No full-text available
To read the full-text of this research,
you can request a copy directly from the author.
Architectural Issues for Compositional Dialog in Games
Abstract
Making a game that supports generative conversation with NPCs involves a number of technical and design challenges, ranging from designing new game mechanics to making a Prolog interpreter run under Unity. I discuss the software architecture of a game, scheduled for initial release in spring 2015, that attempts to solve some of these problems.
... Current authoring practice for videogame dialogue production, in which individuals or even teams of writers tirelessly produce huge amounts of content by hand, is largely seen as both untenable and constraining of the form (Mateas 2007). This issue hinders all games with dialogue, but it is most pronounced in the emerging genre that features freeform naturalistic conversation as core gameplay (Mateas and Stern 2003;Strong and Mateas 2008;Spierling 2011;Joseph 2012;Endrass et al. 2014;Brusk 2014;Evans and Short 2014;Horswill 2014;Lewis and Dill 2015;Treanor, McCoy, and Sullivan 2015;Mohov 2015;Lessard 2016), a format that has been shown to maximize player presence and engagement (Sali et al. 2010). Clearly, any alleviation of this content problem will rely on some amount of generativity, and so natural language generation (NLG) appears as a panacea. ...
... Clearly, any alleviation of this content problem will rely on some amount of generativity, and so natural language generation (NLG) appears as a panacea. But NLG has yet to find its killer application in games (Horswill 2014). Elsewhere, we have argued that traditional NLG-in which heavy interlocking systems are assembled by practitioners with deep expertise-may actually Copyright c 2016, Association for the Advancement of Artificial Intelligence (www.aaai.org). ...
... Bot Colony also employs a traditional NLG pipeline-a first for a commercially released title-particularly one in the style of service-based dialogue systems, made possible by the narrative conceit of NPCs in the game being service robots (Joseph 2012). Even more recently, Ian Horswill's ambitious MKULTRA generates character dialogue from a definite clause grammar specified in Prolog (Horswill 2014). Dunyazad likewise takes a grammar-based approach, in its case toward generating narrational text and choice prompts (Mawhorter 2016). ...
The Expressive Intelligence Studio is developing a new approach to freeform conversational interaction in playable media that combines dialogue management, natural language generation (NLG), and natural language understanding. In this paper, we present our method for dialogue generation, which has been fully implemented in a game we are developing called Talk of the Town. Eschewing a traditional NLG pipeline, we take up a novel approach that combines human language expertise with computer generativity. Specifically, this method utilizes a tool that we have developed for authoring context-free grammars (CFGs) whose productions come packaged with explicit metadata. Instead of terminally expanding top-level symbols — the conventional way of generating from a CFG — we employ an unusual middle-out procedure that targets mid-level symbols and traverses the grammar by both forward chaining and backward chaining, expanding symbols conditionally by testing against the current game state. In this paper, we present our method, discuss a series of associated authoring patterns, and situate our approach against the few earlier projects in this area.
... Current authoring practice for videogame dialogue production, in which individuals or even teams of writers tirelessly produce huge amounts of content by hand, is largely seen as both untenable and constraining of the form (Mateas 2007). This issue hinders all games with dialogue, but it is most pronounced in the emerging genre that features freeform naturalistic conversation as core gameplay (Mateas and Stern 2003;Strong and Mateas 2008;Spierling 2011;Joseph 2012;Endrass et al. 2014;Brusk 2014;Evans and Short 2014;Horswill 2014;Lewis and Dill 2015;Treanor, McCoy, and Sullivan 2015;Mohov 2015;Lessard 2016), a format that has been shown to maximize player presence and engagement (Sali et al. 2010). Clearly, any alleviation of this content problem will rely on some amount of generativity, and so natural language generation (NLG) appears as a panacea. ...
... Clearly, any alleviation of this content problem will rely on some amount of generativity, and so natural language generation (NLG) appears as a panacea. But NLG has yet to find its killer application in games (Horswill 2014). Elsewhere, we have argued that traditional NLG-in which heavy interlocking systems are assembled by practitioners with deep expertise-may actually Copyright c 2016, Association for the Advancement of Artificial Intelligence (www.aaai.org). ...
... Bot Colony also employs a traditional NLG pipeline-a first for a commercially released title-particularly one in the style of service-based dialogue systems, made possible by the narrative conceit of NPCs in the game being service robots (Joseph 2012). Even more recently, Ian Horswill's ambitious MKULTRA generates character dialogue from a definite clause grammar specified in Prolog (Horswill 2014). Dunyazad likewise takes a grammar-based approach, in its case toward generating narrational text and choice prompts (Mawhorter 2016). ...
