BIM 3 Flashcards
What is the broad definition of AI?
AI is the attempt to build computer systems to think and act like humans
What is the narrow definition of AI?
AI is taking data input and processing it to produce outputs. They use different software techniques
What are major drivers of AI?
- development of big data
- increase in computer processing and decrease in cost
- e-commerce
- IoT
- social media
What are expert systems?
Represents the knowledge of experts as a set of rules programmed so a computer can assist human decision makers.
What is the knowledge base of expert systems?
knowledge base is the set of rules used to model human knowledge.
What is the inference engine of expert systems?
A strategy to search through the collection of rules(knowledge base) and formulate conclusions
What are 3 components of expert systems?
- user interface
most crucial part as it takes user queries and sends it to the interface engine - knowledge base
the set of rules from knowledge of experts - interface engine searches through the rules and solves specific problems
What are limitations of expert systems?
- experts dont know how they make decisions
- knowledge base can be chaotic will all the rules and need to be continually updated.
- not useful for unstructured problems
What is machine learning?
ML is software that can identify patterns in a large databse without explicit programming. The main focus is to find patters in data and classify data inputs into known or unknown outputs. example is spotify playlist
What is supervised learning?
Supervised learning is when the system is trained by receiving specific examples of desired inputs and outputs by humans. It uses datasets to train algorithms
what is unsupervised learning?
humans do not feed the system examples but instead the system is asked to process the database and report whatever patterns it finds. It clusters unlabeled data sets. Basically just giving them stuff and seeing how they sort or find patterns
What are neural networks?
Neural networks are algorithms trained to classify objects into known categories based on data inputs. Deep learning uses multiple layers of neural networks to reveal underlying patterns in data.
How does neural network learn patterns
It searches for relationships, models, and corrects over and over again. Humans train networks by feeding it data inputs which outputs are known, to help neural networks learn solutions.
What are limitations of neural networks and ML?
- requires large data sets
- not all patterns are sensible
- difficult to understand how system arrived at solution
- no sense for ethics
What are genetic algorithms
based on evolution, it is used to generate high-quality solutions for optimization and search problems by examining large numbers of solutions to the problem. It searches among solution variables by changing component parts using processes like selection or mutation
