How do graphs aid AI? [4]
- Artificial neural networks can be representing using graphs
-Graphs provide structures for relationships
-AI problems can be defined as finding a path in a graph
-Graphs may be analysed by a range of algorithms eg. Dijkstra’s
Purpose of A* and Dijkstra’s Algorithms [4]
- To find optimal shortest and most cost-effective route between two nodes
-Part of AI that is meant to simulate the function of a human brain
-Key component of machine learning
-Have self learning capabilities
-Can solve complex problems humans cannot solve
What does an Artificial Neural Network do with the multiple hidden layers? [4]
-Enables deep learning to take place
-Needed when the problem you are tying to solve has a higher level of complexity
-Enables the neural network to learn and make decisions on its own
-Improve accuracy of results, more hidden layers, more complex learning capabilities
Describe what is meant by Deep Learning [4]
-Simulates data processing of the human brain to make decisions
-Uses artificial neural networks which are modelled after the human brain
-Use large number of hidden layers (to progressively extract higher level features)
-Is a specialised form of machine learning
-Trained using large amounts of unlabelled data
How do Artificial Neural Networks enable Machine Learning? [4]
-They are intended to replicate the way the human brain works
-Weights/values assigned between nodes
-Output layer provides the results
-Data is input into the input layer and is passed into the system(then analysed by hidden layers)
-Reinforcement learning takes places through repeated training
-Decisions can be made without being specifically programmed
-The deep learning net will have created complex feature detectors
State the possible reasons for using Deep Learning
- Makes good use of unstructured data
-Outperforms other models with large data sizes
-Effective at identifying hidden patterns
-Can provide more accurate outcomes with higher number of hidden layers
Describe what is meant by Reinforcement Learning [4]
-Based on feedback
-For each good action, the AI gets positive feedback and vice-versa
-Node weightings are adjusted to achieve correct outcome
State a possible reason for using Reinforcement Learning [1]
-Enables autonomous learning using feedback without any labeled data
Describe what is meant by Supervised Learning in Machine Learning [3]
-Allows data to be collected
-A known input and associated outputs are given
-Able to predict future outcomes based on past data
Describe what is meant by Unsupervised Learning in Machine Learning [3]
-Helps all kinds of unknown patterns in data to be found
-Only requires input data to be given
-Uses any data - not trained using a right output
