Markov processes play an important role in reliability analysis and particularly in modeling the stochastic evolution of survival/failure behavior of systems. The probability law of Markov processes is described by its generator or the transition rate matrix. In this paper, we suppose that the process is doubly stochastic in the sense that the generator is also stochastic. In our model, we suppose that the entries in the generator change with respect to the changing states of yet another Markov process. This process represents the random environment that the stochastic model operates in. In fact, we have a Markov modulated Markov process which can be modeled as a bivariate Markov process that can be analyzed probabilistically using Markovian analysis. In this setting, however, we are interested in Bayesian inference on model parameters. We present a computationally tractable approach using Gibbs sampling and demonstrate it by numerical illustrations. We also discuss cases that involve complete and partial data sets on both processes.

The architecture of a system is decided at the initial stage of the design. However, the robustness of the system is not usually assessed in detail during the initial stages, and the exploration of alternative system architectures is limited due to the influence of previous designs and opinions. This article presents a novel network generator that enables the analysis of the robustness of alternative system architectures in the initial stages of design. The generator is proposed as a network tool for system architectures dictated by their configuration of source and sink components structured in a way to deliver a particular functionality. Its parameters allow exploration with theoretical patterns to define the main structure and hub structure, vary the number, size, and connectivity of hub components, define source and sink components and directionality at the hub level and adapt a redundancy threshold criterion. The methodology in this article assesses the system architecture patterns through robustness and modularity network based metrics and methods. Two naval distributed engineering system architectures are examined as the basis of reference for the simulated networks. The generator provides the capacity to create alternative complex system architecture options with identifiable patterns and key features, aiding in a broader explorative and analytical, in-depth, time and cost-efficient initial design process.

Subwords have become very popular, but the BERTa and ERNIEbtokenizers often produce surprising results. Byte pair encoding (BPE) trains a dictionary with a simple information theoretic criterion that sidesteps the need for special treatment of unknown words. BPE is more about training (populating a dictionary of word pieces) than inference (parsing an unknown word into word pieces). The parse at inference time can be ambiguous. Which parse should we use? For example, “electroneutral” can be parsed as electron-eu-tral or electro-neutral, and “bidirectional” can be parsed as bid-ire-ction-al and bi-directional. BERT and ERNIE tend to favor the parse with more word pieces. We propose minimizing the number of word pieces. To justify our proposal, a number of criteria will be considered: sound, meaning, etc. The prefix, bi-, has the desired vowel (unlike bid) and the desired meaning (bi is Latin for two, unlike bid, which is Germanic for offer).

The reachability semantics for Petri nets can be studied using open Petri nets. For us, an “open” Petri net is one with certain places designated as inputs and outputs via a cospan of sets. We can compose open Petri nets by gluing the outputs of one to the inputs of another. Open Petri nets can be treated as morphisms of a category Open(Petri), which becomes symmetric monoidal under disjoint union. However, since the composite of open Petri nets is defined only up to isomorphism, it is better to treat them as morphisms of a symmetric monoidal double category ${\mathbb O}$pen(Petri). We describe two forms of semantics for open Petri nets using symmetric monoidal double functors out of ${\mathbb O}$pen(Petri). The first, an operational semantics, gives for each open Petri net a category whose morphisms are the processes that this net can carry out. This is done in a compositional way, so that these categories can be computed on smaller subnets and then glued together. The second, a reachability semantics, simply says which markings of the outputs can be reached from a given marking of the inputs.

