While researchers have examined the effectiveness of various online gloss types on incidental L2 vocabulary learning, little research on online gloss languages has been conducted. Previous attempts which compared the effects of L1 and L2 glosses have reported mixed results. To fill the gaps, this study examined the effectiveness of Chinese and English e-glosses on incidental English vocabulary learning on a less-researched student group in CALL – junior high-school English-as-a-foreign-language (EFL) students. Seventy-eight students with Chinese as their first language read two online passages with either Chinese (L1) or English (L2) glosses. They were divided into four treatment groups: (1) high-proficiency students receiving L1 gloss before L2 gloss (n = 19), (2) high-proficiency students receiving L2 gloss before L1 gloss (n = 19), (3) low-proficiency students receiving L1 gloss before L2 gloss (n = 20), and (4) low-proficiency students receiving L2 gloss before L1 gloss (n = 20). After reading, the students were assessed with a vocabulary test which contained a definition-supply part and a cloze part serving as both post-tests and delayed post-tests. Repeated-measures analyses of variance were utilized to analyze the score data. Significant differences were found not only among the four groups but also between the two post-tests. Overall the high proficiency groups performed better in the post-tests, but the high proficiency group who received English glosses remembered more words in the delayed post-test than the high proficiency group who received Chinese glosses. The results show that as learners’ proficiency increases, they may be able to make better use of the L2 conceptual links for word retention and learning. The conclusions support the usefulness of both Chinese and English e-glosses which should be selected based on learners’ proficiency level.

This study presents novel robotic therapy control algorithms for upper-limb rehabilitation, using newly developed passive and progressive resistance therapy modes. A fuzzy-logic based proportional-integral-derivative (PID) position control strategy, integrating a patient's biomechanical feedback into the control loop, is proposed for passive movements. This allows the robot to smoothly stretch the impaired limb through increasingly rigorous training trajectories. A fuzzy adaptive impedance force controller is addressed in the progressive resistance muscle strength training and the adaptive resistive force is generated according to the impaired limb's muscle strength recovery level, characterized by the online estimated impaired limb's bio-damping and bio-stiffness. The proposed methods are verified with a custom constructed therapeutic robot system featuring a Barrett WAM™ compliant manipulator. Twenty-four recruited stroke subjects were randomly allocated in experimental and control groups and enrolled in a 20-week rehabilitation training program. Preliminary results show that the proposed therapy control strategies can not only improve the impaired limb's joint range of motion but also enhance its muscle strength.

We present a control method for a simple limit-cycle bipedal walker that uses adaptive frequency oscillators (AFOs) to generate stable gaits. Existence of stable limit cycles is demonstrated with an inverted-pendulum model. This model predicts a proportional relationship between hip torque amplitude and stride frequency. The closed-loop walking control incorporates adaptive Fourier analysis to generate a uniform oscillator phase. Gait solutions (fixed points) are predicted via linearization of the walker model, and employed as initial conditions to generate exact solutions via simulation. Global stability is determined via a recursive algorithm that generates the approximate basin of attraction of a fixed point. We also present an initial study on the implementation of AFO-based control on a bipedal walker with realistic mass distribution and articulated knee joints.

Learning probabilistic logic programming languages is receiving an increasing attention, and systems are available for learning the parameters (PRISM, LeProbLog, LFI-ProbLog and EMBLEM) or both structure and parameters (SEM-CP-logic and SLIPCASE) of these languages. In this paper we present the algorithm SLIPCOVER for “Structure LearnIng of Probabilistic logic programs by searChing OVER the clause space.” It performs a beam search in the space of probabilistic clauses and a greedy search in the space of theories using the log likelihood of the data as the guiding heuristics. To estimate the log likelihood, SLIPCOVER performs Expectation Maximization with EMBLEM. The algorithm has been tested on five real world datasets and compared with SLIPCASE, SEM-CP-logic, Aleph and two algorithms for learning Markov Logic Networks (Learning using Structural Motifs (LSM) and ALEPH++ExactL1). SLIPCOVER achieves higher areas under the precision-recall and receiver operating characteristic curves in most cases.

Mobile Ad hoc Networks have attracted much attention in the last years, since they allow the coordination and cooperation between agents belonging to a multi-robot system. However, initially deploying autonomously a wireless sensor robot network in a real environment has not taken the proper attention. Moreover, maintaining the connectivity between agents in real and complex environments is an arduous task since the strength of the connection between two nodes (i.e., robots) can change rapidly in time or even disappear. This paper compares two autonomous and realistic marsupial strategies for initial deployment in unknown scenarios, in the context of swarm exploration: Random and Extended Spiral of Theodorus. These are based on a hierarchical approach, in which exploring agents, named scouts, are autonomously deployed through explicit cooperation with supporting agents, denoted as rangers. Experimental results with a team of heterogeneous robots are conducted using both real and virtual robots. Results show the effectiveness of the methods, using a performance metric based on dispersion. Conclusions drawn in this work pave the way for a whole series of possible new approaches.

