Embedding Context Research Articles

Superior Disembedding in Children with ASD: New Tests Using Abstract, Meaningful, and 3D Contexts

Since its initial development, the embedded figures test (EFT) has been used extensively to measure local–global perceptual style. However, little is known about the perceptual factors that influence target detection. The current study aimed to investigate disembedding in children with and without ASD, aged 8–15 years, using the newly developed, stimulus-controlled L-EFT, M-EFT and D-EFT. Firstly, results revealed superior disembedding for children with ASD, irrespective of the type of target or embedding context, although the ASD group took more time in both the M-EFT and D-EFT. Secondly, the number of target lines continuing into the context proved more of a hindrance for the controls. Taken together, these findings provide strong evidence to support the notion of superior disembedding in ASD.

A Triple-Branch Neural Network for Knowledge Graph Embedding

The distributed representation of knowledge graphs (KGs), which embeds the structured graphs into low-dimensional embedding spaces, is widely used to facilitate various applications of AI, such as information retrieval and question answering. The primary elements of KGs, the entities viewed as nodes and the relations regarded as links between entities, naturally make up the local embedding context for each other, which is called the multi-restriction property of KGs. However, this property is not fully explored by previous models, where either only part of the multi-restriction is captured, or the capability of the embedding model is limited to a fixed function without clear interpretation. To address this issue, we propose TBNN, a triple-branch neural network to learn the embeddings of KGs. In particular, the embedding of any element of a KG is determined by its multi-restriction via an interaction layer followed by parallel branched layers. Thus, the entities and relations can be treated equivalently in spite of their seeming differences in the original KG. We define the loss function of TBNN based on the confidence score of the three elements of each triple. In addition, we propose using the log-sum-exp pairwise loss to smooth the hinge loss, which results in better performance. Empirically, we evaluate our model on the tasks of link prediction and triple classification with the subsets of WordNet and Freebase. Experiment results show that our model performs better than the baselines, especially providing stable performance for relations with different mapping properties.

Draft 2a: Collaboration and Conflict in Three Workplace Teams' Projects

This chapter extends previous research published in 2016 which looked into the embedding contexts of networks of small firms, principally in the EU , and the means by which collaboration between small to medium enterprises (SMEs) was supported inside national and regional clustering structures and incubators initiated in collaboration with university researchers and government agencies. The current chapter drills down further to explore the processes at the level of individual firms analysing and describing how group and individual conflict and/or collaboration was generated or sustained within teams in three different case organizations. In other words, the chapter looks at micro level details of conflict and collaboration as well as the observed socio-emotional dynamics observed. The three organizations were involved with executive education programs and the authors were able to access reflective diaries for 2004 to 2012 enabling the authors to triangulate notes taken with interview data and observations used for this chapter.

Local-global processing bias is not a unitary individual difference in visual processing

A large body of research reports individual differences in local and global visual processing in relation to expertise, culture and psychopathology. However, recent research has suggested that various different measures of local-global processing are not strongly associated with one another, calling its construct validity into question. The current study sought to further explore the validity of local-global processing biases in perception by developing three tasks based on two existing paradigms: the Embedded Figures Test (EFT) and the Navon hierarchical letters task. The newly developed tasks aimed to control for stimulus and response factors that may have impacted upon the reliability of previous research. They were administered to a large sample of undergraduate students (N>100). The results of two new versions of the EFT indicated that disembedding performance is influenced by the structure of the embedding context. In addition, global precedence and interference in the Navon task remained present even when local attentional approaches to global hierarchical stimuli were restricted. Inter-task correlations within the EFT were high but low between the EFT and the Navon task, lending support to the notion that local-global processing is not a monolithic construct, but representative of a number of distinct perceptual abilities and biases. Future research may use these task distinctions to pinpoint more precisely which aspects of perceptual processing characterise specific (clinical) participant populations.

