Set Of Learning Tasks Research Articles

Single loop or double loop learning: English vocabulary learning performance and behavior of students in situated computer games with different guiding strategies

In this study, a situated computer game was implemented for engaging students in English vocabulary contexts. Furthermore, two test-item guiding approaches, the cloze and multiple-choice guiding strategies, were embedded in the game to serve as guidance for the students. To investigate the students’ performance and behaviors of learning English vocabulary with different guiding strategies, two classes of students were included to learn with the two approaches. The experimental results showed that the students using the game with the cloze guiding strategy had significantly better learning achievement with higher cognitive load than those learning with the multiple-choice guiding strategy. Moreover, from the behavioral pattern sequential analysis, it was found that the game with cloze item guidance engaged the students in both single loop situated learning (i.e., repeatedly trying to deal with the same set of learning tasks) and double-loop situated learning (i.e., trying to deal with the same learning tasks after reviewing relevant materials and adjusting their learning strategies). On the other hand, those using the game with multiple-choice item guidance only performed single-loop situated learning during the gaming process, meaning that they seldom reviewed relevant materials or adjusted their learning strategies before trying to solve the same learning tasks again. The findings imply that it is worth considering different test-item guiding approaches when developing English vocabulary games for future studies and applications.

Reinforcement Learning and Wavelet Adapted Vortex Methods for Simulations of Self-propelled Swimmers

We present a numerical method for the simulation of collective hydrodynamics in self-propelled swimmers. Swimmers in a viscous incompressible flow are simulated with a remeshed vortex method coupled with Brinkman penalization and projection approach. The remeshed vortex methods are enhanced via wavelet based adaptivity in space and time. The method is validated on benchmark swimming problems. Furthermore the flow solver is integrated with a reinforcement learning algorithm, such that swimmers can learn to adapt their motion so as to optimally achieve a specified goal, such as fish schooling. The computational efficiency of the wavelet adapted remeshed vortex method is a key aspect for the effective coupling with learning algorithms. The suitability of this approach for the identification of swimming behaviors is assessed on a set of learning tasks.

Longitudinal attentional engagement rescues mice from age-related cognitive declines and cognitive inflexibility

Learning, attentional, and perseverative deficits are characteristic of cognitive aging. In this study, genetically diverse CD-1 mice underwent longitudinal training in a task asserted to tax working memory capacity and its dependence on selective attention. Beginning at 3 mo of age, animals were trained for 12 d to perform in a dual radial-arm maze task that required the mice to remember and operate on two sets of overlapping guidance (spatial) cues. As previously reported, this training resulted in an immediate (at 4 mo of age) improvement in the animals' aggregate performance across a battery of five learning tasks. Subsequently, these animals received an additional 3 d of working memory training at 3-wk intervals for 15 mo (totaling 66 training sessions), and at 18 mo of age were assessed on a selective attention task, a second set of learning tasks, and variations of those tasks that required the animals to modify the previously learned response. Both attentional and learning abilities (on passive avoidance, active avoidance, and reinforced alternation tasks) were impaired in aged animals that had not received working memory training. Likewise, these aged animals exhibited consistent deficits when required to modify a previously instantiated learned response (in reinforced alternation, active avoidance, and spatial water maze). In contrast, these attentional, learning, and perseverative deficits were attenuated in aged animals that had undergone lifelong working memory exercise. These results suggest that general impairments of learning, attention, and cognitive flexibility may be mitigated by a cognitive exercise regimen that requires chronic attentional engagement.

Facilitation of learning after lesions of the tuberomammillary nucleus region in adult and aged rats.

The tuberomammillary nucleus (TM) located in the posterior part of the hypothalamus is the main source of neuronal histamine in the central nervous system. Recent work from our laboratories has indicated an involvement of the TM region in neuronal plasticity and reinforcement processes. In the present study, we investigated the effects of TM lesions on the performance of adult and aged Wistar rats in a set of learning tasks, which differed in terms of complexity and reward contingencies (habituation learning, inhibitory avoidance, discrimination learning, Morris water maze). An improvement was found in every test applied, indicating that TM lesions seem to generally enhance learning and memory capacities independent of the special demands of a given task. Age-related learning deficits were strongly diminished. Immunohistochemistry revealed that the excitotoxic lesions used to destroy the TM region led to a marked decrease in the number of histamine-positive neurons in the vicinity of the injection site, indicating an involvement of the brain histaminergic system in the observed behavioral changes.

Tolerating concept and sampling shift in lazy learning using prediction error context switching

In their unmodified form, lazy-learning algorithms may have difficulty learning and tracking time-varying input/output function maps such as those that occur in concept shift. Extensions of these algorithms, such as Time-Windowed forgetting (TWF), can permit learning of time-varying mappings by deleting older exemplars, but have decreased classification accuracy when the input-space sampling distribution of the learning set is time-varying. Additionally, TWF suffers from lower asymptotic classification accuracy than equivalent non-forgetting algorithms when the input sampling distributions are stationary. Other shift-sensitive algorithms, such as Locally-Weighted forgetting (LWF) avoid the negative effects of time-varying sampling distributions, but still have lower asymptotic classification in non-varying cases. We introduce Prediction Error Context Switching (PECS) which allows lazy-learning algorithms to have good classification accuracy in conditions having a time-varying function mapping and input sampling distributions, while still maintaining their asymptotic classification accuracy in static tasks. PECS works by selecting and re-activating previously stored instances based on their most recent consistency record. The classification accuracy and active learning set sizes for the above algorithms are compared in a set of learning tasks that illustrate the differing time-varying conditions described above. The results show that the PECS algorithm has the best overall classification accuracy over these differing time-varying conditions, while still having asymptotic classification accuracy competitive with unmodified lazy-learners intended for static environments.

On the search for new learning rules for ANNs

In this paper, we present a framework where a learning rule can be optimized within a parametric learning rule space. We define what we callparametric learning rules and present a theoretical study of theirgeneralization properties when estimated from a set of learning tasks and tested over another set of tasks. We corroborate the results of this study with practical experiments.