Artificial Unintelligence: Fverything we did wrong
2021-09-30, 14:30–15:00, Ushuaia

In recent years, the password to get into the club of the “cool kids” in technology has been Artificial Intelligence, also referred to as AI by the “it” group. AI has grown greatly in popularity and application with Geo Gecko also recently jumping on the train by starting to work on some Machine learning models which are a subset of the great AI.
We have been building models to identify different crops using sentinel 1 and sentinel 2 images. This work has given us a front row seat in the implementation of the much-glorified machine learning algorithms. It is from this position that we are able to discuss our insights in regard to how “intelligent” this subset of artificial intelligence really is.
Also having experienced the non-romantic side of machine learning (spoiler alert) which is data accessing, cleaning and preprocessing, we will discuss these in depth, alongside the break throughs we made to overcome them, and the recommendations that we have for the newbies. We intend for this talk to give ML enthusiasts a quick dose of reality so that they can take off the training wheels and get to know what really happens in Machine Learning.


In recent years, the password to get into the club of the “cool kids” in technology has been Artificial Intelligence, also referred to as AI by the “it” group. AI has grown greatly in popularity and application with Geo Gecko also recently jumping on the train by starting to work on some Machine learning models which are a subset of the great AI.
We have been building models to identify different crops using sentinel 1 and sentinel 2 images. This work has given us a front row seat in the implementation of the much-glorified machine learning algorithms. It is from this position that we are able to discuss our insights in regard to how “intelligent” this subset of artificial intelligence really is.
Also having experienced the non-romantic side of machine learning (spoiler alert) which is data accessing, cleaning and preprocessing, we will discuss these in depth, alongside the break throughs we made to overcome them, and the recommendations that we have for the newbies. We intend for this talk to give ML enthusiasts a quick dose of reality so that they can take off the training wheels and get to know what really happens in Machine Learning.


Authors and Affiliations

Tasia Lydia
Naturinda Evet

Track

Use cases & applications

Topic

New trends: IoT, Indoor mapping, drones - UAV (unmanned aerial vehicle), Artificial intelligence - machine learning, deep learning-, geospatial data structures, real time raster analysis

Level

1 - Principiants. No required specific knowledge is needed.

Language of the Presentation

English

Lydia is a “something”. Not yet clear what the something stands for but I am a fan of GIS, Programming, Machine Learning, Business and Hair. Can’t decide on one to define me yet. I am currently working with Geo Gecko doing some Machine learning in the crop classification realm.