| 要旨トップ | 本企画の概要 | | 日本生態学会第64回全国大会 (2017年3月、東京) 講演要旨 ESJ64 Abstract |
シンポジウム S01-6 (Lecture in Symposium)
From non-visible to visible
Ecologists investigate large-scale complex phenomena in which systems are under the influence of numerous parameters and their interactions. What patterns will be found in Ecology? Questions in classical ecological studies were mostly addressed by “already visible” known patterns. For example, alpha diversity in tropics is higher than that of temperate regions. Of course, we ecologists have tried more rigorously to describe such “already visible” patterns by improving data collection technology, updating statistical techniques, and striving to understand the mechanisms that cause the patterns. In contrast, recent big data approaches can reveal unknown phenomena of which primary data were too huge to handle. IT can help us finding new patterns hidden in the flood of data. Like in the age of discovery we expect many surprises and new findings in this approach. Furthermore, in the business field big data and AI are now used to predict future outcomes, often successfully. Similar approaches may compensate the generally-low prediction power of ecological models. However, I must note that the aim of ecology as a basic science is to ever-improve our understanding of nature. Good prediction is not necessary equal to good understanding. Ecologists must extract general laws and mechanisms (that is hopefully as simple as possible) from the algorithms generated by AI approaches to big data analyses. With this perspective I believe big data approach will become real endeavor in the science of complex natural systems.