Showing 1–9 of 9 results for author: Howe, J

EvolvingBehavior: Towards Co-Creative Evolution of Behavior Trees for Game NPCs

Authors: Nathan Partlan, Luis Soto, Jim Howe, Sarthak Shrivastava, Magy Seif El-Nasr, Stacy Marsella

Abstract: To assist game developers in crafting game NPCs, we present EvolvingBehavior, a novel tool for genetic programming to evolve behavior trees in Unreal Engine 4. In an initial evaluation, we compare evolved behavior to hand-crafted trees designed by our researchers, and to randomly-grown trees, in a 3D survival game. We find that EvolvingBehavior is capable of producing behavior approaching the desi… ▽ More To assist game developers in crafting game NPCs, we present EvolvingBehavior, a novel tool for genetic programming to evolve behavior trees in Unreal Engine 4. In an initial evaluation, we compare evolved behavior to hand-crafted trees designed by our researchers, and to randomly-grown trees, in a 3D survival game. We find that EvolvingBehavior is capable of producing behavior approaching the designer's goals in this context. Finally, we discuss implications and future avenues of exploration for co-creative game AI design tools, as well as challenges and difficulties in behavior tree evolution. △ Less

Submitted 1 September, 2022; originally announced September 2022.

Comments: 13 pages, 5 figures. Accepted for publication in Foundations of Digital Games 2022 (FDG '22)

ACM Class: I.2.1; D.2.2

Design-Driven Requirements for Computationally Co-Creative Game AI Design Tools

Authors: Nathan Partlan, Erica Kleinman, Jim Howe, Sabbir Ahmad, Stacy Marsella, Magy Seif El-Nasr

Abstract: Game AI designers must manage complex interactions between the AI character, the game world, and the player, while achieving their design visions. Computational co-creativity tools can aid them, but first, AI and HCI researchers must gather requirements and determine design heuristics to build effective co-creative tools. In this work, we present a participatory design study that categorizes and a… ▽ More Game AI designers must manage complex interactions between the AI character, the game world, and the player, while achieving their design visions. Computational co-creativity tools can aid them, but first, AI and HCI researchers must gather requirements and determine design heuristics to build effective co-creative tools. In this work, we present a participatory design study that categorizes and analyzes game AI designers' workflows, goals, and expectations for such tools. We evince deep connections between game AI design and the design of co-creative tools, and present implications for future co-creativity tool research and development. △ Less

Submitted 29 July, 2021; originally announced July 2021.

Comments: 12 pages, 1 figure. Accepted for publication in Foundations of Digital Games (FDG) 2021

ACM Class: I.2.1; H.5.2

Backjumping is Exception Handling

Authors: Ed Robbins, Andy King, Jacob M. Howe

Abstract: ISO Prolog provides catch and throw to realise the control flow of exception handling. This pearl demonstrates that catch and throw are inconspicuously amenable to the implementation of backjumping. In fact, they have precisely the semantics required: rewinding the search to a specific point, and carrying of a preserved term to that point. The utility of these properties is demonstrated through an… ▽ More ISO Prolog provides catch and throw to realise the control flow of exception handling. This pearl demonstrates that catch and throw are inconspicuously amenable to the implementation of backjumping. In fact, they have precisely the semantics required: rewinding the search to a specific point, and carrying of a preserved term to that point. The utility of these properties is demonstrated through an implementation of graph colouring with backjumping and a backjumping SAT solver that applies Conflict Driven Clause Learning. △ Less

Submitted 11 August, 2020; originally announced August 2020.

