Talking with Robots: Evolving Grounded Language for Embodied Agents. The coming personal robot revolution will be built on robots that have real-world intelligence, with an ability to understand and communicate about the world in the way we humans do. This project extends a previous ARC project, which developed robot-friendly languages for naming places in the world. This new project will develop the robots' abilities and language to understand a comprehensive range of real world objects, places ....Talking with Robots: Evolving Grounded Language for Embodied Agents. The coming personal robot revolution will be built on robots that have real-world intelligence, with an ability to understand and communicate about the world in the way we humans do. This project extends a previous ARC project, which developed robot-friendly languages for naming places in the world. This new project will develop the robots' abilities and language to understand a comprehensive range of real world objects, places, actions, attributes and relationships. This project represents a major advance for Australia in the new and fast growing personal robot industry.Read moreRead less
Enhancing Intelligent Robot Navigation with the Evolution of a Robot-Friendly Language. Personal robots are set to become as popular as personal computers. The key ingredient that has been missing is intelligence - not the kind of intelligence that plays chess, but the kind that allows robots to understand the world in the way that we humans do. This project represents a major advance in that kind of intelligence, giving robots the ability to understand the world and the ability to communicate a ....Enhancing Intelligent Robot Navigation with the Evolution of a Robot-Friendly Language. Personal robots are set to become as popular as personal computers. The key ingredient that has been missing is intelligence - not the kind of intelligence that plays chess, but the kind that allows robots to understand the world in the way that we humans do. This project represents a major advance in that kind of intelligence, giving robots the ability to understand the world and the ability to communicate about their experiences. Armed with this new technology, Australia will have a competitive edge in the new and fast growing personal robot industry.Read moreRead less
Robot Navigation From Nature: Simultaneous Localisation And Mapping Based On Hippocampal Models. This project will create a new method of robot control that allows a robot to learn a map of any area and then navigate using that map. The new method is based on ideas from recent models of rodent brains.
The resulting improvements in robot navigation offer immediate benefits to the emerging service robot industry. In addition, the act of reproducing a high-level brain function in a robot will inc ....Robot Navigation From Nature: Simultaneous Localisation And Mapping Based On Hippocampal Models. This project will create a new method of robot control that allows a robot to learn a map of any area and then navigate using that map. The new method is based on ideas from recent models of rodent brains.
The resulting improvements in robot navigation offer immediate benefits to the emerging service robot industry. In addition, the act of reproducing a high-level brain function in a robot will increase the understanding of memory and learning in mammals, including humans. Consequently, the outcomes of this research will benefit both robot designers and brain researchers.
Read moreRead less
Evolving largest scale concept structures. This project will find new methods of collective intelligence- many small computer programs acting in synergy. Such software has many applications from data mining to networks of sensors, but the main focus will be on one of the Grand Challenges of artificial intelligence -- the Japanese game of Go. Go is at least as difficult as Chess but computers are far from reaching the skill of human experts. Insights into the human brain from autism and savants w ....Evolving largest scale concept structures. This project will find new methods of collective intelligence- many small computer programs acting in synergy. Such software has many applications from data mining to networks of sensors, but the main focus will be on one of the Grand Challenges of artificial intelligence -- the Japanese game of Go. Go is at least as difficult as Chess but computers are far from reaching the skill of human experts. Insights into the human brain from autism and savants will form the foundations of the new computational approaches we will develop.Read moreRead less
Linkage Infrastructure, Equipment And Facilities - Grant ID: LE100100055
Funder
Australian Research Council
Funding Amount
$400,000.00
Summary
State-of-the-art upgrade to multi-transmit multi-receive technology for research dedicated 3 Tesla magnetic resonance imaging (MRI) scanner. Projects requiring the proposed infrastructure are aligned with two National Research Priorities. The research will lead to new methods for imaging and detecting soft tissue changes, identifying developmental, cognitive and degenerative disorders, and pharmacological research. The understanding of the basis of physiological, cognitive and biochemical proces ....State-of-the-art upgrade to multi-transmit multi-receive technology for research dedicated 3 Tesla magnetic resonance imaging (MRI) scanner. Projects requiring the proposed infrastructure are aligned with two National Research Priorities. The research will lead to new methods for imaging and detecting soft tissue changes, identifying developmental, cognitive and degenerative disorders, and pharmacological research. The understanding of the basis of physiological, cognitive and biochemical processes which will be facilitated by the new equipment will contribute to the priority area Promoting and Maintaining Good Health and will underpin an array of subsequent medical research. The new equipment will extend capabilities and training in signal analysis, biomedical engineering and biomedicine, contributing to the priority area Frontier technologies for Building and Transforming Australian Industries.Read moreRead less
