Systems that process and learn from time-series data are of critical interest to the Artificial Intelligence community. The brain is one such system which builds high level concepts from such data; however, significant progress is still required to build systems with brain performance. To realise this goal Artificial Intelligence requires a number of breakthroughs in time-series analytics research. Breakthroughs that allow: multi-modal processing, autonomously discovery of invariants, unsupervised feature learning, multi-timescale structure learning, and operation on asynchronous timescales.
In addition to the unsolved research challenges the explosion of low cost and low power sensing technologies is now driving an ever growing number of applications requiring time-series analytics. These include personal wearables, transport, smart homes, health, energy, manufacturing and other industries. A range of technical challenges arise from the high dimensional, heterogeneous and noisy time series data sets associated with these applications. Development and understanding of these emerging sensing applications will in turn drive future research effort.
We call for papers at the intersection of time-series analytics with applications. Specifically within the following relevant topics:
- Time series machine learning or data mining.
- Application of deep learning methods to time-series.
- Time series classification, detection, regression or prediction.
- Time series representation, compression, feature engineering, feature learning.
- Online learning from time series streams.
- Learning from multi-modal time series.
- Applications of time series.
- Time series with sparse or irregular sampling and/or special types of measurement noise.
- Analytics of time series with special structures including hierarchical, spatio-temporal and relational.
- Analytics for multi-scale, multi-variate, heterogeneous, or asynchronous time series
- New architectures for time-series analytics
All submissions will go through a rigorous double blind review process conducted by the program committee. Accepted papers are to be published by ACM.
Further details at www.wp.csiro.au/tsaa2016