Artificial Intelligence and Language Technologies

Artificial Intelligence Laboratory

Within the Centre for Knowledge and Interaction Technologies (KIT), the Flinders University Artificial Intelligence and Language Technology Laboratory (AILab) focuses on those aspects of Artificial General Intelligence that have to do with language, learning and cognition. Our capabilities and the technologies we develop and deploy are broadly related to the School of Computer Science, Engineering and Mathematics' research concentrations:

A key focus of the AI/LT Lab is learning – learning the language and culture of our society and the world (including grounded semantics and ontology). A very interdisciplinary approach is taken that seeks to model and develop neurologically plausible psycholinguistic theories as well as engineering commercially viable interface technologies. Learning human language goes beyond understanding speech or parsing sentences or disambiguating multiple senses of words and includes understanding body language, gesture, facial expression and a wide variety of emotions as well as research in AudioVisual Speech Recognition. Learning about the world means understanding what’s out there and how it relates to us and represents the original philosophical concept of Ontology which was the term Powers adopted in the late 1970s and throughout the 1980s as a generalization of Semantics that emphasized the need to connect meanings with the real world rather than just chase words round a dictionary – this concept was popularized by Harnad as Symbol Grounding in the 1990s. Semantics and Ontology without Grounding are only a shadow of reality and are insufficient for understanding, according to Powers, Harnad and an increasing number of Cognitive Scientists.

Robots, physical and simulated, play a major role in our attempts to learn a grounded syntax and semantics in which the computer/system/robot really understands what is being talked about. Both grammar and meaning are learned by children in an unsupervised way by learning patterns in context, and we emulate this with our AI systems. The use of multimodal sensors, including touch, vision and sound, allows for a number of interesting enhancements of the way information is communicated to a computer. For example, combining camera and microphone input allows lip-reading to be used to enhance speech recognition under noisy conditions. In addition, this opens the door to the possibility of picking up additional expressional and emotional content, of tracking where a speaker is looking when talking, and conversely of synthesizing appropriate acoustic and facial expressions, and looking at the objects being talked about. This is a major focus of our Thinking and Teaching Head projects, in which area we hold two current ARC grants and have a commercialized spinoff company/product: CleverMe/Clevertar . This research involves interdisciplinary collaboration both extramurally, including with institutions in the US, Germany and China, and within Flinders, including the Flinders Educational Futures Research Institute

Biometric signals are another source of information, and can not only be used as inputs in their own right, but can be used to correlate with, and thus learn and validate, theories and models of language, learning and emotion. The signal processing and learning expertise developed for speech and language is also being applied to developing new techniques in biomedical signal processing, and in particular for the processing EEG in real world conditions. Our Brain Computer Interface has two facets:

  1. allowing us to understand more of what is going on in a person's brain (including their emotional state and their level of skill acquisition or situation awareness), and
  2. allowing a person to interact with a computer or control devices such as a wheelchair or other vehicle.
Our EEG research is an interdisciplinary collabortion with the Centre for Neuroscience, the School of Medicine's Epilepsy Laboratory and Human EEG Unit and our joint Brain Systems Lab.  We currently hold two joint ARC Discovery grants for this work.  We also cooperate with researchers at the University of Southern Queensland in a number of related initiatives.

Artificial Intelligence and Language Technology at Flinders has expertise and capabilities in the following areas:

Relevant Courses and Topics

The following awards and topics link directly to research within the AI Lab and final year projects are supervised for the courses shown (for further degree combinations see the individual course descriptions):

Bachelor of Computer Science (Honours)
Bachelor of Behavioural Science (Psychology)/Bachelor of Computer Science
Bachelor of Engineering (Biomedical)
Bachelor of Engineering (Electronics)
Bachelor of Engineering (Robotics)
Bachelor of Engineering (Software) 
Bachelor of Information Technology (Honours)
Bachelor of Information Technology (Digital Media) (Honours)
Bachelor of Science (Honours)Computer Science and Information Systems
Master of Engineering (Biomedical)
Master of Engineering (Electronics)
 