Journal ref: Theory and Practice of Logic Programming 21 (2021) 125-144

Conditional Generative Adversarial Networks for Data Augmentation and Adaptation in Remotely Sensed Imagery

Authors: Jonathan Howe, Kyle Pula, Aaron A. Reite

Abstract: The difficulty in obtaining labeled data relevant to a given task is among the most common and well-known practical obstacles to applying deep learning techniques to new or even slightly modified domains. The data volumes required by the current generation of supervised learning algorithms typically far exceed what a human needs to learn and complete a given task. We investigate ways to expand a g… ▽ More The difficulty in obtaining labeled data relevant to a given task is among the most common and well-known practical obstacles to applying deep learning techniques to new or even slightly modified domains. The data volumes required by the current generation of supervised learning algorithms typically far exceed what a human needs to learn and complete a given task. We investigate ways to expand a given labeled corpus of remote sensed imagery into a larger corpus using Generative Adversarial Networks (GANs). We then measure how these additional synthetic data affect supervised machine learning performance on an object detection task. Our data driven strategy is to train GANs to (1) generate synthetic segmentation masks and (2) generate plausible synthetic remote sensing imagery corresponding to these segmentation masks. Run sequentially, these GANs allow the generation of synthetic remote sensing imagery complete with segmentation labels. We apply this strategy to the data set from ISPRS' 2D Semantic Labeling Contest - Potsdam, with a follow on vehicle detection task. We find that in scenarios with limited training data, augmenting the available data with such synthetically generated data can improve detector performance. △ Less

Submitted 10 August, 2019; originally announced August 2019.

Efficient predicate invention using shared "NeMuS"

Authors: Edjard Mota, Jacob M. Howe, Ana Schramm, Artur d'Avila Garcez

Abstract: Amao is a cognitive agent framework that tackles the invention of predicates with a different strategy as compared to recent advances in Inductive Logic Programming (ILP) approaches like Meta-Intepretive Learning (MIL) technique. It uses a Neural Multi-Space (NeMuS) graph structure to anti-unify atoms from the Herbrand base, which passes in the inductive momentum check. Inductive Clause Learning (… ▽ More Amao is a cognitive agent framework that tackles the invention of predicates with a different strategy as compared to recent advances in Inductive Logic Programming (ILP) approaches like Meta-Intepretive Learning (MIL) technique. It uses a Neural Multi-Space (NeMuS) graph structure to anti-unify atoms from the Herbrand base, which passes in the inductive momentum check. Inductive Clause Learning (ICL), as it is called, is extended here by using the weights of logical components, already present in NeMuS, to support inductive learning by expanding clause candidates with anti-unified atoms. An efficient invention mechanism is achieved, including the learning of recursive hypotheses, while restricting the shape of the hypothesis by adding bias definitions or idiosyncrasies of the language. △ Less

Submitted 14 June, 2019; originally announced June 2019.

Comments: 7 pages, 5 figures, Proceedings of the 2019 International Workshop on Neural-Symbolic Learning and Reasoning

ACM Class: I.2.0; I.2.1; I.2.3; I.2.4; I.2.8; I.2.11

Legal Area Classification: A Comparative Study of Text Classifiers on Singapore Supreme Court Judgments

Authors: Jerrold Soh Tsin Howe, Lim How Khang, Ian Ernst Chai

Abstract: This paper conducts a comparative study on the performance of various machine learning (``ML'') approaches for classifying judgments into legal areas. Using a novel dataset of 6,227 Singapore Supreme Court judgments, we investigate how state-of-the-art NLP methods compare against traditional statistical models when applied to a legal corpus that comprised few but lengthy documents. All approaches… ▽ More This paper conducts a comparative study on the performance of various machine learning (``ML'') approaches for classifying judgments into legal areas. Using a novel dataset of 6,227 Singapore Supreme Court judgments, we investigate how state-of-the-art NLP methods compare against traditional statistical models when applied to a legal corpus that comprised few but lengthy documents. All approaches tested, including topic model, word embedding, and language model-based classifiers, performed well with as little as a few hundred judgments. However, more work needs to be done to optimize state-of-the-art methods for the legal domain. △ Less

Submitted 12 April, 2019; originally announced April 2019.