This paper presents findings from a study investigating young English language learners (YELLs) in Sweden in 4th grade (N = 76, aged 10–11). Data were collected with the help of a questionnaire and a one-week language diary. The main purpose was to examine the learners’ L2 English language-related activities outside of school in general, and their use of computers and engagement in playing digital games in particular. A comparison is made between language-related activities in English, Swedish, and other languages. Another purpose was to see whether there is a relationship between playing digital games and (a) gender, (b) L1, (c) motivation for learning English, (d) self-assessed English ability, and (e) self-reported strategies for speaking English. In order to do so, the sample was divided into three digital game groups, (1) non-gamers, (2) moderate, and (3) frequent gamers (≥4 hours/week), based on diary data (using self-reported times for playing digital games in English). Results showed that YELLs are extensively involved in extramural English (EE) activities (M = 7.2 hrs/w). There are statistically significant gender differences, boys (11.5 hrs/w) and girls (5.1 hrs/w; p < .01), the reason being boys’ greater time investment in digital gaming and watching films. The girls, on the other hand, spent significantly more time on pastime language-related activities in Swedish (11.5 hrs/w) than the boys (8.0 hrs/w; p < .05), the reason being girls’ greater time investment in facebooking. Investigation of the digital game groups revealed that group (1) was predominantly female, (2) a mix, and (3) predominantly male. YELLs with an L1 other than Swedish were overrepresented in group (3). Motivation and self-assessed English ability were high across all groups. Finally, regarding the self-reported strategies, code-switching to one's L1 was more commonly reported by non- and moderate gamers than frequent gamers.

In this paper, a new sensor-based approach called nonholonomic random replanner (NRR) is presented for motion planning of car-like mobile robots. The robot is incrementally directed toward its destination using a nonholonomic rapidly exploring random tree (RRT) algorithm. At each iteration, the robot's perceived map of the environment is updated using sensor readings and is used for local motion planning. If the goal was not visible to the robot, an approximate path toward the goal is calculated and the robot traces it to an extent within its sensor range. The robot updates its motion to goal through replanning. This procedure is repeated until the goal lies within the scope of the robot, after which it finds a more precise path by sampling in a tighter Goal Region for the nonholonomic RRT. Three main replanning strategies are proposed to decide when to perform a visibility scan and when to replan a new path. Those are named Basic, Deliberative and Greedy strategies, which yield different paths. The NRR was also modified for motion planning of Dubin's car-like robots. The proposed algorithm is probabilistically complete and its effectiveness and efficiency were tested by running several simulations and the resulting runtimes and path lengths were compared to the basic RRT method.

Natural Language Generation systems in interactive settings often face a multitude of choices, given that the communicative effect of each utterance they generate depends crucially on the interplay between its physical circumstances, addressee and interaction history. This is particularly true in interactive and situated settings. In this paper we present a novel approach for situated Natural Language Generation in dialogue that is based on hierarchical reinforcement learning and learns the best utterance for a context by optimisation through trial and error. The model is trained from human–human corpus data and learns particularly to balance the trade-off between efficiency and detail in giving instructions: the user needs to be given sufficient information to execute their task, but without exceeding their cognitive load. We present results from simulation and a task-based human evaluation study comparing two different versions of hierarchical reinforcement learning: One operates using a hierarchy of policies with a large state space and local knowledge, and the other additionally shares knowledge across generation subtasks to enhance performance. Results show that sharing knowledge across subtasks achieves better performance than learning in isolation, leading to smoother and more successful interactions that are better perceived by human users.

Answer Set Programming (ASP) is a popular framework for modelling combinatorial problems. However, ASP cannot be used easily for reasoning about uncertain information. Possibilistic ASP (PASP) is an extension of ASP that combines possibilistic logic and ASP. In PASP a weight is associated with each rule, whereas this weight is interpreted as the certainty with which the conclusion can be established when the body is known to hold. As such, it allows us to model and reason about uncertain information in an intuitive way. In this paper we present new semantics for PASP in which rules are interpreted as constraints on possibility distributions. Special models of these constraints are then identified as possibilistic answer sets. In addition, since ASP is a special case of PASP in which all the rules are entirely certain, we obtain a new characterization of ASP in terms of constraints on possibility distributions. This allows us to uncover a new form of disjunction, called weak disjunction, that has not been previously considered in the literature. In addition to introducing and motivating the semantics of weak disjunction, we also pinpoint its computational complexity. In particular, while the complexity of most reasoning tasks coincides with standard disjunctive ASP, we find that brave reasoning for programs with weak disjunctions is easier.

Commercial Users of Functional Programming (CUFP) is an annual workshop that is aimed at the community of software developers who use functional programming in real-world settings. This scribe report covers the talks that were delivered at the 2012 workshop, which was held in association with International Conference on Functional Programming (ICFP) in Copenhagen, Denmark. The goal of the report is to give the reader a sense of what went on, rather than to reproduce the full details of the talks. Videos and slides from all the talks are available online at http://cufp.org.

A measure of relative importance of network effects in the stochastic actor-oriented model (SAOM) is proposed. The SAOM is a parametric model for statistical inference in longitudinal social networks. The complexity of the model makes the interpretation of inferred results difficult. So far, the focus is on significance tests while the relative importance of effects is usually ignored. Indeed, there is no established measure to determine the relative importance of an effect in a SAOM. We introduce such a measure based on the influence effects have on decisions of individual actors in the network. We demonstrate its utility on empirical data by analyzing an evolving friendship network of university freshmen.

The human connectome has been widely studied over the past decade. A principal finding is that it can be decomposed into communities of densely interconnected brain regions. Past studies have often used single-scale modularity measures in order to infer the connectome's community structure, possibly overlooking interesting structure at other organizational scales. In this report, we used the partition stability framework, which defines communities in terms of a Markov process (random walk), to infer the connectome's multi-scale community structure. Comparing the community structure to observed resting-state functional connectivity revealed communities across a broad range of scales that were closely related to functional connectivity. This result suggests a mapping between communities in structural networks, models of influence-spreading and diffusion, and brain function. It further suggests that the spread of influence among brain regions may not be limited to a single characteristic scale.