Skip to content
 

Buzzwords on Advanced Methods

osgeo

On the OSGeoJapan-discuss mailing list regarding FOSS4G 2019 KOBE.KANSAI :

今年は、Geo-AIのテーマにマッチした、ディープラーニングの勉強が初歩から体験できるコースが2つあります!

  translation: this year, there are two courses that match the theme of Geo-AI, where you can experience deep learning from the beginning!

A OSGeoJapan wiki page says:    今年のテーマは「Geo-AI」です。
    translation: this year’s theme is “Geo-AI”.

Geo-AI a closer look – “geo” is easy, what about AI?

a working definition of AI*:
 Artificial Intelligence is the theory and development of computer systems able to perform tasks which previously required human intelligence.  AI is a broad field, of which Machine Learning ML is a sub-category.

ds_ml_ai

Data Science overlaps Artificial Intelligence, Machine Learning exists in that overlap.

When people speak about Machine Learning these days, they are often speaking of Deep Learning, a subset of Artificial Neural Networks NN.. networks consisting of connected, simple processors called “neurons”  wikipedia -LINK-nn -LINK-dl This blog post -LINK- describes NN training methods, including “weakly supervised” and “transfer learning.”

Since 2009, supervised deep NN‘s have won many awards in international pattern recognition competitions, achieving better-than-human results in some limited domains. — Schmidhuber 2015 Survey of Deep Learning

Deep Learning is especially effective on data sets like images and sound, and improves with the amount of training data available. Deep Learning is not new, but what is new are the volumes of available labelled training data, and the capacity to compute on them. Many Deep Learning programming libraries are in Python, and many libraries run natively on a GPU. -LINK-sklearn

Data Science uses scientific methods, processes, algorithms and systems to extract knowledge and insights from structured and unstructured data. wikipedia -LINK-ds

Machine Learning as a subject, is the study and application of algorithms and statistical models that computers use to execute some tasks without explicit instruction; sometimes known as predictive analytics. -LINK-ml

* ai-and-the-global-economy-mark-carney-2018