【摘要】：對(duì)話管理(DM:Dialogue Management)在人機(jī)對(duì)話系統(tǒng)(DS:Dialogue System)中扮演著重要角色�；隈R氏決策過(guò)程(MDP:Markov Decision Process)的對(duì)話管理建模取得了不少進(jìn)展,但也存在一些問(wèn)題。其中之一是維度災(zāi)難(Curse of Dimensionality),它導(dǎo)致該模型不能應(yīng)用在復(fù)雜的、交換信息相對(duì)龐大的對(duì)話系統(tǒng)中。本文在前人的研究基礎(chǔ)上提出了一種層次MDP(Tier-MDP)對(duì)話管理模型。該模型將對(duì)話管理任務(wù)分為兩層。底層處理一個(gè)或多個(gè)相互獨(dú)立的對(duì)話子任務(wù),每個(gè)子任務(wù)接收對(duì)話理解的輸入,輸出對(duì)話行為,傳遞給上一層。上層將底層輸出的動(dòng)作做一個(gè)函數(shù)轉(zhuǎn)變作為狀態(tài),基于此進(jìn)行決策,獲得最后的對(duì)話行為。與MDP模型相比,Tier-MDP模型通過(guò)分層方式可以有效降低任務(wù)的狀態(tài)空間的規(guī)模,同時(shí)模型復(fù)雜度比分層強(qiáng)化學(xué)習(xí)(HRL:Hierarchical Reinforcement Learning)低。論文進(jìn)一步設(shè)計(jì)實(shí)現(xiàn)了 Tier-MDP對(duì)話管理模型的求解算法。論文實(shí)現(xiàn)了一個(gè)基于Tier-MDP對(duì)話管理模型的人機(jī)對(duì)話系統(tǒng),系統(tǒng)可以執(zhí)行基于人機(jī)對(duì)話的在線會(huì)議室預(yù)定任務(wù),系統(tǒng)具有較好的性能和交互能力,表明了基于Tier-MDP的對(duì)話管理系統(tǒng)具有較好的性能。
[Abstract]:DIALOG MANAGEMENT plays an important role in the man-machine dialog system (DS: Dialogue System). There are some problems in the process of dialogue management based on the Markov Decision Process (MDP), but there are some problems. One of these is the Curse of Dimensionality, which leads to the fact that the model cannot be applied in a complex, exchange-information-relatively large dialog system. In this paper, a hierarchical MDP (Tier-MDP) dialog management model is presented on the basis of previous studies. The model divides the session management task into two layers. The bottom layer processes one or more independent dialog sub-tasks, each sub-task receiving an input of a dialogue understanding, outputting a dialog behavior, and transmitting to the previous layer. The upper layer converts the operation of the bottom layer output into a function transition as a state, and makes a decision based on the function, and obtains the final conversation behavior. Compared with the MDP model, the Tier-MDP model can effectively reduce the scale of the state space of the task by the layering mode, and meanwhile, the model complexity level enhancement learning (HRL: Hierarchy Remedial Learning) is low. The paper further designs the algorithm for solving the Tier-MDP dialog management model. The paper realizes a man-machine conversation system based on the Tier-MDP dialog management model. The system can carry out the online meeting room reservation task based on man-machine conversation, and the system has better performance and interactive ability, which shows that the conversation management system based on the Tier-MDP has good performance.
【學(xué)位授予單位】：北京郵電大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2016
【分類號(hào)】：TP311.52

【參考文獻(xiàn)】

相關(guān)期刊論文 前4條

1 張文彬;;微軟小冰進(jìn)化第三代 擁有圖像識(shí)別能力[J];計(jì)算機(jī)與網(wǎng)絡(luò);2015年15期

2 胡寶潔;趙忠文;曾巒;張永繼;;圖靈機(jī)和圖靈測(cè)試[J];電腦知識(shí)與技術(shù);2006年23期

3 沈晶;顧國(guó)昌;劉海波;;分層強(qiáng)化學(xué)習(xí)中的Option自動(dòng)生成算法[J];計(jì)算機(jī)工程與應(yīng)用;2005年34期

4 王菁華,鐘義信,王樅,劉建毅;口語(yǔ)對(duì)話管理綜述[J];計(jì)算機(jī)應(yīng)用研究;2005年10期

相關(guān)碩士學(xué)位論文 前1條

1 李立云;基于Option自動(dòng)生成的分層強(qiáng)化學(xué)習(xí)方法研究[D];長(zhǎng)沙理工大學(xué);2008年

，

本文編號(hào)：2498808