COMP3722 Theory and Practice of Computation
COMP3742 Intelligent Systems
COMP3751 Interactive Computer Systems
COMP3752 Computer Game Development
COMP4712 Embodied Conversational Agents
COMP4715 Computational Intelligence
COMP4716 Information Retrieval and Text Processing
COMP4717 Mobile Application Development
ENGR3721 Signal Processing
ENGR3741 Physiological Measurement
ENGR3771 Robotic Systems
ENGR4722 Haptic Enabled Systems
ENGR4761 Image Processing
 

For More Information...

For more information on these projects, or if you're interested in joining the group, or enquiring about courses or scholarships, please contact Professor David Powers.

 

Significant Publications by AILab Members

Proceedings and Monographs

Powers, David M. W., Ed., 1998. Proceedings Joint International Conference on New Methods in Language Processing and Computational Natural Language Learning (NeMLaP3/CoNLL98) Somerset NJ: ACL.

Powers, David M. W. and Reeker, David M. W., Eds., 1991. Proceedings of the AAAI Spring Symposium on Machine Learning of Natural Language and Ontology, Document D-91-09, DFKI, Univ. Kaiserslautern FRG. 

Powers, David M. W. and Turk, Christopher, 1989. Machine Learning of Natural Language, Research Monograph, Springer-Verlag (NewYork/Berlin), 1989, ISBN 3-540-19557-2/0-387-19557-2

Book chapters

Powers, David M. W., 2002. Robot babies: what can they teach us about language acquisition? In J. Leather and J. Van Dam, eds, The Ecology of Language Acquisition, Kluwer Academic pp.160-182.

Sharman, Darryl K. and Powers, David M. W., 1999. Hardware System Simulation in R. Buyya, Ed., High Performance Cluster Computing: Programming and Applications, Vol.2, Prentice-Hall, pp.395-417. Republished in Chinese Translation by Zhen Weiming, Shi Wei and Wang Dongshen, Publishing House of Electronics Industry, Beijing, China, pp. 293-308.

Refereed journal articles

Atyabi, A., Luerssen. M.H. & Powers, D.M.W., PSO-Based Dimension Reduction of EEG Recordings: Implications for Subject Transfer in BCI, 2013, Neurocomputing.

Fitzgibbon, S.P., Lewis, T.W., Powers, D.M.W., Whitham, E.M., Willoughby, J.O., & Pope, K., 2012. Surface Laplacian of Central Scalp Electrical Signals is Insensitive to Muscle Contamination. IEEE Transactions on Biomedical Engineering.

Powers, D.M., 2011. Evaluation: From Precision, Recall and F-Measure to ROC, Informedness, Markedness & Correlation. Journal of Machine Learning Technologies, 2(1), 37-63.

Pfitzner, D.M.W., Leibbrandt, R.E., & Powers, D.M., 2009. Characterization and evaluation of similarity measures for pairs of clusterings. Knowledge and Information Systems, 19(3), 361-394.
 

Lewis, T.W. & Powers, D.M.W., 2003. Audio-Visual Speech Recognition using Red Exclusion an Neural Networks. Journal of Research and Practice in Information Technology, 35(1), 41-64.

Powers, David M. W., 1997. Unsupervised learning of linguistic structure: an empirical evaluation", Int'l Journal of Corpus Linguistics 2(1): 91-131.

 (For more recent publications see individual author or project pages.)

