You programmed a never-AI to distinguish a non existent thing?
HerbaciousTea on
Awesome stuff.
This is the sort of application that “AI” in the current sense, a neural network for large data analytics, is highly applicable. What these systems do excellently is produce a highly refined point of criticality for pattern recognition. They can zoom in on that razor’s edge where random noise is just *barely* discernable as a statistically meaningful pattern much more closely than conventional analysis.
Applications like this lean directly into that strength. This is using the tool for the thing it is most suited. The difficulty lies in building a body of data representative of the reality of the thing you are studying, of having enough resolution in the data to minimize noise your collection process has introduced, and generally addressing biases.
2 Comments
You programmed a never-AI to distinguish a non existent thing?
Awesome stuff.
This is the sort of application that “AI” in the current sense, a neural network for large data analytics, is highly applicable. What these systems do excellently is produce a highly refined point of criticality for pattern recognition. They can zoom in on that razor’s edge where random noise is just *barely* discernable as a statistically meaningful pattern much more closely than conventional analysis.
Applications like this lean directly into that strength. This is using the tool for the thing it is most suited. The difficulty lies in building a body of data representative of the reality of the thing you are studying, of having enough resolution in the data to minimize noise your collection process has introduced, and generally addressing biases.