Creating collaborative working and learning experiences has long been at the forefront of computer-assisted language learning research. It is in this context that, in recent years, the integration of social networking sites and Web 2.0 in learning settings has surged, generating new opportunities to establish and explore virtual communities of practice (VCoPs). However, despite the number of studies on the concept, research remains inconclusive on how learners develop a sense of community in a VCoP, and what effect this may have on interaction and learning. This research project proposes to use social network analysis, part of graph theory, to explore the configuration of a set of VCoPs, and presents an empirical approach to determine how interaction in such communities takes shape. The present paper studies the concept of “community” in two VCoPs on Facebook. Participants (Group 1: N = 123, Group 2: N = 34) in both VCoPs are enrolled in English as a foreign language courses at two Belgian institutions of higher education. Social network analysis is used to show how both learner groups establish and develop a network of peers, and how different participants in those groups adopt different roles. Participation matrices reveal that interaction mainly revolves around a number of active key figures and that certain factors such as the incorporation of online and offline assignments and the inclusion of a teacher online result in varying levels of success when establishing collaborative dialogue within the VCoPs. Recommendations are formulated to inform and improve future practice.

The use of robots in performance arts is increasing. But, it is hard for robots to cope with unexpected circumstances during a performance, and it is almost impossible for robots to act fully autonomously in such situations. IROS-HAC is a new challenge in robotics research and a new opportunity for cross-disciplinary collaborative research. In this paper, we describe a practical method for generating different personalities of a robot entertainer. The personalities are created by selecting speech or gestures from a set of options. The selection uses roulette wheel selection to select answers that are more closely aligned with the desired personality. In particular, we focus on a robot magician, as a good magic show includes good interaction with the audience and it may also include other robots and performers. The magician with a variety of personalities increased the audience immersion and appreciation and maintained the audience’s interest. The magic show was awarded first prize in the competition for a comprehensive evaluation of technology, story, and performance. This paper contains both the research methodology and a critical evaluation of our research.

The recent breakthroughs in deep neural architectures across multiple machine learning fields have led to the widespread use of deep neural models. These learners are often applied as black-box models that ignore or insufficiently utilize a wealth of preexisting semantic information. In this study, we focus on the text classification task, investigating methods for augmenting the input to deep neural networks (DNNs) with semantic information. We extract semantics for the words in the preprocessed text from the WordNet semantic graph, in the form of weighted concept terms that form a semantic frequency vector. Concepts are selected via a variety of semantic disambiguation techniques, including a basic, a part-of-speech-based, and a semantic embedding projection method. Additionally, we consider a weight propagation mechanism that exploits semantic relationships in the concept graph and conveys a spreading activation component. We enrich word2vec embeddings with the resulting semantic vector through concatenation or replacement and apply the semantically augmented word embeddings on the classification task via a DNN. Experimental results over established datasets demonstrate that our approach of semantic augmentation in the input space boosts classification performance significantly, with concatenation offering the best performance. We also note additional interesting findings produced by our approach regarding the behavior of term frequency - inverse document frequency normalization on semantic vectors, along with the radical dimensionality reduction potential with negligible performance loss.

Ecodesign has gained significant traction in recent years ranging from academic research to business applications at a global scale. Initial emphasis on the environmental aspect of design has evolved to include economic and social aspects, with projects ranging from small-scale products to large-scale industrial systems. In this paper, the authors re-analyse 10 of their major ecodesign research projects of the past ten years to identify five categories of challenges and promising future directions for ecodesign research. This paper is primarily a retrospective position paper based on the authors’ experience of actual design studies, providing also a relevant literature review and summary of design practices.

