Week 12: The Future of Informatics, Big Data, AI Flashcards
What is Big Data? List 3 characteristics of it.
enormous amounts of data that can drive business, research, clinical policy
have:
- high volume
- high rate of data capture/receiving – ie. data that needs to be processed and managed quickly otherwise the data becomes outdated
- high variety – ie. structured and unstructured data, multiple different sources and formats making one method of processing data difficult
What is metadata?
details about the data (ie. size, format, date received/sent) – ‘data about data’
- indexing/tagging data into ‘metadata’ allows for efficient processing and use
What are discrete data fields?
data that is measureable and reportable
What is artificial intelligence?
computer programs and systems that can perform tasks that would typically require humans
- various techniques that can be used to achieve this
- used broadly such as in chat bots, self-driving cars, others
What is machine learning?
common form of artificial intelligence
- can have various forms of sophistication and ‘training’ of the machine/computer
- example process of training a machine: collect data, process data, train the model, response generation, iterative improvement
What is deep learning?
variant of machine learning – language model that generates responses that sound natural, based on learning sophisticated patterns and relationships between words
- ie. ChatGPT, Bard (Google), others
Compared to rule-based or statistical methods of machine learning, what does deep learning allow? (3)
- learning of patterns and relationships between words
- incorporate wider range of inputs and contexts
- more natural-sounding output
What are the limitations of deep learning? (3)
- can be biased based on learning material
- struggles with incorporating background knowledge or common sense reasoning
- requires human oversight and filtering
