A Detailed Earthquake Catalog for Banda Arc–Australian Plate Collision Zone Using Machine-Learning Phase Picker and an Automated Workflow

The tectonic setting of Timor–Leste and Eastern Indonesia comprises of a complex transition from oceanic lithosphere subduction to arc-continental collision. To better understand the deformation and convergent-zone structure of the region, we derive a new catalog of earthquake hypocenters and magnitudes from a temporary deployment of five years of continuous seismic data using an automated processing procedure. This includes a machine-learning phase picker, EQTransformer, and a sequential earthquake association and location workflow. We detect and locate ∼19,000 events during 2014–2018, which demonstrates that it is possible to characterize earthquake sequences from raw seismic data using a well-trained machine-learning picker for a complex convergent plate setting. This study provides the most complete catalog available for the region for the duration of the temporary deployment, which includes a complex pattern of crustal events across the collision zone and into the back-arc, as well as abundant deep slab seismicity.

Applying cognitive linguistics to second language idiom learning

<p>Idioms are known to cause great difficulty for second language (L2) learners, who may understand the literal meanings of the constituent words of idioms like (be) waiting in the wings, but often fail to interpret the idiomatic, figurative meaning of the expression. Proponents of Cognitive Linguistics (CL) claim that CL provides a pathway to more systematic and insightful learning of figurative expressions like idioms. They advocate that learners should be informed of the literal underpinning of idiomatic expressions and their relationship to the figurative meaning. This is supported by the results of several experimental studies employing ‘etymological elaboration’. However, little is known about how learners actually experience the CL-style explanations, or about how the learning is affected by other factors such as learners’ perceived transparency of the connection between the literal underpinnings and the idiomatic meanings, and their L1. The research reported in this thesis therefore (1) investigates the effectiveness of etymological elaboration in facilitating idiom comprehension and retention; (2) examines the problems that L2 learners, i.e., native-Chinese EFL learners in this study, experience when they encounter English figurative idioms, and identifies the factors influencing success in learning the meanings of idioms. To achieve these objectives, a mixed methods design was employed. Etymological elaboration was implemented in a teaching experiment involving one-on-one interviews, in which 25 Chinese learners of English were presented with idioms whose meaning they were asked to guess first without and then with the aid of information about their literal underpinnings. After the correct figurative meaning was established, participants rated the transparency of the connection between the literal underpinning and the figurative meaning. One week later, the learners were presented with the same idioms and asked to recall their meaning. Follow-up interviews were also conducted to survey the learners’ experience with and awareness of idioms, and their general attitudes and strategies towards idiom learning. Participants’ responses and their recall of idiomatic meanings were scored by three raters. A combination of quantitative and qualitative analyses of the interview data investigated the learning process and the outcomes of the teaching experiment. The major findings are: (1) Etymological elaboration can facilitate the interpretation and meaning retention of L2 idioms to a substantial degree; and the L2 idiom learning involves the interplay of multiple factors, including the transparency of the idioms, L1 transfer and cross-cultural differences, learners’ prior L2 lexical knowledge, and their proficiency levels. (2) The degree of transparency of the literal-figurative connection influences meaning retention, especially for the low proficiency learners. However, the mnemonic effect is not confined to idioms that learners find most transparent, but also affects those that are “far-fetched”. (3) The accuracy of meaning inference during the learning phase has a significant impact on memory for the idioms; many errors can be traced back to wrong guesses made in the prior learning phase, and some relate to false equivalents and partial equivalents in the L1. This suggests that trial-and-error learning potentially induces wrong memory traces and that teaching practices should therefore promote more accurate comprehension from the start, in order to facilitate better long-term memory for idioms. (4) More exposure to and better awareness of idioms help EFL learners foster positive attitudes towards idiom learning, which may facilitate the integration and automatization of figurative multiword expressions like idioms in their bilingual mental lexicon, and in turn lead to higher L2 proficiency. The findings of this study have implications for second language teaching and learning. The innovative research design and advanced statistical analyses contribute to the development of language teaching research methodology.</p>

