Overall Findings
Requires heavy human oversight.
Used to categorize data and make predictions.
Inputs and outputs are determined humans.
Calculated using programs such as R or Python.
Requires minimal human involvement.
Used to find underlying relationships within data sets.
Outputs are unknown and often unpredictable.
Algorithms are more computationally complex.
Artificial intelligence (AI)programs rely on machine learningalgorithmsto perform novel tasks as they take in new information.
Supervised learning uses sets of data labeled by a human to train AI.
That said, it can be difficult for humans to verify the accuracy of outputs.
Both approaches are sometimes used together.
For example, in semi-supervised learning, the initial training set includes labeled and unlabeled data.
Supervised Learning Pros and Cons
Highly accurate.
More transparent than unsupervised learning.
Possible outcomes are already known.
Can’t classify new data own its own.
Takes a lot of time to train.
Requires a human data expert.
The AI then learns how to label future inputs of unlabeled data from the training set.
Supervised learning is best used when you know what inputs and outputs to expect.
Other practical uses include geographic mapping, news curation, marketing, and predicting real estate values.
Unsupervised Learning Pros and Cons
Less expensive than supervised learning.
Can identify patterns humans can’t.
New inputs can be analyzed in real time.
Requires more computing time and power.
Testing for accuracy is difficult.
Less human control over possible outputs.
Unsupervised learning algorithms look for patterns in sets of unlabeled data.
These AI algorithms learn by comparing the similarities and differences between different data points.
Then, it might get more specific, such as sorting shapes based on their number of sides.
The outputs of unsupervised learning can be hard to predict, and verifying their accuracy can also prove difficult.
This approach to AI has practical applications in cybersecurity, computer vision, quality assurance, and even healthcare.
Unsupervised learning algorithms are useful when you don’t know what outputs to expect.
Detecting abnormalities for quality assurance is a good example since you could’t predict abnormalities.
Another example is recommendation engines for streaming services since new content is always added.
FAQ
In self-supervised learning, the model works independently without people correcting errors.
Humans do this as children during playtime.
Your credit card company likely uses some unsupervised learning to detect fraud detection or spending habits.
People are very good at seeing patterns, but at very large scales, computers can do it faster.
