Readings Week 1 Flashcards
Ground truth in AI training:
labels assigned to the data to train a ML model to link inputs and outputs.
Lebovitz et al (2021) - “Is AI ground truth really “true”? The dangers of training and evaluation AI tools based on experts’ know-what.”:
What is the additional risk of relying too much on AI?
Human data → used to teach AI → AI gets very reliable → only AI is used → only AI based data output → AI data used to teach AI → might inhibit learning and process improvement)
What is a smart object?
Objects that are able to collect, process and communicate data concerning its functionality and operating environment
Furtak et al (2015) Designing Information Systems with Predictive Analytics (DISPA) (4 steps)
- Goal definition
- Develop / build
- Justify / evaluate
- Process evaluation and conclusions
What is a knowledge graph according to Chaudhri et al (2021)?
“A knowledge graph is a directed labeled graph in which we have associated domain specific meanings with nodes and edges.
Anything can act as a node, for example, people, company, computer, etc. An edge label captures the relationship of interest between the nodes, for example, a friendship relationship between two people, a customer relationship between a company and person, or a network connection between two computers, etc.”
Why use KGs for data integration (3)? Chaudhri et al (2021)
- Reduces the costs of starting with a data integration project
- Easy to adapt
- Very suited for answering questions that require examining relationships across the graph
Characteristics of modern KGs (3)? Chaudhri et al (2021)
- Bigger scale
- Bottom up design approach
- More automation
Front end technologies: (4) Frank et al (2019)
- Smart manufacturing
- Smart products
- Smart supply chain
- Smart working
Base technologies (4): Frank et al (2019)
- Internet of things
- Cloud services
- Big Data
- Analytics