Applying cognitive linguistics to second language idiom learning

<p>Idioms are known to cause great difficulty for second language (L2) learners, who may understand the literal meanings of the constituent words of idioms like (be) waiting in the wings, but often fail to interpret the idiomatic, figurative meaning of the expression. Proponents of Cognitive Linguistics (CL) claim that CL provides a pathway to more systematic and insightful learning of figurative expressions like idioms. They advocate that learners should be informed of the literal underpinning of idiomatic expressions and their relationship to the figurative meaning. This is supported by the results of several experimental studies employing ‘etymological elaboration’. However, little is known about how learners actually experience the CL-style explanations, or about how the learning is affected by other factors such as learners’ perceived transparency of the connection between the literal underpinnings and the idiomatic meanings, and their L1. The research reported in this thesis therefore (1) investigates the effectiveness of etymological elaboration in facilitating idiom comprehension and retention; (2) examines the problems that L2 learners, i.e., native-Chinese EFL learners in this study, experience when they encounter English figurative idioms, and identifies the factors influencing success in learning the meanings of idioms. To achieve these objectives, a mixed methods design was employed. Etymological elaboration was implemented in a teaching experiment involving one-on-one interviews, in which 25 Chinese learners of English were presented with idioms whose meaning they were asked to guess first without and then with the aid of information about their literal underpinnings. After the correct figurative meaning was established, participants rated the transparency of the connection between the literal underpinning and the figurative meaning. One week later, the learners were presented with the same idioms and asked to recall their meaning. Follow-up interviews were also conducted to survey the learners’ experience with and awareness of idioms, and their general attitudes and strategies towards idiom learning. Participants’ responses and their recall of idiomatic meanings were scored by three raters. A combination of quantitative and qualitative analyses of the interview data investigated the learning process and the outcomes of the teaching experiment. The major findings are: (1) Etymological elaboration can facilitate the interpretation and meaning retention of L2 idioms to a substantial degree; and the L2 idiom learning involves the interplay of multiple factors, including the transparency of the idioms, L1 transfer and cross-cultural differences, learners’ prior L2 lexical knowledge, and their proficiency levels. (2) The degree of transparency of the literal-figurative connection influences meaning retention, especially for the low proficiency learners. However, the mnemonic effect is not confined to idioms that learners find most transparent, but also affects those that are “far-fetched”. (3) The accuracy of meaning inference during the learning phase has a significant impact on memory for the idioms; many errors can be traced back to wrong guesses made in the prior learning phase, and some relate to false equivalents and partial equivalents in the L1. This suggests that trial-and-error learning potentially induces wrong memory traces and that teaching practices should therefore promote more accurate comprehension from the start, in order to facilitate better long-term memory for idioms. (4) More exposure to and better awareness of idioms help EFL learners foster positive attitudes towards idiom learning, which may facilitate the integration and automatization of figurative multiword expressions like idioms in their bilingual mental lexicon, and in turn lead to higher L2 proficiency. The findings of this study have implications for second language teaching and learning. The innovative research design and advanced statistical analyses contribute to the development of language teaching research methodology.</p>

Learning to diagnose X-rays: a neuroscientific study of practice-related activation changes in the prefrontal cortex

Objectives Medical expertise manifests itself by the ability of a physician to rapidly diagnose patients. How this expertise develops from a neural-activation perspective is not well understood. The objective of the present study was to investigate practice-related activation changes in the prefrontal cortex (PFC) as medical students learn to diagnose chest X-rays. Methods The experimental paradigm consisted of a learning and a test phase. During the learning phase, 26 medical students were trained to diagnose four out of eight chest X-rays. These four cases were presented repeatedly and corrective feedback was provided. During the test phase, all eight cases were presented together with near- and far-transfer cases to examine whether participants’ diagnostic learning went beyond simple rote recognition of the trained X-rays. During both phases, participants’ PFC was scanned using functional near-infrared spectroscopy. Response time and diagnostic accuracy were recorded as behavioural indicators. One-way repeated measures ANOVA were conducted to analyse the data. Results Results revealed that participants’ diagnostic accuracy significantly increased during the learning phase (F=6.72, p<0.01), whereas their response time significantly decreased (F=16.69, p<0.001). Learning to diagnose chest X-rays was associated with a significant decrease in PFC activity (F=33.21, p<0.001) in the left dorsolateral prefrontal cortex, the orbitofrontal area, the frontopolar area and the frontal eye field. Further, the results of the test phase indicated that participants’ diagnostic accuracy was significantly higher for the four trained cases, second highest for the near-transfer, third highest for the far-transfer cases and lowest for the untrained cases (F=167.20, p<0.001) and response time was lowest for the trained cases, second lowest for the near-transfer, third lowest for the far-transfer cases and highest for the untrained cases (F=9.72, p<0.001). In addition, PFC activity was lowest for the trained and near-transfer cases, followed by the far-transfer cases and highest for the untrained cases (F=282.38, p<0.001). Conclusions The results suggest that learning to diagnose X-rays is associated with a significant decrease in PFC activity. In terms of dual-process theory, these findings support the notion that students initially rely more on slow analytical system-2 reasoning. As expertise develops, system-2 reasoning transitions into faster and automatic system-1 reasoning.

When color coding backfires: A guidance reversal effect when learning with realistic visualizations

Digital learning increasingly makes use of realistic visualizations, although realism can be demanding for learners. Color coding is a popular way of helping learners understand visualizations and has been found to aid in learning with detailed visualizations. However, previous research has shown that color coding must not always be an effective aid, and that it even may reduce retention when used with simple visualizations. This study assessed whether the presence of color coding in learning tests has an effect after having learned using a detailed visualization that either featured color cues or one that did not. The results indicate that color coding helps learners the most if the learning tests also feature color coding. Importantly, learning with color-coded visualizations and being tested without color cues leads to the worst results in retention and transfer tests. Regarding transfer, color coding in the testing visualization boosts performance regardless of the presence of color cues in the learning phase. The results of this study challenge popular perspectives aiming at optimizing learning by removing potential sources of difficulty. Depending on the learning test, it may be more effective to keep a certain level of difficulty in the learning task when learning with digital media.

