AI

Question 1

what is AI

Answer 1

science of making machines do things that would require intelligence if done by men

• consist of implementing a number of techniques aimed at enabling machines to imitate a real form if intelligence

Question 2

what is machine learning, give some examples

Answer 2

• identifies patterns with statistical learning and computers by unearthing boundaries in data sets
• can be used to make predictions

e.g speech recognition, fraud detection, online chatbot, recommended engine

Question 3

how does machine learning make predictions

Answer 3

via decision trees

• uses a series of if-then statements to identify boundaries and define patterns in data

Question 4

what is ‘overfitting’ in machine learning and how do you test this

Answer 4

when some boundaries are based on distinctions that dont make a difference

• tested by having test data flow thru the model

Question 5

what is deep learning in ai

Answer 5

literal machine learning, deep learning models taught to perform classification tasks and recognise patterns in photos, texts, audio and various data.

• also used to automate tasks usually needing human intelligence e.g describing images

Question 6

what is neural network in ai

Answer 6

Neural networks are a method of machine learning that tries to mimic how a brain would work

• e.g goes through endless trial and errors. what it gets correct, it remembers and thus connects the correct answers (forming a type of neural network that strengthens over time as it continually gets more correct) (‘neural network’ between incorrect answers get weaker) (eventually it will be able to distinguish between right and wrong
• after endless attempts, it’ll eventually have very few failures

Question 7

what is AI assisted image processing

Answer 7

AI driven image processing platform that gives diagnostic quality imaging from the very first x ray

Question 8

what is ‘FAST’ planning?

Answer 8

FAST Planning facilitates immediate, organ-based setting of scan and recon ranges aiming for a safer, faster and more standardized workflow at the scanner.

Question 9

what are 4 main benefits of using AI in imaging

Answer 9

1. consistent image quality
2. exceptional detail
3. noise reduction
4. metal implant handling (see bone-metal interface without halo artifact) (a scatter correction artifact which derive from high organ-to-background activity ratios)

Question 10

what is veye lung nodule

Answer 10

ai system that runs silently through CT scan and detects and measures lung nodules (rapidly)

Question 11

what is a big issue with data collection and annotation in ai

Answer 11

• bias added unintentionally from data collection
• small change in input = big change in output (ai can jump to conclusion if you aren’t extremely specific with the data u input)

Question 12

what is black box processing / the issue with this idea

Answer 12

This inability for us to see how deep learning systems make their decisions is known as the “black box problem,” and it’s a big deal for a couple of different reasons. First, this quality makes it difficult to fix deep learning systems when they produce unwanted outcomes.

• In science, computing, and engineering, a black box is a device, system, or object which produces useful information without revealing any information about its internal workings. The explanations for its conclusions remain opaque or “black.”

Question 13

Concerns in AI:
Ethics
ai for good or bad?
fair/unbiased?
explainable? (black box)
privacy invasion?