., and Ulmus americana L.) were containerized and subsequently subjected to two treatments representing common landscape site stresses: drought and elevated soil pH. To minimize any atmospheric effects on reflectance values, the trees were placed in enclosed, artificially lit chambers during the detection process. The normalized difference vegetation index (NDVI) was utilized as the metric for analysis. Results showed the single imager sensor derived NDVI values were reliable indicators of tree stress within the species groups. Results also revealed that the derived values were not reliable indicators of tree stress across the species groups.

Micropropagation is a preferred method to propagate clean, clonal stock plants. Subculture is labor intensive and costly. In vitro hedging can reduce hood labor and was demonstrated with cannabis (Cannabis sativa L. ‘US Nursery Cherry 1’) using both apical and nodal explants at four different photosynthetic photon flux densities (25–167 μmol m−2 s−1) in vessels with vented or non-vented closures. The numbers of harvested shoot tips from four repeated 3-wk cutting cycles without subculture and the quality of the harvested shoot tips during ex vitro rooting in phenolic foam plugs were observed. The number of shoot tips harvested in non-vented vessels increased over four repeated cycles; however, there was a decline in number of shoot tips harvested from vented vessels in the fourth cycle due to excessive drying and collapse of the agar matric. The number of shoot tips harvested through repeated cycles was increased with light intensity. Nearly 100% of the micro-cuttings from optimal light and ventilation rooted ex vitro. Most plantlets had roots that penetrated the exterior surfaces of the plug after 2 wk. The number of leaves per rooted plantlet ex vitro increased with light intensity in vitro. Micropropagation labor efficiency could be improved by using a multi-cycle cutting process which would allow the same material to be repeatedly cut from rooted bases instead of subcultured. In a system with strong apical dominance, enhanced axillary divisions of shoot tip explants over successive cycles of cutting, without the need of exogenous PGRs, were further increased by appropriate in vitro factors, light, and ventilation, presenting labor saving potential compared to standard, single-cut system in common use.

Design and design management are increasingly called to respond to the world’s complex, dynamic problems. Yet, no standards or methodology exists to help designers understand, model, and design solutions for complex wicked problems. Program theory and social innovation promote the use of theory of change models to develop linear pathways of outcomes to show how a change initiative will have its desired effects. However, critics of these models accuse them of being simplistic and reductively linear. Systems thinking models use influence maps and causal loop diagrams to create maps of systems that show their behaviour in their full, dynamic complexity. However, these diagrams are sometimes complicated, overwhelming to read and therefore impractical. In this paper, we combine these tools with a novel technique from systemic design called “leverage analysis” to help identify crucial features of a complex problem and help designers develop practical theories of systemic change.

Crowdsourcing is a method of completing a task by engaging a large group of heterogeneous contributors. Data crowdsourcing is crowdsourcing of data collection. In this paper, we demonstrate how data crowdsourcing projects can be differentiated along five dimensions: (1) the extent to which tasks are well-defined; (2) the duration of the task; (3) the type of value generated by the consumers of crowdsourcing data; (4) the variety of contribution allowed when completing the task; and (5) the relative value of each contribution. We argue that the quality of information created by a crowd depends on the granularity of contributions contributors are able to make. Finally, we propose a set of principles for designing crowdsourcing system to align the level of granularity of contributions with project objectives.

Many systemic design processes include the development and analysis of systems models that represent the issue(s) at hand. In causal loop diagram models, phenomena are graphed as nodes, with connections between them indicating a control relationship. Such models provide mechanisms for stakeholder collaboration, problem finding and generative insight and are powerful . These functions are valued in design thinking, but the potential of these models may yet be unfulfilled. We introduce the notion of “leverage measures” to systemic design, adapting techniques from social network analysis and systems dynamics to uncover key structures, relationships and latent leverage positions of modelled phenomena. We demonstrate their utility in a pilot study. By rethinking the logics of leverage, we make better arguments for change and find the place from which to move the world.

Data crowdsourcing is the mobilization of large groups of contributors—often volunteers via the Internet— to collect and/or analyze data. Research on data crowdsourcing often prioritizes the data consumer or project sponsor. Significant gaps remain in understanding how to address design issues from the perspective of data crowdsourcing contributors. A systematic literature review is an ideal method for identifying gaps in how researchers conceptualize contributions in data crowdsourcing. This project presents a protocol for such a systematic literature review of data crowdsourcing. We will use the protocol to guide a subsequent systematic literature review and the construction of a data-information-knowledgewisdom chart that identifies critical gaps and opportunities for research in data crowdsourcing systems.