Context-sensitive spatiotemporal simulation model for movement

GIScience 2016 Short Paper Proceedings Context-sensitive spatiotemporal simulation model for movement S. Dodge Department of Geography, Environment, and Society, University of Minnesota Email: sdodge@umn.edu Abstract This paper presents a context-sensitive spatiotemporal model to simulate movement trajectories. The model incorporates both the correlated random walk and time-geography theories to generate a more realistic trajectory of an agent within its environment. 1. Introduction Movement is an essential form of temporal change that is an integral characteristic of dynamic entities (e.g. humans, animals, vehicles, diseases). It is the focus of research in a range of application domains such as transportation, movement ecology, environmental studies, and human health. Movement models help us to better understand the characteristics of movement, enable us to simulate movement and predict its patterns (Dodge 2016). Examples of existing movement models include the random walk and its variations (Codling et al. 2008, Technitis et al. 2015), time-geography (Miller 2005, Song and Miller 2014), and Brownian Bridge (Horne et al. 2007) models. These models either generate trajectories using a set of geometric movement parameters (turn angle, distance), or they identify a visitation probability surface for an agent considering its speed and time budget. Existing models often disregard the characteristics of the environment or the context within which the movement takes place. Simulation of movement in relation to its embedding context is an essential problem that is applied to generate trajectories to fill gaps in low-resolution tracking datasets, or to examine behavioral responses of moving agents to environmental changes. This paper introduces a context-sensitive spatiotemporal simulation model based on a correlated random walk with external biases and is controlled by time-geography constraints of the moving agent. The novelty of the model is that at each step the simulation is driven by behavior and the contextual factors (i.e. environment, geography) that influence the local movement of the agent. As a case study, this research uses GPS observations of a tiger to parameterize the model and to simulate the tiger’s movement between actual GPS observations. 2. Movement Simulation The overall goal is to generate a trajectory (a sequence of spatiotemporal points) from a start location and time !(# $ , & $ , ' $ ) to an end location and time )(# * , & * , ' * ). The simulation uses a correlated random walk from ! with an external bias to move towards ) (i.e. global constraints). The local movement at each step is driven by agent’s behavior and contextual factors. The model specifications are: (1) the maximum movement speed is determined by behavior (e.g. patrolling, hunting, foraging, biking), (2) the global movement path and speed are controlled by the actual time-budget to reach the end-point, and (3) the path is influenced by agent’s local choices based on context (environmental drivers and spatial constraints, -e.g. general movement direction, slope preferences, trail network). The simulation algorithm runs on regular time intervals defined by the user, named step time , to ensure the global movement occurs within the time-budget ('+ = ' * − ' $ ). The maximum speed of the agent (/ 012 ) is determined based on expert knowledge or derived

Novel Tools for Product-service System Engineering

The objective of the research is to develop novel tools for engineering of Product Service Systems around products, which possess cyber physical features. The products with cyber physical features, such as vehicles, machines, home appliances, offer large spectrum of opportunities to provide new services around such products. The analysis of the state of the art indicate that engineering tools to support effective engineering of services around such products, are still missing. Two new engineering tools directly supporting development of the product service systems are developed a tool for selection /definition of sensors and intelligent features at the products for building services and tools for embedding context sensitivity into products and services. The research is driven by industrial application scenarios addressing different aspects of product-services systems as well as products development in different sectors (automotive, home appliances, automation equipment etc.). In this paper the focus is on an application scenario in automation equipment industry.

Grammatical constructions and cross-text generalizations

This paper investigates two tense-based constructions in English and Greek and one complementation construction in Greek, whose import is to effect a deictic shift and allow narration to proceed from the point of view of the narrated events and a participant therein. In addition to the individual formal and discourse-pragmatic properties of the patterns at hand, I focus on properties of the embedding context, showing that these unrelated constructions impose similar formal and interpretational requirements. This, in turn, supports the statement of generalizations at the level of genre, in this case empathetic narration as a special kind of narration that departs from the default past narrative which is deictically anchored to the narrator and the conversational coordinates. While the analysis adopts a bottom-up, language-driven approach to genre, it also refutes its exhaustive equation with linguistic conventions, arguing that a Bakhtinian view of genre, which includes both linguistic and socio-cultural dimensions, is more appropriate for the data at hand.

Similarity Measures for Chinese Short Text Based on Representation Learning

Similarity measure in Chinese short text is an important prerequisite for many content-based texts or documents retrieval tasks. In this paper, we propose a fast method for representing Chinese short texts to calculate the similarity between texts. The method is based on the representation of Chinese words. First, Chinese word representation is learned by a deep neural network with local context embedding and global context. Then, the words in short text are replaced by the learned representations of Chinese words and the short text is represented by dynamic average-weighted function depending on target text. Next, the cosine similarity method is used for the similarity measurement between texts. Last, experiment shows the semantic by visualizing the result of Chinese word representation learning and the experiment on similarity measure demonstrates the effectiveness of our short text representation method.

