136879 research outputs found
Sort by
Cuban Economic Policy: A Quantitative Comparison to Vietnam and the Dominican Republic
Get PDFCuba began a shift towards a more market-oriented economy in 2011 with the announcement of new economic policies allowing for less state control in certain areas. Using a regression analysis, this study explores Cuba���s potential economic growth due to greater economic freedom by noting relationships between current economic factors and GDP, comparing those regression results to those in the Dominican Republic and Vietnam, and then drawing conclusions about short and long-term economic growth prospects. The analysis shows that the Dominican Republic and Cuba are economically similar, while Vietnam���s economy experiences different impacts on GDP from exports, tourism, and remittances compared to Cuba and the DR. If Cuba efficiently implements market-oriented policies to create a favorable environment for innovation, it could realize long-term growth
Dialouge Based Evaluation of Conversational Taskbots
Get PDFThe increasing popularity of intelligent systems has led to the rise of new types of conversational agents. Some of these systems are classified as task-oriented dialogue systems, or Taskbots. Previously, these systems focused primarily on completing digital tasks such as booking a hotel, or technical support. With the Alexa Challenge, University teams across the world have the chance to design conversational Taskbots which assist users with physical tasks such as cooking or DIY tasks. We use data from conversations between our bot and real users to determine the factors which determine the quality of a conversation with a Taskbot. We then use these factors to construct models which classify conversations based on predicted quality. We find that several factors are correlated with conversation quality, and that those factors can be used to predict the quality of a conversation. These results give insight into what factors users value in a conversation and reveal that conversation quality is not entirely determined by task completion. The predictive models developed in this work can help correct negative conversations as they happen and provide a baseline for the evaluation of future Taskbots
