Why in AI

3 flashcards · Computer Science · by J M COOPER

1. Why is uncertainty important in AI? Give 4 reasons with examples.

   Ignorant, weak and lazy? Act now! Epistemic and aleatoric uncertainty (ignorance) There may be gaps in our knowledge that we are unaware of (epistemic uncertainty) or inaccuracies in our data (aleatoric uncertainty) which must be accounted for when reasoning about the world. For example, for a forest fire predictor to be accurate it must take into account that there might be a mad arsonist at large, but we can't say for sure either way. Logical systems can draw weak conclusions Logical systems…

2. Why is search important in AI? Give 4 reasons with examples.

   Model on trial, applied gradient eyeshadow. Problems can be modelled as state space Many AI problems can be modelled as state spaces, and so we can try to solve those problems using search - for example the travelling salesperson problem which has useful real world applications. Agents that learn by trial and error When a sequence of steps to reach a given goal state (e.g. tying a certain knot in a rope) is not known in advance, agents can use search to explore different possible routes to the…

3. Why is machine learning important in AI? Give 4 reasons with examples.

   experts data code commercial When we don't have experts If there are no human experts in a particular domain - for example if we are trying to classify unknown types of stars from a huge data set - machine learning can identify classes and provide novel insights that we would not otherwise have. A way of using big data The world is full of data now. Machine learning enables us to leverage these large data sets in productive and helpful ways - for example, by training a neural network to recogni…