Natural Language Generation in Dialog Systems

Recent advances in Automatic Speech Recognition technology have put the goal of naturally sounding dialog systems within reach. However, the improved speech recognition has brought to light a new problem: as dialog systems understand more of what the user tells them, they need to be more sophisticated at responding to the user. The issue of system response to users has been extensively studied by the natural language generation community, though rarely in the context of dialog systems. We show how research in generation can be adapted to dialog systems, and how the high cost of hand-crafting knowledge-based generation systems can be overcome by employing machine learning techniques. 1. DIALOG SYSTEMS AND GENERATION Recent advances in Automatic Speech Recognition (ASR) technology have put the goal of naturally sounding dialog systems within reach. However, the improved ASR has brought to light a new problem: as dialog systems understand more of what the user tells them, they need to be more sophisticated at responding to the user. If ASR is limited in quality, dialog systems typically employ a system-initiative dialog strategy in which the dialog system prompts the user for specific information and then presents some information to the user. In this paradigm, the range of user input at any time is limited (thus facilitating ASR), and the range of system output at any time is also limited. However, such interactions are not very natural. In a more natural interaction, the user can supply more and different information at any time in the dialog. The dialog system must then support a mixed-initiative dialog strategy. While this strategy places greater requirements on ASR, it also increases the range of system responses and the requirements on their quality in terms of informativeness and of adaptation to the context. For a long time, the issue of system response to users has been studied by the Natural Language Generation (NLG) community, though rarely in the context of dialog systems. What have emerged from this work are a “consensus architecture” [17] which modularizes the large number of tasks performed during NLG in a parThe work reported in this paper was partially funded by DARPA contract MDA972-99-3-0003. . ticular way, and a range of linguistic representations which can be used in accomplishing these tasks. Many systems have been built using NLG technology, including report generators [8, 7], system description generators [10], and systems that attempt to convince the user of a particular view through argumentation [20, 4]. In this paper, we claim that the work in NLG is relevant to dialog systems as well. We show how the results can be incorporated, and report on some initial work in adapting NLG approaches to dialog systems and their special needs. The dialog system we use is the AT&T Communicator travel planning system.We use machine learning and stochastic approaches where hand-crafting appears to be too complex an option, but we also use insight gained during previous work on NLG in order to develop models of what should be learned. In this respect, the work reported in this paper differs from other recent work on generation in the context of dialog systems [12, 16], which does not modularize the generation process and proposes a single stochastic model for the entire process. We start out by reviewing the generation architecture (Section 2). In Section 3, we discuss the issue of text planning for Communicator. In Section 4, we summarize some initial work in using machine learning for sentence planning [19]. Finally, in Section 5 we summarize work using stochastic tree models in generation [2]. 2. TEXT GENERATION ARCHITECTURE . NLG is conceptualized as a process leading from a high-level communicative goal to a sequence of communicative acts which accomplish this communicative goal. A communicative goal is a goal to affect the user’s cognitive state, e.g., his or her beliefs about the world, desires with respect to the world, or intentions about his or her actions in the world. Following (at least) [13], it has been customary to divide the generation process into three phases, the first two of which are planning phases. Reiter [17] calls this architecture a “consensus architecture” in NLG. During text planning, a high-level communicative goal is broken down into a structured representation of atomic communicative goals, i.e., goals that can be attained with a single communicative act (in language, by uttering a single clause). The atomic communicative goals may be linked by rhetorical relations which show how attaining the atomic goals contributes to attaining the high-level goal. During sentence planning, abstract linguistic resources are chosen to achieve the atomic communicative goals. This includes choosing meaning-bearing lexemes, and how the meaning-bearing lexemes are connected through abstract grammatical constructions (basically, lexical predicate-argument