Bibtex citations for all of these publications can be found here.


Publications by Graeme Mann.

"Opportunities for Enhanced Robot Control Along the Adjustable Autonomy Scale"
Graham A. Mann and Nicolas Small, In 5th International Conference on Human Systems Interaction, June 2012.

Abstract: As we move along the scale of adjustable autonomy for the control of robots from direct teleoperation at one extreme to full automation at the other, several opportunities for improvement in control quality, user feedback and machine learning suggest themselves. We describe three experiments, in telerobotics, the provision of situational awareness, and the acquisition of knowledge for automation from the human operator, and explain our concept of explicit, assigned responsibility as an organising principle for flexible work-sharing between humans and robots. A novel design for an interface based on this principle is outlined.
Keywords: human-robot interaction; robot control; context awareness; telerobotics
PDF download:


"A Machine That Daydreams"
Graham A. Mann, In Cultural Computing. Proceedings of the 2nd IFIP TC 14 Entertainment Computing Symposium, eds: Ryohei Nakatsu, Naoko Tosa, Fazel Naghdy, Kok Kwai Wong, Philippe Codognet, pages 21-35. Springer, Germany, September 2010.

Abstract: Some aspects of human cognitive processing across experiential cases in episodic memory seem quite different from conventional artificial reasoning by logical rules, such as that seen in CBR systems. One difference is that in humans, linkages between particular experiences can apparently be made in a number of qualitatively different ways, forming recollective chains of memories along different dimensions. Data-driven, creative, free-association from one memory to the next does not appear to be economically described by rules. Efforts to enable computers to deal with cultural content such as narratives in new ways could benefit from sequential indexing of this kind, provided that the conceptual representations are rich enough, and that a way can be found of modeling the emotional impact each elicits. A conceptual-graph-based FGP (Fetch, Generalise, Project) machine using a knowledgebase of archetypical narratives enhanced with affect is described, which demonstrates how such emotive memory-walks can be computed.
Keywords: cultural computing, narratives, conceptual graphs, affective computing
PDF download:


"Quantitative Evaluation of Human-Robot Options for Maintenance Tasks during Analogue Surface Operations"
Graham A. Mann, In Proceedings of the 8th Australian Mars Exploration Conference, eds: Colin Pain, pages 26-34. July 2008.


"The Low Cost Manufacture of High Technology Machines"
Graham A. Mann, In Proceedings of the 7th Australian Mars Exploration Conference, eds: Graham A. Mann, July 2007.


"Generalising Power of Learned Hidden Hierarchical Learning Model"
Owen Lamont and Graham A. Mann, In Proceedings of 8th IASTED International Conference on Intelligent Systems and Control (ISC 2005), eds: M.H. Hamza, pages 291-296. Anaheim, 2005.


"Crossroads of Meaning: The Four Pole Theory of Symbolic Reference"
Graham A. Mann, In Proceedings of the Computers and Philosophy Conference (CAP 2003), 2003 .


"Delivering on the Promise of Domestic Robots"
Graham A. Mann, School of Information Technology, IT/02/01, 2002 .


"A Dynamically Balancing Walking Biped"
Graham A. Mann, Bruce Armstrong, Philip Preston, and Barry Drake, School of Computer Science and Egineering, UNSW-CSE-TR-0110, 2001.


"Stability Control in an Actively Balanced Bipedal Robot"
Graham A. Mann, In Proceedings of the Australian Conference on Robotics and Automation (ACRA 2000), pages 191-195. Room 121, Building 69, Wellington Road, CLAYTON, VIC, 3168, August 2000.


"Chemical Trail Guidance for a Simple Floor Cleaning Machine"
Graham A. Mann and George Katz, In Proceedings of the 2nd International Conference on Field and Service Robotics, eds: John Bares and Howie Choset, pages 305-309. Pittsburgh, PA, 1999.


"Procedural Renunciation and The Semi-Automatic Trap"
Graham A. Mann, In Conceptual Structures: Theory, Tools and Applications, eds: Mungier, M-L. and Chien, M., pages 319-333. 1998.