We consider a simple and overarching representation for permutation-invariant functions of sequences (or multiset functions). Our approach, which we call Janossy pooling, expresses a permutation-invariant function as the average of a permutation-sensitive function applied to all reorderings of the input sequence. This allows us to leverage the rich and mature literature on permutation-sensitive functions to construct novel and flexible permutation-invariant functions. If carried out naively, Janossy pooling can be computationally prohibitive. To allow computational tractability, we consider three kinds of approximations: canonical orderings of sequences, functions with k-order interactions, and stochastic optimization algorithms with random permutations. Our framework unifies a variety of existing work in the literature, and suggests possible modeling and algorithmic extensions. We explore a few in our experiments, which demonstrate improved performance over current state-of-the-art methods.

Coronary artery ectasia (CAE) is an uncommon pathology, which is sometimes incidentally found on left heart catheterization (LHC). CAE is occasionally treated with systemic anticoagulation to prevent thrombosis or progression of the clot in the coronary arteries. We present a 63-year-old male with known CAE on warfarin who presented to the hospital with myocardial infarction after a routine colonoscopy for which anticoagulation was held. His myocardial infarction was attributed to a likely coronary thromboembolic event. This case highlights the need for consideration of bridging anticoagulation therapy before and after procedures in patients with CAE to prevent adverse coronary events.

This work generalizes graph neural networks (GNNs) beyond those based on the Weisfeiler-Lehman (WL) algorithm, graph Laplacians, and diffusions. Our approach, denoted Relational Pooling (RP), draws from the theory of finite partial exchangeability to provide a framework with maximal representation power for graphs. RP can work with existing graph representation models and, somewhat counterintuitively, can make them even more powerful than the original WL isomorphism test. Additionally, RP allows architectures like Recurrent Neural Networks and Convolutional Neural Networks to be used in a theoretically sound approach for graph classification. We demonstrate improved performance of RP-based graph representations over state-of-the-art methods on a number of tasks.

In recent years, there has been a large push in the U.S. Department of Defense (DoD) to more rapidly respond and adapt to changing technology advancements and emerging communications systems threats. One issue has been that traditional DoD waveform development has been stove-piped in that processing blocks are implemented to serve a single function instead of made generic and configurable to support multiple waveforms. Developing new waveforms typically requires starting from scratch almost everytime. In this paper, we present Common Hardware-modem Integrated Library (CHIL). CHIL is both a library of configurable processing blocks as well as a framework that ties these blocks together to rapidly instantiate new DoD and non-DoD waveforms. We present an overview of the CHIL framework and various components as well as provide a test-case where CHIL was leveraged to instantiate a DoD waveform.

An abstract is not available for this content. As you have access to this content, full HTML content is provided on this page. A PDF of this content is also available in through the ‘Save PDF’ action button.

Integrated continuous simulation-optimization models can be effective predictors of a process-based responses for cost-benefit optimization of best management practices (BMPs) selection and placement. However, practical application of simulation-optimization model is computationally prohibitive for large-scale systems. This study proposes an enhanced Nonlinearity Interval Mapping Scheme (NIMS) to solve large-scale watershed simulation-optimization problems several orders of magnitude faster than other commonly used algorithms. An efficient interval response coefficient (IRC) derivation method was incorporated into the NIMS framework to overcome a computational bottleneck. The proposed algorithm was evaluated using a case study watershed in the Los Angeles County Flood Control District. Using a continuous simulation watershed/stream-transport model, Loading Simulation Program in C++ (LSPC), three nested in-stream compliance points (CP)each with multiple Total Maximum Daily Loads (TMDL) targetswere selected to derive optimal treatment levels for each of the 28 subwatersheds, so that the TMDL targets at all the CP were met with the lowest possible BMP implementation cost. Genetic Algorithm (GA) and NIMS were both applied and compared. The results showed that the NIMS took 11 iterations (about 11 min) to complete with the resulting optimal solution having a total cost of $67.2 million, while each of the multiple GA executions took 21-38 days to reach near optimal solutions. The best solution obtained among all the GA executions compared had a minimized cost of $67.7 millionmarginally higher, but approximately equal to that of the NIMS solution. The results highlight the utility for decision making in large-scale watershed simulation-optimization formulations.