ANFIS soft sensing model of SMB chromatographic separation process based on new adaptive population evolution particle swarm optimization algorithm

Aiming at predicting the purity of the extract and raffinate components in the simulated moving bed (SMB) chromatographic separation process, a soft-sensor modeling method was proposed by adoptig the hybrid learning algorithm based on an improved particle swarm optimization (PSO) algorithm and the least means squares (LMS) method to optimize the adaptive neural fuzzy inference system (ANFIS) parameters. The hybrid learning algorithm includes a premise parameter learning phase and a conclusion parameter learning phase. In the premise parameter learning stage, the input data space division of the SMB chromatographic separation process and the initialization of the premise parameters are realized based on the fuzzy C-means (FCM) clustering algorithm. Then, the improved PSO algorithm is used to calculate the excitation intensity and normalized excitation intensity of all the rules for each individual in the population. In the conclusion parameter learning phase, these linear parameters are identified by the LMS method. In order to improve population diversity and convergence accuracy, the population evolution rate function was defined. According to the relationship between population diversity, population fitness function and particle position change, a new adaptive population evolution particle swarm optimization (NAPEPSO) algorithm was proposed. The inertia weight is adaptively adjusted according to the evolution of the population and the change of the particle position, thereby improving the diversity of the particle swarm and the ability of the algorithm to jump out of the local optimal solution. The simulation results show that the proposed soft-sensor model can effectively predict the key economic and technical indicators of the SMB chromatographic separation process so as to meet the real-time and efficient operation of the SMB chromatographic separation process.

Garner_et_al_when_learning_hurts_control

Performing two tasks concurrently typically leads to performance costs. Historically, multitasking costs have been assumed to reflect fundamental constraints of cognitive architectures. A new perspective proposes that multitasking costs reflect information sharing between constituent tasks; shared information gains representational efficiency, at the expense of multitasking capability. We test this theory by determining whether increasing cross-task information harms multitasking. 48 participants performed multitasks where they mapped keypresses to four shapes. In a subsequent statistical learning task, these shapes then formed pairs that were predictive or non-predictive of an upcoming target judgement. When participants again responded to these shapes in the multitasking context, performance was poorer when the shape pair had been predictive of target outcomes in the learning phase, relative to non-predictive. Thus, associating common information to shape pairings transferred to negatively impact multitasking performance, providing the first causal evidence for the shared representational account of multitasking performance.

Machine Learning for Species Identification: The HebelomaProject from database to website

Attempts to use machine learning (ML) for species identification of macrofungi have usually involved the use of image recognition to deduce the species from photographs, sometimes combining this with collection metadata. Our approach is different: we use a set of quantified morphological characters (for example, the average length of the spores) and locality (GPS coordinates). Using this data alone, the machine can learn to differentiate between species. Our case study is the genus Hebeloma, fungi within the order Agaricales, where species determination is renowned as a difficult problem. Whether it is as a result of recent speciation, the plasticity of the species, hybridization or stasis is a difficult question to answer. What is sure is that this has led to difficulties with species delimitation and consequently a controversial taxonomy. The Hebeloma Project—our attempt to solve this problem by rigorously understanding the genus—has been evolving for over 20 years. We began organizing collections in a database in 2003. The database now has over 10,000 collections, from around the world, with not only metadata but also morphological descriptions and photographs, both macroscopic and microscopic, as well as molecular data including at least an internal transcribed spacer (ITS) sequence (generally, but not universally, accepted as a DNA barcode marker for fungi (Schoch et al. 2012)), and in many cases sequences of several loci. Included within this set of collections are almost all type specimens worldwide. The collections on the database have been analysed and compared. The analysis uses both the morphological and molecular data as well as information about habitat and location. In this way, almost all collections are assigned to a species. This development has been enabled and assisted by citizen scientists from around the globe, collecting and recording information about their finds as well as preserving material. From this database, we have built a website, which updates as the database updates. The website (hebeloma.org) is currently undergoing beta testing prior to a public launch. It includes up-to-date species descriptions, which are generated by amalgamating the data from the collections of each species in the database. Additional tools allow the user to explore those species with similar habitat preferences, or those from a particular biogeographic area. The user is also able to compare a range of characters of different species via an interactive plotter. The ML-based species identifier is featured on the website. The standardised storage of the collection data on the database forms the backbone for the identifier. A portion of the collections on the database are (almost) randomly selected as a training set for the learning phase of the algorithm. The learning is “supervised” in the sense that collections in the training set have been pre-assigned to a species by expert analysis. With the learning phase complete, the remainder of the database collections may then be used for testing. To use the species identifier on the website, a user inputs the same small number of morphological characters used to train the tool and it promptly returns the most likely species represented, ranked in order of probability. As well as describing the neural network behind the species identifier tool, we will demonstrate it in action on the website, present the successful results it has had in testing to date and discuss its current limitations and possible generalizations.