Adaptive multi-objective artificial immune system based virtual network embedding

In network virtualization, there are two decoupled roles involved: (i) infrastructure providers who manage the substrate network, and (ii) service providers who request virtual networks to the infrastructure providers. Embedding virtual networks to a shared substrate network, which is termed as virtual network embedding problem, is widely believed as one of the most significant challenges in such context. For this problem, prior work primarily focuses on either (i) maximizing the revenues by accommodating more virtual network requests or (ii) minimizing the energy consumption by consolidating the virtual networks into minimum number of substrate nodes. In this paper, we aim at achieving these two goals simultaneously. We first formulate the virtual network embedding problem into a multi-objective integer linear programming. We then design an artificial immune system based algorithm to solve this programming. In this algorithm, (i) we design a discrete approach to encode the virtual node mapping solution as an antibody; (ii) to initialize the antibodies, we design two adaptive revenue and energy aware strategies for the node and link mapping, respectively, to strike a balance between revenue and energy costs; (iii) we design corresponding customized strategies in the cloning, crossover and mutation process of artificial immune system in virtual network embedding context; (iv) for the generated antibodies, we leverage the Pareto optimality for evaluating their quality. Through extensive simulations, we show that our algorithm outperforms the state-of-the-art algorithms in terms of the revenue and the energy consumption.

Bringing Environmental Psychology Into Action

Environmental psychology is by definition one of the main disciplines for individual behavioral change with regard to sustainability issues. However, it has not been as visible or successful as one could wish. We present a four-step process that describes the passage of knowledge from science to policy: Basic research generates pieces of reliable knowledge; transdisciplinary research includes the views and knowledge of the stakeholders that are concerned, and tests results in the field; dissemination brings results, tools, and techniques to the attention of the public; finally, policy creates structures that provide guiding behavioral contexts. The various frames of understanding in the sectors of basic and transdisciplinary science, the media, and policy making are discussed, as well as ways of addressing them. Some attention is brought to the embedding contexts that foster the generation of compatible and useful scientific results. We conclude with a discussion of the relation between scientists and societal actors with a focus on trustworthiness.

A Method for Embedding Context to Sound-based Life Log

A Method for Embedding Context to Sound-based Life Log

Fluidity in Institutional Embedding of Proto-Institutions

Previous studies establish that rather than merely overcoming inertia and rigidity, institutional change also goes through stages of fluidity during which transformation seems to proceed rapidly without many contradictions, effective opposition, and failed attempts. Yet, we lack coherent understanding on what underlies such fluidity. We focus on proto-institutions, that is, institutions-in-the-making, and examine how aspects in their embedding enable fluid institutional change. Drawing on a longitudinal study of a project to establish and formalize the first foundation-based university in Finland, spearheading a wider university reform, we develop a framework for fluid institutional change. We find that collaboration among influential communities provides a context for effectual encoding and embedding of a proto-institution. Fluidity emerged from unexpected lack of resistance resulting from two processes. First, failed prior attempts for change, while constructing the issue and eroding institutions, simultaneously generated a perception that any further proto-institution will also fail, and thus resistance is unnecessary. Second, we find that the way in which proponents were able to encode and pace the proto-institution gives rise to fait accompli – a perception that a particular change is inevitable and resistance is futile. This represents initial objectification and internalization of the proto-institution among the key participating communities.

The 'consciousness' of the living state: Extending the Baars global broadcast model across physiological subsystems

AbstractFor the broad spectrum of cognitive biological phenomena having 'dual' information sources, isolation from crosstalk between them requires more metabolic free energy than permitting correlation. This allows an evolutionary exaptation leading to dynamic global broadcasts at multiple scales, similar to the well-studied exaptation of noise to trigger stochastic resonance amplification in physiological systems. Entropy gradient models adapted from nonequilibrium thermodynamics lead to an index theorem in which analytic solutions of empirical equations describe different possible topological modes. Not only is the living state characterized by cognition at every scale and level of organization, but by multiple, shifting, tunable, cooperative broadcasts that link selected subsets of those structures to address problems. As for animal consciousness, the 'stream' of cognitive physiological broadcast is constrained by the riverbanks of embedding context, expressed in terms of systematic regularities of an enveloping environmental information source and the limitations imposed by developmental trajectory.

A unified enhanced particle swarm optimization‐based virtual network embedding algorithm

SUMMARYVirtual network (VN) embedding is a major challenge in network virtualization. In this paper, we aim to increase the acceptance ratio of VNs and the revenue of infrastructure providers by optimizing VN embedding costs. We first establish two models for VN embedding: an integer linear programming model for a substrate network that does not support path splitting and a mixed integer programming model when path splitting is supported. Then we propose a unified enhanced particle swarm optimization‐based VN embedding algorithm, called VNE‐UEPSO, to solve these two models irrespective of the support for path splitting. In VNE‐UEPSO, the parameters and operations of the particles are well redefined according to the VN embedding context. To reduce the time complexity of the link mapping stage, we use shortest path algorithm for link mapping when path splitting is unsupported and propose greedy k‐shortest paths algorithm for the other case. Furthermore, a large to large and small to small preferred node mapping strategy is proposed to achieve better convergence and load balance of the substrate network. The simulation results show that our algorithm significantly outperforms previous approaches in terms of the VN acceptance ratio and long‐term average revenue. Copyright © 2012 John Wiley & Sons, Ltd.