The Expressive Intelligence Studio is developing a new approach to freeform conversational interaction in playable media that combines dialogue management, natural language generation (NLG), and natural language understanding. In this paper, we present our method for dialogue generation, which has been fully implemented in a game we are developing called Talk of the Town. Eschewing a traditional NLG pipeline, we take up a novel approach that combines human language expertise with computer generativity. Specifically, this method utilizes a tool that we have developed for author-ing context-free grammars (CFGs) whose productions come packaged with explicit metadata. Instead of terminally expanding top-level symbols—the conventional way of generating from a CFG—we employ an unusual middle-out procedure that targets mid-level symbols and traverses the grammar by both forward chaining and backward chaining, expanding symbols conditionally by testing against the current game state. In this paper, we present our method, discuss a series of associated authoring patterns, and situate our approach against the few earlier projects in this area.
... The mostly context-free formalism Tracery has spurred the creation of new communities of grammar authors (Compton, Kybartas, and Mateas 2015), and Expressionist and its tagged-grammars approach is being used in a variety of videogame projects (Ryan et al. 2016b;Lessard et al. 2017) and in the middleware technology of Spirit AI (Spirit AI 2017). Beyond applications of Tracery and Expressionist, others are employing grammar-based approaches in games and interactive storytelling, too (Horswill 2014;Mawhorter 2016;Dias 2016b;Togelius, Shaker, and Dormans 2016;Lewis 2017). As an alternative to conventional natural language generation pipelines-the dominant approach in mainstream text generation-grammars are appealing because they are easier to author (especially for writers who do not code) while still being very generative (due to an inherent combinatorial explosion). 1 The core appeal of Expressionist in particular, the approach we focus on in this paper, is that its tagging mechanism enables targeted generation (at runtime, content can be requested by specifying the meanings it should express) and content understanding (the meanings expressed by generated content are known, which means a larger system can understand and act on generated content). ...
... Beyond applications of Tracery and Expressionist, recent projects by Horswill, Mawhorter, and Dias, respectively, have also utilized generative grammars for expressive text generation (Horswill 2014;Mawhorter 2016;Dias 2016b). By employing generative grammars, these systems harness the power of templated dialogue, a pattern used in Prom ...
We extend the Expressionist project, and thereby the re-emerging area of grammar-based text generation, by applying a technique from software verification to a critical search problem related to content generation from grammars. In Expressionist, authors attach tags (corresponding to pertinent meanings) to nonterminal symbols in a context-free grammar, which enables the targeted generation of content that expresses requested meanings (i.e., has the requested tags). While previous work has demonstrated methods for requesting content with a single required tag, requests for multiple tags yields a search task over domains that may realistically span quintillions or more elements. In this paper, we reduce Expressionist grammars to symbolic visibly pushdown automata, which allows us to locate in massive search spaces generable outputs that satisfy moderately complex criteria related to tags. While the satisficing of more complex tag criteria is still not feasible using this technique, we forecast a number of opportunities for future directions.
... The obvious reason for this is that NLG is an extremely difficult task. There is a huge amount of domain knowledge that its practitioners must have, and beyond this there remain several major technical challenges in incorporating NLG into playable media (Horswill 2014). We contend, however, that another major reason that NLG has yet to be popularly or even successfully incorporated into these media is that the current state of the art is simply inadequate. ...
Natural language generation (NLG) has been featured in at most a handful of shipped games and interactive stories. This is certainly due to it being a very specialized practice, but another contributing factor is that the state of the art today, in terms of content quality, is simply inadequate. The major benefits of NLG are its alleviation of authorial burden and the capability it gives to a system of generating state-bespoke content, but we believe we can have these benefits without actually employing a full NLG pipeline. In this paper, we present the preliminary design of Expressionist, an in-development mixed-initiative authoring tool that instantiates an authoring scheme residing somewhere between conventional NLG and conventional human content authoring. In this scheme, a human author plays the part of an NLG module in that she starts from a set of deep representations constructed for the game or story domain and proceeds to specify dialogic content that may express those representations. Rather than authoring static dialogue, the author defines a probabilistic context-free grammar that yields templated dialogue. This allows a human author to still harness a computer's generativity, but in a capacity in which it can be trusted: operating over probabilities and treelike control structures. Additional features of Expressionist's design include arbitrary markup and realtime feedback showing currently valid derivations.
... Beyond applications of Tracery and Expressionist, recent projects by Horswill, Mawhorter, and Dias, respectively, have also utilized generative grammars for expressive text generation (Horswill 2014;Mawhorter 2016;Dias 2016b). By employing generative grammars, these systems harness the power of templated dialogue, a pattern used in Prom ...
