Week 1 Flashcards
What is Artificial Intelligence?
The creation of intelligent machines that react like humans
What is Data Science?
The use of artificial intelligence to solve real problems using data
What is Weak AI?
AI designed for specific tasks
No self-awareness or understanding beyond its trained domain
Examples:
Siri or Alexa can answer questions or set reminders but cant reason
Chess AI can play chess but cant do unrelated tasks such as driving
LLMs, such as GPT-4 or similiar systems
What is Strong AI?
Hypothetical AI with human-like general reasoning
Can learn and adapt to new situations without extra training
(Not yet achieved in reality)
What is Data Mining?
Subset of Artificial Intelligence
Focuses on extracing knowledge from data - how to identify previous unknown patterns, relationships or anomalies in the large data sets that humans can then use to solve a complex problem
This is a manual process that requires human intervention and decision making
What example does not count as Data Mining?
Simple search and query processing
What is Machine Learning?
“Machine learning is a field of study that gives computers the ability to learn without being explicitly programmed
Focuses on teaching computer how to learn to analyse large datasets and “learn” the patterns in it (from training data) to then make predictions on new data
What is Supervised Learning?
Type of machine learning where models are trained using labeled data. Each training sample includes inputs and the corresponding correct outputs (labels), allowing the model to learn by comparing its predictions to the actual outcomes
What is Unsupervised Learning?
Involves training models using data that has not been labeled, categorized, or classified. The model seeks to identify patterns and relationships in the data on its own
What is the difference between Machine and Deep learning?
For machine learning the features must be extracted by a human, very time consuming and difficult process
For deep learning model learns directly from input signal, no human needed at all
What is the Image Segmentation task?
Used to divide image into different subsets where each subset has special characterizations such as texture, shape or colour
What is Data Visualization?
The graphical representation of information and data.
By using visual elements like charts, graphs or maps allows us to understand trends, outliers and patterns in the data
Why is Data Visualization important?
Good way to communicate complex information
Criticial tool in AI, providing an effective way to identify summaries, structure, relationships, differences and abnormalities in the data
What is a good chart to show the composition of data?
Pie Charts as they show categories as a proportion or a percentage of the whole
What is the Grammar of Graphics?
Theory that describes components of a graphic and how they work together
This theory has influenced various visualization tools and libraries such as ggplot2 in R which allows users to create complex graphics through deep understanding of this foundational grammar
Why is Grammar important?
Learning the grammar will not only help you create graphics that you know about know, but will also help you to think about new graphics that could be even better
Without grammar there is no underlying theory and existing graphics packages are just a big collection of special cases
What are the components in Grammar of Graphics?
Data - The data itself
Aesthetic - Maps from variables to aesthetic attributes on graph
Geometries - Geometric object such as line or curve in the plot
Facets - Way to break up the data into subsets and show relationship between subsets
Statistics
Coordinates
Themes - Something you can use to polish visualization of the graph to make it look clearer or nicer
What Grammar of Graphics module do we use in this course?
ggplot2
(gg stands for grammar of graphics)
What are examples of Data Quality Issues?
Missing Values
Duplicate Data
Inconsistent Data
Noise
Outliers
What are ways to address Data Quality Issues?
Remove data with missing values
Merge duplicate records
Generate best estimate for invalid records
What is Feature Transformation?
The process of mapping set of values for a feature to a new set of values to make the representation of the data easier for the analysis
