Search

Abstract Reinforcement learning (RL) is an effective method in training dialogue policies to steer the conversation towards successful task completion. However, most RL-based methods only rely on semantic inputs that lack empathy as they ignore the user emotional information. Moreover, these methods suffer from delayed rewards caused by the user simulator returning valuable results only at dialogue end. Recently, some methods have been proposed to learn the reward function together with user emotions, but they omit considering user emotion in each dialogue turn. In this paper, we proposed an emotion-sensitive dialogue policy model (ESDP), it incorporates user emotions information into dialogue policy and selects the optimal action by the combination of top-k actions with the user emotions. The user emotion information in each turn is used as an immediate reward for the current dialogue state to solve sparse rewards and the dependency on termination. Extensive experiments validate that our method outperforms the baseline approaches when combined with different Q-Learning algorithms, and also surpasses other popular existing dialog policies’ performance.

Aluminum alloy LF 21 has a strong ability to reflect electromagnetic waves. LF 21 waveguide slit array structure is widely used in waveguide radar antenna. The stiffness of the slit array structure is relatively weak. So, the structure is prone to deformation under the cutting force in the conventional milling process. Micro-milling technology can realize high-precision machining of mesoscale parts/structures and is a potential effective machining technology for the waveguide slit array structure. However, the diameter of the micro-milling cutter is small, and the feed per tooth is comparable to the arc radius of the cutting edge, so the micro-milling cutter is prone to wear. In addition, the effects of elastic recovery of material, the minimum cutting thickness and friction of cutting dead zone on micro-milling force cannot be ignored. A simulation model of micro-milling aluminum alloy LF 21 processes based on DEFORM 3D is built by combining the theory of cutting and the technology of process simulation. Prediction of tool wear is achieved. The quantitative relationship between the arc radius of the cutting edge and tool wear is clarified for the first time. The authors built an improved cutting force model in micro-milling LF 21 considering tool wear and cutter runout with the minimum cutting thickness as the boundary. The validity of the built micro-milling force model is verified by experiments.