Artificial Intelligence and Human and Machine Teaming
There has been a lot of excitement about the use of computer vision (CV) and machine learning (ML) technologies for geospatial imagery analysis. As there are many use cases where these technologies can be employed effectively to solve real problems, it is tempting to try and apply them on any problem involving the use of imagery. CV and ML solutions do not typically yield perfect answers, and therefore human+machine teaming becomes a natural approach for a solution. However, if machines and humans are brought together naively, where humans search for and fix mistakes made by machines, any efficiency gains produced by automation can be lost because often, the effort of finding and fixing mistakes is just as laborious as a full-on manual solution. In this presentation, we will talk about various factors that determine whether CV/ML technology can be put to good use on a given problem. We will discuss rules of thumb that allow one to pick problems likely to be solvable with the technology.
Training Goals:
1. Understand the opportunities and limits of AI, CV, and deep learning (DL) technology.
2. Learn how to assess whether a problem is solvable with AI, CV, and DL.
3. Learn ways in which humans can interface with the technology to achieve a mission.
Prerequisites:
• No prerequisites other than an interest in human and machine teaming using computer vision.
• This is an intermediate-level workshop for geospatial professionals or analysts. Users should be familiar with core geo concepts (raster vs. vector data, coordinate systems, etc.), and should be comfortable reading and executing Python code in Jupyter notebooks.
Relevant Government Agencies
Intelligence Agencies, Other Federal Agencies, State Government, Federal Government, State & Local Government
Event Type
Webcast
This event has no exhibitor/sponsor opportunities
When
Tue, Jul 28, 2020, 1:00pm - 2:00pm
ET
Cost
Complimentary: $ 0.00
Where
Webcast
Website
Click here to visit event website
Event Sponsors
Organizer
USGIF