Module 1 Flashcards
Provide a simple definition of AI
Machines performing tasks that normally require human intelligence
What test did Allan Turing develop
He developed a test to determine whether or not a machine is intelligent
What did Allan Turing believe a machine had to do to be considered AI
Its must provide responses that fooled an interviewer into thinking it was human
What are the common elements in most definitions of AI
Technology – the use of technology and specific objectives for the technology to achieve
Autonomy – some level of autonomy by the technology to achieve those objectives
Human involvement – need for human input to train the technology and identify objectives for it to follow
Output – the technology produces output such as performing tasks, solving problems, or producing content
What is Machine Learning?
The process of training machines to display AI behaviour
Why is AI considered a socio-technical system?
Because AI influences society and society influences AI
Why is there risk in AI?
- Complexity of AI systems
- AI data will change over time
- Usually implemented in very complex environments
What does the OECD Framework help organizations do?
- Classify AI systems
- Examine risks
List the 5 dimensions in the OECD Framework
- People and planet
- Economic context
- Data and input
- AI model
- Tasks and output
Describe the “People and planet” dimension of the OECD Framework
Considers the individuals and groups that may be affected by the system
Focused on how the system impacts things like human rights, environment, and society in general
Describe the “Economic context” dimension of the OECD Framework
The system is looked at according to the economic and sectoral environment which it operates
- Sector
- Business function or model
- Is it critical to operations
- How it was deployed
- Impact of the deployment
- Scale of the system
- Technological maturity of the system – a newer system hasn’t been tested on as much data over time and as it becomes more mature it may grow more effective
Describe the “Data and input” dimension of the OECD Framework
Considers:
- What type of data was used
- Whether expert input was used (human knowledge that gets codified into rules)
- How the data was collected
- Data collection method (machine or human)
- Structure of the data
- Data format
Describe the “AI model” dimension of the OECD Framework
Considers:
- Technical type
- How the model is built
- How it is used
Describe the “Tasks and outputs” dimension of the OECD Framework
Considers:
- Tasks the system performs
- It’s outputs
- Resulting actions from those outputs
- System tasks, systems that combine tasks and actions together
- Evaluation methods that are used to look at how the system performs the tasks that it does
List some common AI use cases
- Recognition (images, speech, face)
- Event detection
- Forecasting
- Personalization
- Interaction support
- Goal-driven optimization
- Recommendation
List examples of the “Recognition” AI use case
- Determine if the individual’s face can be matched to another picture
- Retailer product matches – individual can take a picture of the product they want
- Teaching manufacturing machines to detect defects
- Plagiarism detection
Provide examples of the “Event detection” AI use case
- Fraud detection – ex. credit cards, government programs – they are looking for patterns of fraudulent behaviour
- Sports video – when a particular activity occurred (such as when a touchdown was scored)
- Cyber event & systems management
List examples of the “Forecasting” AI use case
- Business forecasting – sales, revenue, demand
- Ride sharing – when there might be the most demand, and price adjusting
- Weather forecasting
List examples of the “Personalization” AI use case
- Unique customer profiles are created to provide relevant experiences to individuals
- Provides a unique experience for each customer and ultimately can increase sales in retail
List examples of the “Interaction support” AI use case
- Virtual assistants
- Chatbots
List examples of the “Goal-driven optimization” AI use case
- Finding many solutions
- Optimizing supply chain issues
- Driving routes and idle time
List examples of the “Recommendation” AI use case
- Product recommendations
- Viewing recommendations
- Decision support systems
- Healthcare providers making diagnosing diagnosis
- Government for adjudicating disability cases, help identify the best benefits according to the case
What 3 high-level categories can AI be grouped into?
- Artificial Narrow Intelligence (ANI)
- Artificial General Intelligence (AGI)
- Artificial Super Intelligence (ASI)
Describe Artificial Narrow Intelligence (ANI)
- Designed to perform a single or a narrow set of related tasks at a high level of proficiency
- May seem intelligent; however, they operate under a narrow set of constraints and limitations, which is why this type of AI is commonly referred to as weak AI
- While limited in scope, artificial narrow intelligence systems can help boost productivity and efficiency by automating repetitive tasks, enabling smarter decision making and optimization through trend analysis.
- A system designed to play chess is an example of artificial narrow intelligence.
Describe Artificial General Intelligence (AGI)
- Also known as Strong, Deep or Full AI
- Intended to closely mimic human intelligence
- Has been a goal of AI development for decades but, as of today, it remains beyond our reach
- 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
Describe Artificial Super Intelligence (ASI)
- A category of AI systems with intellectual powers beyond those of humans across a comprehensive range of categories and fields of endeavor
- Capable of outperforming humans
- Does not yet exist
- However, experts expect this type of system would be self-aware, capable of understanding human emotions and experiences and evoking its own, thus experiencing 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
- 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
How does Machine Learning work?
- Leverages the use of data and algorithms to enable systems to learn and make decisions repeatedly
- Improves over time without being explicitly instructed or programmed to do so
How are Machine Learning technologies categorized?
