Module 1: Foundations of AI: Types of AI and Machine Learning Flashcards
What are the 3 broad categories of AI?
1) Artificial Narrow Intelligence
2) Artificial General Intelligence
3) Artificial Super Intelligence
What is Artificial Narrow Intelligence? What are the benefits? What are the drawbacks? Provide examples.
AI designed to perform a single or a narrow set of related tasks at a high level of proficiency. (AKA “Weak AI”)
Benefits:
- Boost productivity and efficiency by automating repetitive tasks
- Enabling smarter decision making
- Optimization through trend analysis
Drawbacks:
- Cannot transfer information.
Example: Voice assistants, systems designed to play chess.
What is Artificial General Intelligence?
AI designed to closely mimic human intelligence (still has not been achieved). Experts expect AGI systems will do the following things at a level that is similar to or on par with human capabilities:
- Have strong generalization capabilities
- Be able to think, understand, learn and perform complex tasks
- Achieve goals in different contexts and environments
What is Artificial Super Intelligence?
A category of AI systems with intellectual powers beyond those of humans across a comprehensive range of categories and fields of endeavor (still has not been achieved). Once developed, the expectation is that it would be:
- Self-aware
- Capable of understanding human emotions and experiences
- Able to experience reality like humans
What is Broad Artificial Intelligence?
A category of AI more advanced in scope than artificial narrow intelligence, capable of performing a broader set of tasks, but not sophisticated enough to be considered AGI. Broad artificial intelligence often involves reliance on a group of artificial intelligence systems, capable of working together and combining decision-making capabilities, but still lacking the full human-like capabilities experts expect of AGI.
What is Machine Learning?
A branch of AI that leverages the use of data and algorithms to enable systems to learn and make decisions repeatedly. It improves over time without being explicitly instructed or programmed to do so.
What are the 3 main types of Machine Learning?
1) Supervised Learning
2) Unsupervised Learning
3) Reinforcement Learning
What is the definition of Supervised Learning?
Labeled data grouped or classified into categories via the AI system. (examples: email spam detection, fraudulent transaction data)
What is the goal of Supervised Learning?
Accurate predictions of the output of new data.
What is needed in order for Supervised Learning to predict accurate results?
High quality labeled data
What are some challenges with Supervised Learning?
- It requires a large amount of labeled data, which can be expensive.
- Labeling data can introduce bias.
What are the characteristics of a Supervised Learning Model? Provide an example.
A Machine Learning model which learns from a pre-labeled and classified data set. As part of a supervised learning training process, an algorithm analyzes the input data and associated labels to produce an inferred function, which can then become the basis for the system to make predictions based on new, previously unseen inputs. Supervised learning models can also compare their outputs with the correct or intended output, to identify errors and improve their prediction skills.
Example: A model that analyzes images of road signs labeled to define the sign’s meaning or purpose.
What are the subcategories of Supervised Learning?
1) Classification Models
The models produce outputs in the form of a specific categorical response; for example, whether an image contains a puppy.
2) Prediction Models predict a continuous value; for example, estimating a stock price.
Name the 2 types of Prediction Models
1) Regression Models: a type of prediction model that predict output values by identifying linear relationships between values. For example, a regression model could predict the price of a car based on data such as its year, model, make, features and mileage.
2) Classification Models: a type of prediction model that predict categorical output variables by labeling input data. For example, a classification model could predict whether a future email is spam based on prior emails being labelled “Spam” or “Not spam.”
Name a couple of widely used Supervised Learning Models.
1) Support Vector Machine (SVM):
Used for classification and regression tasks but most widely used for classification objectives.
2) Support Vector Regression (SVR):
Most commonly used to produce continuous values.
What is the definition of Unsupervised Learning?
Unlabeled data; typically used for pattern detection. (examples: customer segmentation - purchasing preferences; anomaly detection)
What is the goal of unsupervised learning?
To find patterns, structures or relationships without predefined targets.
What is a challenge with Unsupervised Learning?
Interpretations may be subjective.
What are the characteristics of an Unsupervised Learning Model? What are the weaknesses?
Unsupervised learning models do not rely on labeled datasets. Instead, they are designed to identify differences, similarities and other patterns without the aid of human supervision. These models tend to be more cost-efficient and require less effort, but are susceptible to producing less accurate outputs and can display unpredictable behaviors.
What are the categories of Unsupervised Learning Models? Name some use cases.
1) Clustering:
Automatically groups data points that share similar or identical attributes; for example, looking for similarities or patterns in DNA samples.
2) Association Rule Learning: Identifies relationships and associations between data points; for example, understanding consumer buying habits.
Use cases:
- Anomaly detection for mechanical faults
- Fraud detection
- Consumer segmentation
- Marketing strategies
- Genetics
What is a strength of using Unsupervised Learning?
Discovering hidden patterns and insights in data.
Describe two models that are in between Supervised and Unsupervised.
1) Semi-supervised learning models: This approach generally uses a small amount of labeled data and a large amount of unlabeled data. (Examples: Image analysis, speech recognition, and categorization and ranking of web page search results.)
2) Large Language Models (LLMs):
Often rely on semi-supervised learning models. They are a form of AI using deep learning algorithms to create models trained on massive text data sets to analyze and learn patterns and relationships among characters, words and phrases.
What is a benefit of using Semi-Supervised Learning?
It reduces the need for expensive manual labeling.
What are some challenges with using Semi-Supervised Learning?
- The quality and consistency of the labeled data may be subpar.
- It can be difficult to choose an effective algorithm that can leverage labeled and unlabeled data.