Embedding effects in choice experiment valuations of environmental preservation projects

This paper addresses the question, whether attribute values derived from a CE study of one nature restoration and preservation project depend on the overall scale of nature preservation activities in which it is embedded. A split-sample CE study was undertaken in which a particular nature preservation project was evaluated in three plausible and strictly different embedding contexts. Respondents' attention was drawn to a varying number of forthcoming substitute preservation projects to be implemented prior to the one in question. Results show that while WTP for the project in focus is sensitive to the presentation of substitute projects as such, there does not seem to be any clear and unidirectional relationship between WTP and the number of substitute projects. Furthermore, effects vary across the project's attributes.

Chinese Radical Detection Also Depends on Linguistic Role

This study tests whether the linguistic role of a character in Chinese text affects detection of its constituent radicals. Linguistic role or status of an embedding context (e.g., word) has been shown to affect detection of embedded constituent units (e.g., letters) when reading alphabetic texts. Given significant differences between alphabetic and logographic systems such as English and Chinese, respectively, it is not clear whether the embedding linguistic role has a similar influence on constituent unit detection in Chinese. Findings here reveal that the linguistic role of the embedding context in Chinese contributes significantly to constituent detection. The outcome is consistent with the structural hypothesis, which posits that during reading, the prominence of linguistic units rapidly shifts from items contributing to structure to those that contribute meaning. Of note, the present study shows that this shifting importance of linguistic units applies to logographic as well as alphabetic systems.

Protein folding disorders: Toward a basic biological paradigm

Mechanistic ‘physics’ models of protein folding fail to account for the observed spectrum and rate of protein folding and aggregation disorders in human populations, showing that more appropriately in vivo paradigms reflecting biological and other embedding contexts are needed for understanding the etiology, prevention, and treatment of these diseases. Here, a topological rate distortion analysis is applied to the problem that is analogous to Tlusty (2007) elegant exploration of the genetic code. A ‘developmental’ perspective sees the rate distortion function as a temperature analog in a spontaneous symmetry breaking argument, and permits incorporation of external factors as catalysts, driving the system to different possible outcomes via a nonequilibrium empirical Onsager treatment, viewed as a kind of dynamic regression equation. The formalism produces large-scale, quasi-equilibrium ‘resilience’ states representing normal and pathological protein folding. Generalization to long times produces diffusion models of protein folding disorders in which epigenetic or life history factors determine the rate of onset of dysfunction.

Formalizing Homogeneous Language Embeddings

The cost of implementing syntactically distinct Domain Specific Languages (DSLs) can be reduced by homogeneously embedding them in a host language in cooperation with its compiler. Current homogeneous embedding approaches either restrict the embedding of multiple DSLs in order to provide safety guarantees, or allow multiple DSLs to be embedded but force the user to deal with the interoperability burden. In this paper we present the μ-calculus which allows parameterisable language embeddings to be specified and analysed. By reducing the problem to its core essentials we are able to show how multiple, expressive language embeddings can be defined in a homogeneous embedding context. We further show how variant calculi with safety guarantees can be defined.

Teaching & Learning Guide for: Once Upon a Time: Explaining Cultural Differences in Episodic Specificity

Teaching & Learning Guide for: Once Upon a Time: Explaining Cultural Differences in Episodic Specificity

Modality and Evidentiality

Linguistic expressions of modality and evidentiality have been the object of much active and exciting research for many years, but continue to pose challenges in all areas of semantic theory. Key aspects of their behaviour at the syntax–semantics interface, such as interactions with quantifiers and other operators or the calculation of modal implicatures triggered in embedding contexts, are not yet fully understood. Likewise, both the semantic interpretation and the pragmatic role of non-declarative sentences in modal constructions have only recently begun to attract significant attention, partly because the standard formal apparatus requires substantial augmentations in order to be applicable in that domain. But even one of the most fundamental and least controversial aspects of that formal apparatus, the use of possible worlds in the model theory, is subject to criticism and controversy. The articles in this collection address open issues in each of these three areas. The first paper, Yurie Hara's ‘Evidentiality of Discourse Items and Because-clauses’, addresses the problem of implicature computation involved in Japanese, focusing on contrastive wa-marking and sooda, a marker of ‘hearsay’ evidentiality. The use of contrastive wa-marking presupposes a contextually given stronger alternative, e.g. Mary and John came in (1b). (All glosses in this section are Hara's. ‘Con’ glosses contrastive wa.) When this presupposition is satisfied, the sentence conventionally implicates that the speaker does not know that the stronger alternative is true, or knows that it is false. ...