Abstract
One of the biggest challenges that scientists are facing is understanding
complex dynamic processes. Such processes can be
measured as multidimensional and multimodal data in different
spatial and temporal dimensions, at different scales of time and
space. A fundamental question is: Can essential patterns that
characterise the dynamics (the changes) of these processes be
revealed through machine learning from data and can these patterns
be represented as knowledge representation, comprehensible
by humans? If such patterns can be learned and interpreted as
new knowledge, then our ability to explain phenomena in nature,
understand mechanisms of human cognition, and predict future
events will be significantly improved.
complex dynamic processes. Such processes can be
measured as multidimensional and multimodal data in different
spatial and temporal dimensions, at different scales of time and
space. A fundamental question is: Can essential patterns that
characterise the dynamics (the changes) of these processes be
revealed through machine learning from data and can these patterns
be represented as knowledge representation, comprehensible
by humans? If such patterns can be learned and interpreted as
new knowledge, then our ability to explain phenomena in nature,
understand mechanisms of human cognition, and predict future
events will be significantly improved.
Original language | English |
---|---|
Pages (from-to) | 341-342 |
Number of pages | 2 |
Journal | Neural Networks |
Volume | 119 |
Early online date | 23 Aug 2019 |
DOIs | |
Publication status | Published online - 23 Aug 2019 |
Keywords
- spiking neural networks
- knowledge representaion
- machine learning