We extend the Expressionist project, and thereby the re-emerging area of grammar-based text generation, by applying a technique from software verification to a critical search problem related to content generation from grammars. In Expressionist, authors attach tags (corresponding to pertinent meanings) to nonterminal symbols in a context-free grammar , which enables the targeted generation of content that expresses requested meanings (i.e., has the requested tags). While previous work has demonstrated methods for requesting content with a single required tag, requests for multiple tags yields a search task over domains that may realistically span quintillions or more elements. In this paper, we reduce Expressionist grammars to symbolic visibly pushdown automata, which allows us to locate in massive search spaces generable outputs that satisfy moderately complex criteria related to tags. While the satisficing of more complex tag criteria is still not feasible using this technique, we forecast a number of opportunities for future directions.
... By virtue of its markup affordance, Expressionist also provides more authorial control, for the reasons described above. In addition to Tracery and Expressionist, two other recent systems, MKULTRA [9] and Dunyazad [14], have employed grammar-based approaches. Instead of CFGs, these systems use variants of a definite-clause grammar. ...
We present Expressionist, an authoring tool for in-game text generation that combines the raw generative power of context-free grammars (CFGs) with the expressive power of free-text markup. Specifically, authors use the tool to define CFGs whose nonterminal symbols may be annotated using arbitrary author-defined tagsets. Any content generated by the CFG comes packaged with explicit metadata in the form of the markup attributed to all the symbols that were expanded to produce the content. Expressionist has already been utilized in two released games and it is currently being used in two ongoing projects. In this paper, we describe the tool and discuss these usage examples in a series of case studies. Expressionist is planned for release in late 2016.
... The obvious reason for this is that NLG is an extremely difficult task. There is a huge amount of domain knowledge that its practitioners must have, and beyond this there remain several major technical challenges in incorporating NLG into playable media (Horswill 2014). We contend, however, that another major reason that NLG has yet to be popularly or even successfully incorporated into these media is that the current state of the art is simply inadequate. ...
Natural language generation (NLG) has been featured in at most a handful of shipped games and interactive stories. This is certainly due to it being a very specialized practice, but another contributing factor is that the state of the art today, in terms of content quality, is simply inadequate. The major benefits of NLG are its alleviation of authorial burden and the capability it gives to a system of generating state-bespoke content, but we believe we can have these benefits without actually employing a full NLG pipeline. In this paper, we present the preliminary design of Expressionist, an in-development mixed-initiative authoring tool that instantiates an authoring scheme residing somewhere between conventional NLG and conventional human content authoring. In this scheme, a human author plays the part of an NLG module in that she starts from a set of deep representations constructed for the game or story domain and proceeds to specify dialogic content that may express those representations. Rather than authoring static dialogue, the author defines a probabilistic context-free grammar that yields templated dialogue. This allows a human author to still harness a computer's generativity, but in a capacity in which it can be trusted: operating over probabilities and treelike control structures. Additional features of Expressionist's design include arbitrary markup and realtime feedback showing currently valid derivations.
... This is perhaps the hardest challenge we present in this paper, but we propose one potential compositional representational strategy that employs generative dialogue and procedural animation. A number of projects have explored using techniques from natural language generation to facilitate generative dialogue in interactive narrative systems [20,69,63,51,35]. A handful of other projects have explored expressive procedural character animation [57,36]. ...
We introduce a research framework for the design of interactive experiences in the domain of emergent narrative, an application area of computational narrative in which stories emerge bottom-up from the behavior of autonomous characters in a simulated storyworld. Prior work in this area has largely concerned the development and tuning of the simulations themselves from which interesting stories may reliably emerge, but this approach will not necessarily improve system performance at its most crucial level—the actual interactive experience. Looking to completed experiences, namely simulation games like Dwarf Fortress and The Sims, we identify a series of shortcomings that yield four design challenges at the level of interaction: modular content, compo-sitional representational strategies, story recognition, and story support. In this paper, we motivate and discuss each of these design challenges and, for each, summarize prior work and propose new approaches that future work might take.
We present Bronco: an in-development authoring language for Turing-complete procedural text generation. Our language emerged from a close examination of existing tools. This analysis led to our desire of supporting users in specifying yielding grammars, a formalism we invented that is more expressive than what several popular and available solutions offer. With this formalism as our basis, we detail the qualities of Bronco that expose its power in author-focused ways.KeywordsText generationAuthoring toolsProgramming languageGrammars
Through interaction with the virtual environment and virtual characters, users are able to influence the storyline of many games. The design choice for the style of interactivity can thereby have a crucial influence on the user's experience. However, only a few approaches evaluate different interaction modalities for one system to investigate the impact of design choice on the users' experience. In this paper, we present an experimental approach in which we first reflect on design alternatives concerning a specific element of interactive narratives—user-character dialog—and then investigate user responses to different design options (round-based dialog versus continuous dialog). Results of an experimental evaluation study show that users tend to prefer continuous interaction in a soap-opera-like game environment using typed text input to communicate with virtual characters that act and react using speech output, although the recognition rate of user utterances of the continuous version was slightly worse compared to the round-based version.
