Sequence-to-sequence neural generation models have achieved promising performance on short text conversation tasks. However, they tend to generate generic/dull responses, leading to unsatisfying dialogue experience. We observe that in conversation tasks, each query could have multiple responses, which forms a 1-ton or m-ton relationship in the view of the total corpus. The objective function used in standard sequence-to-sequence models will be dominated by loss terms with generic patterns. Inspired by this observation, we introduce a statistical re-weighting method that assigns different weights for the multiple responses of the same query, and trains the standard neural generation model with the weights. Experimental results on a large Chinese dialogue corpus show that our method improves the acceptance rate of generated responses compared with several baseline models and significantly reduces the number of generated generic responses.
The proposed approach can learn transparent behavior models represented as Behavior Trees, which could be used to alleviate the heaven endeavor of manual agent programming in game and simulation.
In the context of the rapid growth in electricity demand and the expanding of the power grid, grid dispatching and maintenance is becoming increasingly complex. To improve the dispatching and maintenance scheduling in distribution network, the “Intelligent Dispatching and Maintenance Arrangement System for 110kV Power Grid” was developed. This paper describes the design framework and functions of the system, and also briefly introduces the visual development tools-TWaver. Finally, the implementation of visualizing the real-time dynamic display of network power flow as well as the optimization result of maintenance schedule is given and realized a visual and intelligent mode of dispatching and maintenance scheduling effectively. The system has been put into operation in a grid with a stable and reliable performance, and shows the practicality and effectiveness of the system.
Abstract.Wireless communication technologies such as public wireless network, power private wireless network, wireless mesh network and LPWAN (Low Power Wide Area Network) are widely used in power grid communication in recent years, but the application principle of these technologies are still not clear when they are used in the communication network for distribution automation. On the base of the contrastive analysis of the technical characters of different wireless communication technologies, this paper gives an application suggestion of wireless communication technologies in distribution automation, which can guide the planning and construction of the communication network for distribution automation.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.