Mechanisms of Recovery after Extinction of Conditioned Behaviour. Old habits die hard and may never die at all. My previous ARC-funded research has revealed that extinguished learning can be recovered rapidly and in unsuspected ways. This project is aimed at building a neural network to explain how old learning can recovered. In practical terms, rapid recovery has both benefits, e.g., our ability to regain old skills with brief refresher training, and costs, e.g., relapse after therapies for anx ....Mechanisms of Recovery after Extinction of Conditioned Behaviour. Old habits die hard and may never die at all. My previous ARC-funded research has revealed that extinguished learning can be recovered rapidly and in unsuspected ways. This project is aimed at building a neural network to explain how old learning can recovered. In practical terms, rapid recovery has both benefits, e.g., our ability to regain old skills with brief refresher training, and costs, e.g., relapse after therapies for anxiety disorders and substance abuse. In theoretical terms, understanding recovery in biological systems will inform research concerning both the neural basis of memory and the design of robots.Read moreRead less
The Analog Mind: Connectionism, Consciousness and Mental Representation. The aim of our project is to explore novel conjectures about the way information is coded and processed in the brain, and the manner in which the brain manufactures consciousness. This work may eventually pay rich dividends by reducing the incidence of mental illnesses, such as schizophrenia, that disturb conscious experience. It may also lead to the development of more powerful forms of computation in artificial neural ne ....The Analog Mind: Connectionism, Consciousness and Mental Representation. The aim of our project is to explore novel conjectures about the way information is coded and processed in the brain, and the manner in which the brain manufactures consciousness. This work may eventually pay rich dividends by reducing the incidence of mental illnesses, such as schizophrenia, that disturb conscious experience. It may also lead to the development of more powerful forms of computation in artificial neural networks. This will have a significant impact on the artificial intelligence industry, with obvious economic benefits for Australia.
Read moreRead less
A unified theory of performance in absolute identification tasks. The ability to identify stimuli is fundamentally important in human cognition and is studied in absolute identification tasks, where people must identify one out of a number of stimuli, varying on a single dimension, with an appropriate label. A remarkable finding is that people cannot reliably identify more than about seven different stimuli. This limit imposes severe practical restrictions on our ability to categorise stimuli an ....A unified theory of performance in absolute identification tasks. The ability to identify stimuli is fundamentally important in human cognition and is studied in absolute identification tasks, where people must identify one out of a number of stimuli, varying on a single dimension, with an appropriate label. A remarkable finding is that people cannot reliably identify more than about seven different stimuli. This limit imposes severe practical restrictions on our ability to categorise stimuli and constitutes a perplexing problem for cognitive theory. This project involves an international collaborative effort by five leading researchers in mathematical psychology to develop a comprehensive, integrative model of human performance in absolute identification tasks.
Read moreRead less
Unconscious processing: To what extent, how flexible and how smart? We are unaware of the sophisticated pattern analyses conducted by our perceptual systems that enable us to operate in a complex environment. Less clear is whether unconsciously presented information itself can influence our behaviour or indeed whether unconscious cognition occurs. The proposed studies provide the foundation and structure for a pioneering examination of the extent, flexibility and computational power of unconscio ....Unconscious processing: To what extent, how flexible and how smart? We are unaware of the sophisticated pattern analyses conducted by our perceptual systems that enable us to operate in a complex environment. Less clear is whether unconsciously presented information itself can influence our behaviour or indeed whether unconscious cognition occurs. The proposed studies provide the foundation and structure for a pioneering examination of the extent, flexibility and computational power of unconscious processing at the level of neural response and behaviour. The impact and benefits of the project will be in its contribution to theory with outcomes revealing the boundary conditions governing unconscious cognition and providing a first benchmark for how such conditions might vary across the lifespan.Read moreRead less
Models of Knowledge Restructuring. In a world of rapid change, people must frequently adapt cognitive skills to new technologies. This often requires restructuring of knowledge, from simple initial strategies to increasingly complex ways of mastering the task. This is not always easy and people sometimes resist knowledge restructuring. We present a network model of knowledge restructuring that combines attentional processes with error-driven learning. The model accounts for several existing resu ....Models of Knowledge Restructuring. In a world of rapid change, people must frequently adapt cognitive skills to new technologies. This often requires restructuring of knowledge, from simple initial strategies to increasingly complex ways of mastering the task. This is not always easy and people sometimes resist knowledge restructuring. We present a network model of knowledge restructuring that combines attentional processes with error-driven learning. The model accounts for several existing results and makes predictions that guide the proposed research. Three factors will be examined empirically: (1) the role of directing attention to alternative strategies; (2) the imperfection of existing knowledge; and (3) the complexity of alternative strategies.Read moreRead less