Repetition-Based Approach for Task Adaptation in Imitation Learning.

Transfer learning is an effective approach for adapting an autonomous agent to a new target task by transferring knowledge learned from the previously learned source task. The major problem with traditional transfer learning is that it only focuses on optimizing learning performance on the target task. Thus, the performance on the target task may be improved in exchange for the deterioration of the source task's performance, resulting in an agent that is not able to revisit the earlier task. Therefore, transfer learning methods are still far from being comparable with the learning capability of humans, as humans can perform well on both source and new target tasks. In order to address this limitation, a task adaptation method for imitation learning is proposed in this paper. Being inspired by the idea of repetition learning in neuroscience, the proposed adaptation method enables the agent to repeatedly review the learned knowledge of the source task, while learning the new knowledge of the target task. This ensures that the learning performance on the target task is high, while the deterioration of the learning performance on the source task is small. A comprehensive evaluation over several simulated tasks with varying difficulty levels shows that the proposed method can provide high and consistent performance on both source and target tasks, outperforming existing transfer learning methods.

Domain Adaptation for Imitation Learning Using Generative Adversarial Network.

Imitation learning is an effective approach for an autonomous agent to learn control policies when an explicit reward function is unavailable, using demonstrations provided from an expert. However, standard imitation learning methods assume that the agents and the demonstrations provided by the expert are in the same domain configuration. Such an assumption has made the learned policies difficult to apply in another distinct domain. The problem is formalized as domain adaptive imitation learning, which is the process of learning how to perform a task optimally in a learner domain, given demonstrations of the task in a distinct expert domain. We address the problem by proposing a model based on Generative Adversarial Network. The model aims to learn both domain-shared and domain-specific features and utilizes it to find an optimal policy across domains. The experimental results show the effectiveness of our model in a number of tasks ranging from low to complex high-dimensional.

Adsorption and removal of arsenic from water by iron ore mining waste.

There is a global need to develop low-cost technologies to remove arsenic from water for individual household water supply. In this study, a purified and enriched waste material (treated magnetite waste, TMW) from the Trai Cau's iron ore mine in the Thai Nguyen Province in Vietnam was examined for its capacity to remove arsenic. The treatment system was packed with TMW that consisted of 75% of ferrous-ferric oxide (Fe(3)O(4)) and had a large surface area of 89.7 m(2)/g. The experiments were conducted at a filtration rate of 0.05 m/h to treat groundwater with an arsenic concentration of 380 microg/L and iron, manganese and phosphate concentrations of 2.07 mg/L, 0.093 mg/L and 1.6 mg/L respectively. The batch experimental results show that this new material was able to absorb up to 0.74 mg arsenic/g. The results also indicated that the treatment system removed more than 90% arsenic giving an effluent with an arsenic concentration of less than 30 microg/L while achieving a removal efficiency of about 80% for Mn(2 + ) and PO(4) (3-). This could be a promising and cost-effective new material for capturing arsenic as well as other metals from groundwater.

Arsenic removal by iron oxide coated sponge: treatment and waste management.

One of the problems in drinking water that raises concern over the world is that millions of people still have to use arsenic-contaminated water. There is a worldwide need to develop appropriate technologies to remove arsenic from water for household and community water supply systems. In this study, a new material namely iron oxide coated sponge (IOCSp) was developed and used to remove arsenic (As) from contaminated groundwater in Vietnam. The results indicated that IOCSp has a high capacity in removing both As (V) and As (III). The adsorption capacity of IOCSp was up to 4.6 mg As/g IOCSp, showing better than many other materials. It was observed from a pilot study that a small quantity of IOCSp (180 g) could reduce As concentration of 480 microg/L in 1.5 m3 of contaminated natural water to below 40 microg/L. In addition, an exhausted IOCSp, containing a large amount of arsenic (up to 0.42 wt %) could safely be disposed through the solidification/stabilization with cement. Addition of fly ash also reduced the amount of arsenic in the leachate.