Turn-level user satisfaction is one of the most important performance metrics for conversational agents. It can be used to monitor the agent's performance and provide insights about defective user experiences. While end-to-end deep learning has shown
promising results, having access to a large number of reliable annotated samples required by these methods remains challenging. In a large-scale conversational system, there is a growing number of newly developed skills, making the traditional data collection, annotation, and modeling process impractical due to the required annotation costs and the turnaround times. In this paper, we suggest a self-supervised contrastive learning approach that leverages the pool of unlabeled data to learn user-agent interactions. We show that the pre-trained models using the self-supervised objective are transferable to the user satisfaction prediction. In addition, we propose a novel few-shot transfer learning approach that ensures better transferability for very small sample sizes. The suggested few-shot method does not require any inner loop optimization process and is scalable to very large datasets and complex models. Based on our experiments using real data from a large-scale commercial system, the suggested approach is able to significantly reduce the required number of annotations, while improving the generalization on unseen skills.
Most of the villages and rural towns, in general, and in Latakia, in particular, suffer from
poor transport organizing and Serving of people so that travelling has become a great
suffering for most people. After studying the transport situation bet
ween Jableh and
Latakia, we found that this city suffers from transport problems like other cities. This is
because of transport poor management and regulation including timing, the number of
vehicles available and inability to meet demand, which causes a deterioration in service
quality.
The research provided a comprehensive method to evaluate the performance of public
transport on the road by evaluating travel time, the journey speed, and the buses volume
and frequency. Considering these indicators as a criterion for evaluating the performance
of public transport to determine the most effective factors regarding the quality of
performance of public transport and improve this performance in future. A field and
practical study of the research has been conducted, which includes selecting of study sites,
collecting engineering and traffic data, designing of questionnaires and distributing them
on random passengers and drivers working on Jableh - Lattakia transportation road to find
out their opinions. Then the data has been inserted into SPSS Statistics Base program for
analyzing the results.
The research has concluded that the performance of public transport between Jableh and
Latakia is generally weak, and the largest proportion of passengers depend on their
mobility on minibuses, which does not satisfy the required transportation level. Therefore,
a range of solutions has been developed depending on local conditions which will ease the
burden of mobility and raise the transportation performance level.