"Rational and Affective Linking Across Conceptual Cases - Without Rules"
Graham A. Mann, In Conceptual Structures: Fulfilling Peirce\'\is Dream, eds: Lukose, D. and Delugach, H. and Keeler, M. and Searle, L. and Sowa, J., pages 460-473. 1997.


"Control of a Navigating Rational Agent by Natural Language"
Graham A. Mann, PhD thesis, Computer Science and Engineering, The University of New South Wales, Kensington, New South Wales 2052, December 1996.

Abstract: This work describes the design, construction and testing of BEELINE, an integrated, language-using rational agent, capable of accepting raw, paragraph-length instructions and using them to navigate in a simulation of the physical world. In the development of this software, some difficult problems of natural language understanding are directly confronted. To overcome a knowledge system's potential disengagement from the world, the thesis offers theoretical and practical solutions to the problems of building, situating and grounding conceptual graph knowledge structures, which serve as internal representations of the world and the language used to describe it. It is argued that language cannot be properly understood as an abstraction, but that it must be considered as only one source of knowledge informing a more broadly defined cognitive system immersed in an environment. To be robust enough to deal with unprocessed text, an actor-based conceptual parser should be designed to accept the potential contribution of, but not depend on, a syntactical chart parser using a constituent grammar. Semantic parsing processes are to be kept manageable by arbitrary bounding of the text handling and inferencing. A novel pragmatic theory of semantics combines these elements into a model of knowledge interaction. The theory is then extended to allow behavioural control of a rational agent, which interprets and executes the conceptual output of the parser. This model is implemented as BEELINE. Construction of the conceptual graph processor, conceptual parser, rational agent heuristic, and physical world model are discussed. The active components are evaluated on engineering criteria, and their strengths and weaknesses are discussed. By virtue of the agent's placement in a simulated environment, the machine's use of language can also be operationally evaluated using behavioural methods. Measurements of this kind show that with natural language directions, the agent can find a destination significantly more efficiently than it could without them. The strengths, weaknesses and implications of this work for further development of machine understanding are discussed.
Keywords: natural language, conceptual graphs, rational agent, robot navigation
PDF download:


"Conceptual Strutures: Knowledge Representation as Interlingua"
Graham A. Mann, Springer, 1996.


"BEELINE: A Situated, Bounded Conceptual Knowledge System"
Graham A. Mann, In International Journal of Systems Research and Information Science, pages 37-53. 1995.


"What Conceptual Graphs Need for Natural Language Processing"
Graham A. Mann, In Proceeding of the 4th Workshop on PEIRCE, pages 80-87. 1995.


"Navigation Learning by a Mobile Robot"
Nasser Esmail, Graham A. Mann, and Claude Sammut, In Proceedings of the 2nd Iranian Conference on Electronic Engineering (ICEE-94), May 1994.


"Results of a Conceptual Parse Can Inform a Navigating Rational Agent"
Graham A. Mann, In Proceedings of the AAAI Spring Symposium, eds: Tepfenhart, J. and Dick, J.P. and Sowa, J.F., pages 113-126. 1994.


"A Rational Goal-Seeking Agent Using Conceptual Graphs"
Graham A. Mann, In Conceptual Structures: Current Practices, eds: Tepfenhart, J. and Dick, J.P. and Sowa, J.F., pages 113-126. 1994.


"Conceptual Graphs for Natural Language Representation"
Graham A. Mann, The University of New South Wales, 9311, 1993.


"Assembly of Conceptual Graphs from Natural Language by means of multiple knowledge specialists"
Graham A. Mann, In Conceptual Structures: Theory and Implementation, eds: Pfeiffer, H. and Nagle. T., pages 275-282. 1992.


AI-related publications by Graeme Bell.

"Strengthening CAPTCHA-based Web security"
Graeme Baxter Bell, First Monday, 17 (2-6): 33, February 2012.

Summary: How to make any CAPTCHA more resistant to computer attacks without being harder for humans.
Keywords: Artificial Intelligence, CAPTCHA.
PDF download:


"A New CAPTCHA Interface Design for Mobile Devices"
Rosa Lin, Shih-Yu Huang, Graeme B. Bell, and Yeuan-Kean Lee. In: Australasian User Interface Conference, Australasian Computer Science Week (ACSW2011), Perth, Australia, and ACM digital library, 2011.

