1. Valuing your data
1.1. where data come from?
1.1.1. from users: work, play, interact with services or platforms
1.1.2. Users often allow organisations to capture and store their data: free access to tools
1.1.3. webpage, cookies, tracking moments on the web
1.1.4. Machine learning provide users targeted advertising, to report on trends, to help business on produce developments
1.1.5. Many no financial cost apps and sites used daily
2. Using "dark patterns" online
2.1. Machine learning models
2.1.1. insights: interfere with human decision making processes
2.1.2. to determine person's act: use information to manipulate the actions
2.2. facilitaing dark patterns
2.2.1. dark patterns: tricks used by user interface (UI) or user experience (UX) --> to manipulate users into doing things: buying, signing up
2.2.2. by machine learning activities
3. Machine learning: reinforcing biases
3.1. Bias
3.1.1. skews the results of an algorithms in favour, against, an idea
3.1.2. systematic error due to incorrect assumptions
3.2. learning from past mistakes
3.2.1. models often rely on data
3.2.2. interrogating and preparing data
3.2.3. unintended biases
3.2.4. randomly sampled data will often include biases
3.3. recognising feeback loops
3.3.1. using historical data
3.3.2. constantly updated and trained with new datasets --> ensure validity remains up to date
4. Artificial intelligence (AI)
4.1. Definitions
4.1.1. computer systems modelling
4.1.2. intelligence of the human mind
4.1.3. Machine, device, mathematical model or software
4.1.4. mimicking the cognitive functions of humans
4.1.5. inform decision making
4.1.6. smart algorithms to understand human speech
4.2. Examples
4.2.1. Automated financial investing
4.2.2. Virtual travel booking agent
4.2.3. Social media monitoring
4.2.4. Conversational marketing bot
4.2.5. Proactive healthcare management
4.2.6. Natural Language Processing (NLP) tools
4.2.7. Manufacturing robots
4.2.8. Self-driving cars
4.2.9. Smart assistants
4.2.10. Disease mapping
4.2.11. Inter-team chat tool
5. Applications of AI
5.1. Types of AI
5.1.1. two categories
5.1.1.1. technologies
5.1.1.2. AI applications
5.1.2. three subfields
5.1.2.1. sensing
5.1.2.1.1. "sense" light, movement, sound, temperature, or other elements
5.1.2.2. comprehending
5.1.2.2.1. to understand inputs and analyse data
5.1.2.3. acting
5.1.2.3.1. to interact with people or other systems
5.2. AI defined
6. Machine learning
6.1. a type of AI, learn patterns from data, subsequently improve future experience
6.2. creating models (algorithms) --> providing insights
6.3. More data --> better prediction
6.4. Supervised learning
6.4.1. target value: e.g., offer credit card to customer
6.4.2. knowns in advance the values it wants to obtain
6.4.3. categories of values
6.4.3.1. approved
6.4.3.2. unapproved
6.4.3.3. high risk
6.4.3.4. low risk
6.4.4. expected values can be numbers
6.5. Unsupervised learning
6.5.1. don't have target value
6.5.2. to find previously unrealised patterns, relationships, or structures
6.5.3. don't know in advance what values are seeking
6.6. Reinforcement learning techniques: models learn through experience
6.6.1. powerful, increase accuracy or predictive power of algorithms
6.6.2. example: self-driving cars