Comments: Accepted to the 1st Workshop on Natural Legal Language Processing (co-located with NAACL2019)

Worst-Case Groundness Analysis Using Definite Boolean Functions

Authors: Samir Genaim, Michael Codish, Jacob M. Howe

Abstract: This note illustrates theoretical worst-case scenarios for groundness analyses obtained through abstract interpretation over the abstract domains of definite (Def) and positive (Pos) Boolean functions. For Def, an example is given for which any Def-based abstract interpretation for groundness analysis follows a chain which is exponential in the number of argument positions as well as in the numb… ▽ More This note illustrates theoretical worst-case scenarios for groundness analyses obtained through abstract interpretation over the abstract domains of definite (Def) and positive (Pos) Boolean functions. For Def, an example is given for which any Def-based abstract interpretation for groundness analysis follows a chain which is exponential in the number of argument positions as well as in the number of clauses but sub-exponential in the size of the program. For Pos, we strengthen a previous result by illustrating an example for which any Pos-based abstract interpretation for groundness analysis follows a chain which is exponential in the size of the program. It remains an open problem to determine if the worst case for Def is really as bad as that for Pos. △ Less

Submitted 26 May, 2004; originally announced May 2004.

Comments: Appeared in Theory and Practice of Logic Programming, vol. 1, no. 5, 2001

ACM Class: D.1.6; D.3.2

Journal ref: Theory and Practice of Logic Programming, vol. 1, no. 5, 2001

Three Optimisations for Sharing

Authors: Jacob M. Howe, Andy King

Abstract: In order to improve precision and efficiency sharing analysis should track both freeness and linearity. The abstract unification algorithms for these combined domains are suboptimal, hence there is scope for improving precision. This paper proposes three optimisations for tracing sharing in combination with freeness and linearity. A novel connection between equations and sharing abstractions is… ▽ More In order to improve precision and efficiency sharing analysis should track both freeness and linearity. The abstract unification algorithms for these combined domains are suboptimal, hence there is scope for improving precision. This paper proposes three optimisations for tracing sharing in combination with freeness and linearity. A novel connection between equations and sharing abstractions is used to establish correctness of these optimisations even in the presence of rational trees. A method for pruning intermediate sharing abstractions to improve efficiency is also proposed. The optimisations are lightweight and therefore some, if not all, of these optimisations will be of interest to the implementor. △ Less

Submitted 18 March, 2002; originally announced March 2002.

Comments: To appear in Theiry and Practice of Logic Programming

ACM Class: D.1.6; f.3.2

Efficient Groundness Analysis in Prolog

Authors: Jacob M. Howe, Andy King

Abstract: Boolean functions can be used to express the groundness of, and trace grounding dependencies between, program variables in (constraint) logic programs. In this paper, a variety of issues pertaining to the efficient Prolog implementation of groundness analysis are investigated, focusing on the domain of definite Boolean functions, Def. The systematic design of the representation of an abstract do… ▽ More Boolean functions can be used to express the groundness of, and trace grounding dependencies between, program variables in (constraint) logic programs. In this paper, a variety of issues pertaining to the efficient Prolog implementation of groundness analysis are investigated, focusing on the domain of definite Boolean functions, Def. The systematic design of the representation of an abstract domain is discussed in relation to its impact on the algorithmic complexity of the domain operations; the most frequently called operations should be the most lightweight. This methodology is applied to Def, resulting in a new representation, together with new algorithms for its domain operations utilising previously unexploited properties of Def -- for instance, quadratic-time entailment checking. The iteration strategy driving the analysis is also discussed and a simple, but very effective, optimisation of induced magic is described. The analysis can be implemented straightforwardly in Prolog and the use of a non-ground representation results in an efficient, scalable tool which does not require widening to be invoked, even on the largest benchmarks. An extensive experimental evaluation is given △ Less

Submitted 16 January, 2002; originally announced January 2002.

Comments: 31 pages To appear in Theory and Practice of Logic Programming

ACM Class: D.1.6; F.3.2