Summary: A new, secure, and user-friendly approach to graphical CAPTCHAs that is resistant to machine learning and image-recognition approaches, and is suitable for mobile devices.
Keywords: CAPTCHA, HCI, Artificial Intelligence.
PDF download:


"A method for automatic identification of signatures of steganography software"
Graeme B. Bell and Yeuan-Kuen Lee. IEEE Transactions on Information Forensics and Security, 5 (2): 354-358, Jun 2010.

Summary: Binary fingerprinting is amazingly effective against many steganography programs.
Keywords: Steganalysis, Steganography, Artificial Intelligence, Image Processing, Digital Forensics.
PDF download:


"An Efficient Segmentation Algorithm for CAPTCHAs with Line Cluttering and Character Warping"
Shih-Yu Huang, Yeuan-Kuen Lee, Graeme B. Bell, and Zhan-he Ou. Multimedia Tools and Applications, 48 (2): 267-289, 2010.

Summary: An efficient image-processing technique for breaking various types of CAPTCHAs.
Keywords: CAPTCHA, Segmentation, Recognition, Turing test, Image Processing, Artificial Intelligence.
PDF download:


"A Projection-based Segmentation Algorithm for Breaking MSN and YAHOO CAPTCHAs"
Shih-Yu Huang, Yeuan-Kuen Lee, Graeme B. Bell, and Zhan-He Ou. In Proceedings of the 2008 International Conference of Signal and Image Engineering (ICSIE'08), London, UK, pages 1-. Jul 2008.

Summary: An improved segmentation algorithm for attacking MSN/Yahoo CAPTCHAs.
Keywords: CAPTCHA breaking, segmentation, recognition, CAPTCHA, Image Processing, Artificial Intelligence.
PDF download:


"The Existence of Local Minima in Local-Minimum-Free Potential Surfaces"
Graeme B. Bell and Michael Livesey. In: Towards Autonomous Robotic Systems (TAROS 2005), London, UK, 2005.

Summary: There are mathematical objects that behave like local minima but which do not conform to the normal ideas of local minima. This is particularly important for techniques that rely on minima-detection heuristics.
Keywords: Artificial Intelligence, Potential Fields, Local Minimum, Navigation, Robotics, Computer Games.
PDF download:


"Forward Chaining for Potential Field Based Navigation"
Graeme B. Bell. PhD thesis, University of St Andrews, 2005 .

Summary: A general introduction to navigation, potential fields and a new fast, simple potential field technique for navigation in robotics and computer games. It doesn't rely on minima-free fields (inefficient) or random behaviour (inefficient, unreliable), and doesn't make reference to any information outside the potential field. The family of techniques is based on a modified discrete gradient descent algorithm.
Keywords: Artificial Intelligence, Navigation, Potential Fields, Robotics, Computer Games, Local Minimum, Cinematic Agents.
PDF download:


"Agent Navigation Using Potential Fields and Forward Chaining"
Graeme B. Bell and Michael Weir. In PREP 2004 Postgraduate Research Conference in Electronics, Photonics, Communications and Networks and Computing Science. 2004.

Summary: Increasingly competent heuristics produce the novel Forward Chaining heuristic, allowing navigation past obstacle configurations that are impossible for traditional potential field methods.
Keywords: Artificial Intelligence, Navigation, Potential Fields, Robotics, Computer Games, Local Minimum.
PDF download:


"Forward Chaining for Robot and Agent Navigation using Potential Fields"
Graeme B. Bell and Michael Weir. In 27th Australasian Computer Science Conference (ACSC2004), Dunedin, New Zealand. 2004.

Summary: Introducing a fast, efficient potential fields-based navigation technique that isn't vulnerable to the local minima problem in navigation. Suitable for robots and games.
Keywords: Artificial Intelligence, Navigation, Potential Fields, Robotics, Computer Games, Local Minimum.
PDF download:


"The Incorporation of Intentional Action into Robots"
Michael Weir and Graeme B. Bell. School of Computer Science, University of St Andrews. 2001.

Summary: How does a robot become able to act intentionally?
Keywords: Robotics, Behaviour, Determinacy, Plasticity, Potential Fields, Artificial Intelligence.
PDF download: