Speech recognition adaptation for low resource populations. Automatic speech recognition is an essential attribute of mobile devices and consumer electronics. Unfortunately, as these systems are trained with adult speech, they perform poorly when used by children and people with speaking difficulties. The lack of available training speech from these groups makes developing models for them difficult. We will investigate efficient model adaptation methods that use minimal training data to adapt ex ....Speech recognition adaptation for low resource populations. Automatic speech recognition is an essential attribute of mobile devices and consumer electronics. Unfortunately, as these systems are trained with adult speech, they perform poorly when used by children and people with speaking difficulties. The lack of available training speech from these groups makes developing models for them difficult. We will investigate efficient model adaptation methods that use minimal training data to adapt existing adult speech recognition models for use with children and people with speaking difficulties. The intended outcomes will improve access to automatic speech recognition systems for Australians whose communication with speech-controlled environmental and educational devices is currently restricted.Read moreRead less
Estimation and Control of Noisy Riemannian Systems. Many application areas such as satellite control, computer vision, coordination of rigid bodies, require the estimation and control of systems subject to geometric constraints. Most current algorithms for doing this are deterministic and can fail catastrophically in the presence of noise. This project aims to provide:
(i) Methods for analysing and then redesigning deterministic algorithms to ensure stability in the presence of noise;
(ii) New ....Estimation and Control of Noisy Riemannian Systems. Many application areas such as satellite control, computer vision, coordination of rigid bodies, require the estimation and control of systems subject to geometric constraints. Most current algorithms for doing this are deterministic and can fail catastrophically in the presence of noise. This project aims to provide:
(i) Methods for analysing and then redesigning deterministic algorithms to ensure stability in the presence of noise;
(ii) New design methods that deal with noise in an optimal way;
(iii) Noise resistant methods for distributed consensus seeking systems and cooperative control systems.
The outcomes will advance and benefit spatio-temporal data analysis and coordination in areas such as transport, health and video-security.Read moreRead less
Vector network system identification. This machine learning project aims to provide more reliable ways to identify the structure and function of dynamic networks from both continuous and discrete network data. The project will use all the data and create principled new metrics. This could enable early diagnosis of network faults across a range of applications for example in power systems or diseased human brains. It could also enable discovery of critical functional subnetworks affecting reliabl ....Vector network system identification. This machine learning project aims to provide more reliable ways to identify the structure and function of dynamic networks from both continuous and discrete network data. The project will use all the data and create principled new metrics. This could enable early diagnosis of network faults across a range of applications for example in power systems or diseased human brains. It could also enable discovery of critical functional subnetworks affecting reliable operation in large complex human systems (such as financial systems) or natural systems (such as gene regulatory networks).Read moreRead less