Recent Grant Income

  • Burnham, D. K., Dale, R., Stevens, C. J., Powers, D. M., Davis, C. W., Buchholz, J. M., Kuratate, T., Kim, J., Paine, G. C., Kitamura, C. M., Wagner, M., Moeller, S., Black, A. W., Schultz, T. and Bothe, H. H. (2006-2010). From Talking Heads to Thinking Heads: A Research Platform for Human Communication . ARC Thinking Systems: $3,400,000.
  • Powers, D. M., Clark, C. R., Pope, K. J. and Willoughby, J. O. (2009-2011). Heterodensity neuroimaging techniques for spatiotemporal identification and localization. ARC Discovery: $301,000.
  • Broberg, I. M., Pope, K. J., Shuttleworth, C. W. and Willoughby, J. O. (2010). Cell swelling and EEG ripples: components in the transformation of normal brain activity into seizure. NH&MRC: $357,125.
  • Burnham, D., Cox, F., Butcher, A., Fletcher, J., Wagner, M., Epps, J., Ingram, J., Arciuli, J., Togneri, R., Rose, P., Kemp, N., Cutler, A., Dale, R., Kuratate, T., Powers, D., Cassidy, S., Grayden, D., Loakes, D., Bennamoun, M., Lewis, T., Goecke, R., Best, C., Bird, S., Ambikairajah, E., Hajek, J., Ishihara, S., Kinoshita, Y., Tran, D., Chetty, G. and Onslow, M. (2010). The Big Australian Speech Corpus: An audio-visual speech corpus of Australian English. ARC LIEF: $650,000.
  • Pope, K., Powers, D., Lewis, T. and Willoughby, J. (2011-2013). Enhanced brain and muscle signal separation verified by electrical scalp recordings from paralysed awake humans. ARC Discovery: $225,000.

Further Information

We would be please to supply further information about our activities.  In particular, opportunities exist for high achieving postgraduates to join the program.  Please contact Professor David Powers.

Spin outs & Products

  • Clevertar (CleverMe)   – A talking head helper on iPhone/iPad/Android, initial commecialization funded by a Researcher in Business award.

  • Clipsal Homespeak (I2Net Orion) – Control your home or C-bus installation by talking, initially developed by undergraduates, popular for people with disabilities.

  • Head X – Freely Available Customizable Virtual Head developed under the ARC Thinking Systems SRI, "From Talking Heads to Thinking Heads".

  • YourAmigo – The world leader in organic search optimization and deep web search, Yahoo: “Nobody searches the deep web like YourAmigo”. Winner of numerous grants and industry awards with a worldwide presence and a clientelle of Fortune 500 companies.

 

People

 

Feature

Embodied vs Enheaded

We are often asked if our Talking Head is an ECA, an Embodied Conversational Agent. Well, if you regard your head as part of your body, then I guess it is, but if you regard your body as what is below your neck, then I guess it doesn't quite fit. These correspond to the 1st and 3rd definition in the Macquarie Dictionary – the 2nd one, "corpse", unfortunately may be a better fit for some ECAs.

But really it is not about enheaded vs embodied, but about being embedded in the world, in a culture, in human society, and experiencing and learning from and interacting with this environment.  This is very much what all our Thinking Head work is about!  Sometimes our simulated heads do have bodies in the sense of being mounted on a wheeled vehicle or a robot baby or an articulated robot arm. At other times our focus is on looking at the world with cameras and other sensors, and understanding it, and sometimes it is on the drives and emotions, the expressions and the gestures, that are part of our interaction with the world.

Sometimes we operate in what we call a hybrid world – the Learning or Teaching Head has toys to play with, props to illustrate points, and the Human Teacher or Learner has a corresponding set of toys or other props.  Whatever is in the virtual environment can be replicated in the real world with our 3D printers, and whatever is in the real world can be replicated in the virtual world with our 3D scanners. What's more, our industrial robot arms have to be kept in a cage because of the damage they could cause if something went wrong. 

Of course we also connect lightweight robot arms to our robots, but still real robots and robot arms are much more fragile and finickety than those of the robot worlds we have been working in for the last 30 years! The robot worlds and hybrid worlds allow us to concentrate on the Artificial Intelligence, Computational Intelligence, Intelligent Systems and Cognitive Science aspects of communication, understanding and grounding, rather than getting caught up in the hardware side of engineering.

inspiring achievement