This paper investigates whether Chomsky-like grammar representations are useful for learning cost-effective, comprehensible predictors of members of biological sequence families. The Inductive Logic Programming (ILP) Bayesian approach to learning from positive examples is used to generate a grammar for recognising a class of proteins known as human neuropeptide precursors (NPPs). Collectively, ve of the co-authors of this paper, have extensive expertise on NPPs and general bioinformatics methods. Their motivation for generating a NPP grammar was that none of the existing bioinformatics methods could provide suf cient cost-savings during the search for new NPPs. Prior to this project experienced specialists at SmithKline Beecham had tried for many months to hand-code such a grammar but without success. Our best predictor makes the search for novel NPPs more than 100 times more ef cient than randomly selecting proteins for synthesis and testing them for biological activity. As far as these authors are aware, this is both the rst biological grammar learnt using ILP and the rst real-world scienti c application of the ILP Bayesian approach to learning from positive examples. A group of features is derived from this grammar. Other groups of features of NPPs are derived using other learning strategies. Amalgams of these groups are formed. A recognition model is generated for each amalgam using C4.5 and C4.5rules and its performance is measured using both predictive accuracy and a new cost function, Relative Advantage (RA). The highest RA was achieved by a model which includes grammar-derived features. This RA is signi cantly higher than the best RA achieved without the use of the grammar-derived features. Predictive accuracy is not a good measure of performance for this domain because it does not discriminate well between NPP recognition models: despite covering varying numbers of (the rare) positives, all the models are awarded a similar (high) score by predictive accuracy because they all exclude most of the abundant negatives.
Game programmers rely on artificial intelligence techniques to encode characters' behaviors initially specified by game designers. Although significant efforts have been made to assist their collaboration, the formalization of behaviors remains a time-consuming process during the early stages of game development. We propose an authoring tool allowing game designers to formalize, visualize, modify, and validate game level solutions in the form of automatically generated storyboards. This system uses planning techniques to produce a level solution consistent with gameplay constraints. The main planning agent corresponds to the player character, and the system uses the game actions as planning operators and level objectives as goals to plan the level solutions. Generated solutions are presented as 2-D storyboards similar to comic strips. We present in this paper the first version of a fully implemented prototype as well as examples of generated storyboards, adapted from the original design documents of the blockbuster game Hitman.
Interactive storytelling has been present at the heart of digital entertainment media for over thirty years, however the breadth of its narrative scope has remained stifled. As computational boundaries are eased, so are many of the perceived technical obstacles to generated narrative content. Furthermore, there is a sense that notable commercial successes are thawing the professional bias towards authored content. Powerful tools that permit vast and complex worlds to be built have mined gameplay in the sandbox genre. With much of the content generated procedurally the designers have still maintained a strong authorial voice. Presenting similar solutions within the narrative scope can win further converts provided that they are sensitive to the commercial requirements. In this presentation we will explore what the digital entertainment industry has done in the field of interactive storytelling, explore where successes might be reinforced and imagine what it might achieve in the immediate future.
Interactive Storytelling techniques are attracting much interest for their potential to develop new game genres but also as another form of procedural content generation, specifically dedicated to game events rather than objects or characters. However, one issue constantly raised by game developers, when discussing gameplay implications of Interactive Storytelling techniques, is the possible loss of designer control over the dynamically generated storyline. Joint research with industry has suggested a new potential use for Interactive Storytelling technologies, which stands precisely as an assistance to game design. Its basic philosophy is to generate various/all possible solutions to a given game level using the player character as the main agent, and gameplay actions as the basic elements of solution generation. We present a fully-implemented prototype which uses the blockbuster game Hitman as an application. This system uses Heuristic Search Planning to generate level solutions, each legal game action being described as a planning operator. The description of the initial state, the level's objective as well as the level layout, constitute the input data. Other parameters for the simulation include the Hitman's style, which influences the choice of certain actions and privileges a certain style of solution (e.g. stealth versus violent). As a design tool, it seemed appropriate to generate visual output which would be consistent with the current design process. In order to achieve this, we have adapted original Hitman storyboards for their use with a generated solution: we attach elements of storyboards to the planning operators so that a complete solution generates a comic strip similar to an instantiated storyboard for the solution generated. We illustrate system behaviour with specific examples of solution generation.
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.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.