Based on the type of training model they rely on
List the types of Machine Learning training models
- Supervised
- Unsupervised
- Reinforcement
Describe supervised ML models
- Supervised learning models learn from a pre-labeled and classified data set
- 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
Provide examples of supervised ML model
- A model that analyzes images of road signs labeled to define the sign’s meaning or purpose
- Having a bunch of images labelled Cat and Dog, having a model identify whether a new image is of a cat or a dog
- Text recognition and spam detection
List the 2 categories of supervised learning models
- Classification
- Regression
Explain what the classification category of supervised learning models does
Produce outputs in the form of a specific categorical response; for example, whether an image contains a puppy
Explain what the regression category of supervised learning models does
Predict a continuous value; for example, estimating a stock price
List 2 widely used supervised learning models
- Support Vector Machine (SVM), used for classification and regression tasks but most widely used for classification objectives
- Support Vector Regression (SVR), most commonly used to produce continuous values.
What does SVM stand for?
Support Vector Machine
What does SVR stand for?
Support Vector Regression
Describe unsupervised ML models
- Do not rely on labeled datasets
- Designed to identify differences, similarities, and other patterns without the aid of human supervision
- Tend to be more cost-efficient and require less effort but are susceptible to producing less accurate outputs and can display unpredictable behaviours
List the 2 categories of unsupervised learning models
- Clustering
- Association rule learning
Explain what the clustering category of unsupervised learning models does
Automatically groups data points that share similar or identical attributes; for example, looking for similarities or patterns in DNA samples
Explain what the association rule learning category of unsupervised learning models does
Identifies relationships and associations between data points; for example, understanding consumer buying habits
Describe reinforcement ML models
- Use a reward and punishment matrix to determine a correct or optimal outcome
- Rely on trial and error to determine what to do or not to do and are rewarded or punished accordingly
- Do not ingest pre-labeled data sets and learn solely through action and repetition, changing or not changing state or by getting feedback from their environment
Provide an example of reinforcement ML model
- Self-driving cars
- Robot navigating a maze or organizing and stocking shelves in a large warehouse
- Generating predictive text (making the model mimic human writing based on feedback)
- Improve the placement of online ads in real-time bidding
- Real example: Amazon’s Warehouse Supply Chain Optimization
How do reinforcement ML models learn?
Actions and decisions that result in a reward reinforce the triggering behavior, incentivizing the model to follow the same tactic in the future. Conversely, errors trigger a penalty and reduce the reward, proportional in size to the scale of the error.
Provide an example of unsupervised ML model
- Detecting fraudulent behaviour in banking data
What do ANI, AGI and ASI stand for?
- Artificial Narrow Intelligence
- Artificial General Intelligence
- Artificial Super Intelligence
What is semi-supervised learning?
- Using a combination of supervised and unsupervised learning processes
- Generally uses a small amount of labeled data and a large amount of unlabeled data
- Aim is to leverage the benefits of both models, improving reliability while reducing costs
In what situations are semi-supervised learning models very useful?
- Scenarios where it is challenging to find or create a large, pre-labeled dataset
Provide examples of semi-supervised learning models
- Image and speech analysis or categorization and ranking of web page search results
- Large Language Models, or LLMs, often rely on semi-supervised learning models
What are Large Language 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 does LLM stand for?
Large Language Model
Describe robotics
- Multi-disciplinary field that encompasses the design, construction, operation and programming of robots
- Stems from engineering and computer science and aims to design machines that can perform tasks, usually specific tasks or duties, without human intervention
What benefit does AI bring to robotics?
AI can introduce efficiency and effectiveness to exponentially improve robotic processes
What benefit do robotics bring to AI?
Robotics can allow AI systems and software to interact with the physical world without human intervention
Provide an example of robotics working with AI
Rumba
What is the Fourth Industrial Revolution or Industry 4.0?
The next stage of industry and manufacturing advancements, enabled by increased interconnectivity and smart automation
What is machine perception?
- Where systems are trained to process sensory information and mimic human senses
- Robotic sensors can provide relevant data through cameras, microphones, pressure sensors, 3D scanners, motion detectors and thermal imaging
Provide an example of using machine perception
A system that can touch, smell and taste produce could improve food production, preparation, or storage
What does RPA stand for?
Robotic Process Automation
What is Robotic Process Automation?
An evolving technology using software robots to automate repetitive and rule-based tasks in business processes, mimicking human actions (like data entry and form processing)
What type of AI models can enhance RPA?
AI such as natural language processing or machine learning
What is an expert system?
A form of AI intended to mimic the decision-making abilities of a human expert in a specific field
List the three main elements that distinguish an expert system from other AIs
- Knowledge base
- Inference engine
- User interface
Describe the knowledge base of an expert system
- Typically consists of an organized collection of facts and information from human experts
- Focused on a specific field or domain
- In some cases, the system is also allowed to gather additional information from external sources
Describe how the inference engine of an expert system works
- Extracts relevant information from the knowledge base and uses it appropriately to solve a problem
- Normally uses a rule-based approach that maps data from the knowledge base to a series of rules, which the system uses to make decisions in response to input