This paper describes an implementation of the 3T robot architecture which has been under development for the last eight years. The architecture uses three levels of abstraction and description languages which are compatible between levels. The makeup of the architecture helps to coordinate planful activities with real-time behaviours for dealing with dynamic environments. In recent years, other architectures have been created with similar attributes but two features distinguish the 3T architecture: (1) a variety of useful software tools have been created to help implement this architecture on multiple real robots; and (2) this architecture, or parts of it, have been implemented on a variety of very different robot systems using different processors, operating systems, effectors and sensor suites.
Meaningful choice has often been identified as a key component in a player's engagement with an interactive narrative, but branching stories require tremendous amounts of hand-authored content, in amounts that increase exponentially rather than linearly as more choice points are added. Previous approaches to reducing authorial burden for computer RPGs have relied on creating better tools to manage existing unwieldy structures of quests and dialogue trees. We hypothesize that reducing authorial burden and increasing agency are two sides of the same coin, requiring specific advancements in two related areas of design and technology research: (1) dynamic story management architecture that represents story events abstractly and allows story elements to be selected and re-ordered in response to player choices, and (2) dynamic dialogue generation to allow a single story event to be revealed differently by different characters and in the context of dynamic relationships between those characters and the player. This paper describes SpyFeet, a playable prototype of a storytelling system designed to test this hypothesis.
We combine state-of-the-art techniques from computational linguistics and theorem proving to buiM an engine for playing text adventures, computer games with which the player interacts purely through natural language. The system employs a parser for dependency grammar and a generation system based on TAG, and has a component for resolving referring expressions. Most of these modules make heavy use of inferences offered by a modem theorem prover for description logic. Our game engine solves some problems inherent in classical text adventures, and is an interesting test case for the interaction between natural language processing and inference.
Authoring interactive stories where the player is afforded a wide range of social interactions results in a very large space of possible social and story situations. The amount of effort required to individually author for each of these circumstances can quickly become intractable. The social AI system Comme il Faut (CiF) aims to reduce the burden on the author by providing a playable model of social interaction where the author provides reusable and recombinable representations of social norms and social interactions. Motivated through examples from an in-development video game, Prom Week, this paper provides a detailed description of the structures with which CiF represents social knowledge and how this knowledge is employed to simulate social interactions between characters.
A general method for the generation of natural language narrative is described. It allows the expression, or narrative discourse, to vary independently of the underlying events and existents that are the narrative's content. Specifically, this variation is accomplished in an interactive fiction (IF) system which replies to typed input by narrating what has happened in a simulated world. IF works have existed for about 30 years as forms of text-based computer simulation, instances of dialog systems, and examples of literary art. Theorists of narrative have carefully distinguished between the level of underlying content (corresponding to the simulated world in interactive fiction) and that of expression (corresponding to the textual exchange between computer and user) since the mid-1960s, when the field of narratology began to develop, but IF systems have not yet made use of this distinction. The current project contributes new techniques for automatic narration by building on work done in computational linguistics, specifically natural language generation, and in narratology. First, types of narrative variation that are possible in IF are identified and formalized in a way that is suitable for a natural language generation system. An architecture for an IF system is then described and implemented; the result allows multiple works of interactive fiction to be realized and, using a general plan for narrating, allows them to be narrated in different ways during interaction. The system's ability to generate text is considered in a pilot evaluation. Plans for future work are also discussed. They include publicly released systems for IF development and narratology education, adding a planning capability that uses actors' individual perspectives, and adapting automatic narration to different sorts of interactive systems.
Typescript. Thesis (Ph. D.)--University of Massachusetts at Amherst, 1991. Includes bibliographical references (leaves 260-268).
A behavior language is a reactive planning language, based on the Oz Project language Hap, designed specifically for authoring believable agents-characters that express rich personality, and that, in this case, play roles in an interactive story called Facade.
Colossal Cave Adventure
- W Crowther
- D Woods
Crowther, W., & Woods, D. (1976). Colossal Cave Adventure.
Introducing Exclusion Logic as a Deontic Logic
- R Evans
Evans, R. (2010). Introducing Exclusion Logic as a Deontic
Logic. In Deontic Logic in Computer Science, Proceedings of the
10th International Conference, DEON 2010, Lecture Notes in
Computer Science Volume 6181 (pp. 179-195). Fiesole, Italy:
Springer.
The proper treatment of quantification in ordinary English
- R Montague
Montague, R. (1973). The proper treatment of quantification in
ordinary English. In P. Suppes, J. Moravcsik, & J. Hintikka
(Eds.), Approaches to Natural Language: Proceedings of the
1970 Stanford Workshop on Grammar and Semantics (pp. 221-242). Dordrect, NL.