Translations are generally assumed to share universal features that distinguish them from texts that are originally written in the same language. Thus, we can argue that these translations constitute their own variety of a language, often called translationese. However, translations are also influenced by their source languages and thus show different characteristics depending on the source language. Consequently, we argue that these variants constitute different “dialects” of translations into the same target language. Studies using machine learning techniques on Indo-European languages have investigated the universal characteristics of translationese and how translations from various source languages differ. However, for typologically very different languages such as Chinese, there are only few corpus studies that tap into the intricate relation between translations and the originals, as well as into the relations among translations themselves. In this contribution, we investigate the following questions: (1) What are the characteristics of Chinese translationese, both in general and with respect to different source languages? (2) Can we find differences not only at the lexical but also on the syntactic level? and (3) Based on the characteristics found in the previous questions, which of the proposed laws and universals can we corroborate based on our evidence from Chinese? We use machine learning to operationalize determining the importance of different characteristics and comparing their importance for our Chinese dataset with characteristics previously reported in studies on English. In addition, our methodology allows us to add syntactic features, which have rarely been used to study translations into Chinese. Our results show that Chinese translations as a whole can be reliably distinguished from non-translations, even based on only five features. More interestingly, typological traces from the source languages can often be found in their translations, therefore creating what we call dialects of translationese. For instance, translations from two Altaic languages exhibit more noun repetition and less frequent use of pronouns. Additionally, some characteristics that are not discriminative for English work well for Chinese, possibly because the distance between Chinese and the source languages is greater than that in English studies.

A person’s egonet, the set of others with whom that person is connected, is a personal sample of society which especially influences that person’s experience and perceptions of society. We show that egonets systematically misrepresent the general population because each person is included in as many egonets as that person has “friends.” Previous research has recognized that this unequal weighting in egonets leads many people to find that their friends have more friends than they themselves have. This paper builds upon that research to show that people’s egonets provide them with systematically biased samples of the population more generally. We discuss how this ubiquitous egonet bias may have far reaching implications for people’s experiences and perceptions of frequencies of other people’s ties and traits in ways that may influence their own feelings and behaviors. In particular, these egonet biases may help explain people’s tendencies to disproportionately experience and overestimate the prevalence of certain types of deviance and other social behaviors and consequently be influenced toward them. We illustrate egonet bias with analyses of all friends among 63,731 Facebook users. We call for further empirical investigation of egonet biases and their consequences for individuals and society.

Ego networks are thought to be influenced by the opportunities provided to associate with others given by our master statuses (e.g., race or sex), by the preferences individuals possess for interaction given our personality traits (e.g., extroverted or neurotic), and by the capacity to manage interactions on an ongoing basis given our cognitive ability to recall network information. However, prior research has been unable to examine all three classes of predictors concurrently. We rectify this deficiency in the literature by using a novel dataset of nearly 1000 respondents collected using controlled laboratory designs; using this dataset, we can simultaneously examine the impact of master statuses, personality traits, and social cognitive competencies on ego network size, structure (i.e., density), and composition (i.e., diversity). We find that all classes of predictors influence our ego networks, though in different ways, and point to new avenues for research into human sociability.

In recent times, satisfiability modulo theories (SMT) techniques gained increasing attention and obtained remarkable success in model-checking infinite-state systems. Still, we believe that whenever more expressivity is needed in order to specify the systems to be verified, more and more support is needed from mathematical logic and model theory. This is the case of the applications considered in this paper: we study verification over a general model of relational, data-aware processes, to assess (parameterized) safety properties irrespectively of the initial database (DB) instance. Toward this goal, we take inspiration from array-based systems and tackle safety algorithmically via backward reachability. To enable the adoption of this technique in our rich setting, we make use of the model-theoretic machinery of model completion, which surprisingly turns out to be an effective tool for verification of relational systems and represents the main original contribution of this paper. In this way, we pursue a twofold purpose. On the one hand, we isolate three notable classes for which backward reachability terminates, in turn witnessing decidability. Two of such classes relate our approach to conditions singled out in the literature, whereas the third one is genuinely novel. On the other hand, we are able to exploit SMT technology in implementations, building on the well-known MCMT (Model Checker Modulo Theories) model checker for array-based systems and extending it to make all our foundational results fully operational. All in all, the present contribution is deeply rooted in the long-standing tradition of the application of model theory in computer science. In particular, this paper applies these ideas in an original mathematical context and shows how these techniques can be used for the first time to empower algorithmic techniques for the verification of infinite-state systems based on arrays, so as to make such techniques applicable to the timely, challenging settings of data-aware processes.