Foundations of AI
This is the broad goal of AI.
What is accomplishing tasks like a human?
These are the various forms or types that data can take.
What are numbers, text, audio, images, and video?
This type of learning allows you to determine something belongs to a category or not.
What is supervised learning?
This way of extracting value from AI allows companies to market their products or services to targeted audiences at the best times.
What is promotion?
The goal of supervised learning is to do what with data?
What is classification?
This set of algorithms developed by Deepmind are based on using the most studied game in the world as the environment.
What is AlphaGo?
What is defining the learning problem?
This type of learning uses no labels to identify patterns in data
What is unsupervised learning?
Extracting value from AI can be done by maximizing the amount of time systems remain in operation with minimal time on maintenance.
What is to 'produce' ?
During model training, this issue may arise if the model learns the data too well, essentially memorizing the data.
What is overfitting?
This test was used as an intelligence indicator; if the computer passed, it was intelligent. But, this test only proved that for a system to be “intelligent,” it only needed to fool a human.
What is Turing's test?
This final stage of the machine learning process involves continuous feedback on the learning system to identify any changes that might compromise the model's performance.
What are deploying and monitoring?
When a machine learns, it uses rewards to determine if is learning correctly or not.
What is reinforcement learning?
When customers engage with services and/or products, AI adds values by providing a curated individualistic experience
What is personalizing?
These two ways of housing data are distinguished by their organization, or lack thereof, and one is a major inhibitor in using organization's data.
What are structured and unstructured data?
These 3 terms are used interchangeably, but are distinguished by meaning intelligence; specifically, intelligence that we can currently implement with algorithms on computers, and the information manipulation and processing.
What are AI, ML, and data science?
This trial-and-error stage of the machine learning process leads to the identification of 2-3 models that best fit the problem and data.
What is experimenting with algorithms?
What unsupervised algorithm allows you to find a certain number of nearby data points?
What is k-nearest neighbors?
What is projecting/forecasting?
This is a mathematical representation of how to do something/how to behave, based on data/experience.
What is a model?
These "old" school and "new" schools of thought for development of AI/ML have led to the current way of using computers to learn.
What are hard-coding & using algorithms to learn from experience?
These two parts of the data stage describe the data that you have and the data that you need.
What are 'data identification' & 'data assessment' ?
Each of these describes a characteristic of data.
What is a feature?
What are the mobile internet, automation and IoT?
This type of unsupervised learning algorithm means you can extract a web of relationships between data.
Social Network Analysis