Intro to AI Flashcards
How does AI work?
Leverages computers and machines to mimic the problem-solving and decision-making capabilities of the human mind
What % of world storgae consists of medical data?
30%
What was the first machine learning application?
The spam filter
What is the recommender system?
A product that a client might be interested in, based on past purchases eg netflix
What are 4 machine learning benefits?
faster detection
higher accuracy
improved effciency with larger data
real-time processing
What is machine learning?
Science of programming computers so they can learn from data
What are the 4 major categories of machine learning systems?
Supervised learning
unsupervised learning
semi-supervised learning
reinforcement learning
What is supervised machine learning?
The training set feed to the algorithm includes the desired solutions called labels
Typical supervised learning task is classification
What are 3 examples of supervised learning?
k-nearest neighbours
linear regression
logistic regression
What is unsupervised ML?
training data is unlabeled and system tries to learn without a teacher
typical tasks are clustering, anomaly detection and dimensionality reduction
What is semi-supervised ML?
algorithms that can deal with data that is partially labelled eg google photos recognising faces
What is reinforcement learning?
Learning system called an agent can observe the environment, select and perform actions and get rewards/penalties in return
What is a policy in reinforcement learning?
Defines what action the agent should choose when it is in a given situation
What is deep learning a subdomain of?
Machine learning
What does deep learning use?
Deep artificial neural networks to mimic the learning process of the human brain
What type of learning is the most popular technique in the medical imaging field?
Deep learning for image segmentation, image and text classification and object detection
What are CNNS?
Convolutional neural networks which perform most image-based tasks in medical imaging
What are the 3 layers in CNN?
Convolutional
pooling
fully-connected
What is convolution?
A kernal is applied to the input data to extract features and produce a feature map
What is pooling?
Down-sampling operation that reduces the in-plane dimension of the feature maps
What is the fully connected layers?
Maps features extracted by both convolutional layers and pooling layers to the final outputs of the model
How is AI used in general radiography?
A 3D camera utilising DL (CNN) automatically detects the thorax, even when patients are clothed, and collimates accordingly
What is the GE critical care suite 2.0?
Collection of AI algorithms embedded on mobile x-ray systems for assessing AP CXPs
How is AI used in automated equipment maintenance?
ML gives an opportunity to turn long, unexpected downtimes into scheduled maintenance by predicting equipmemnt failure in advance
What are 4 legal and ethical issues of AI?
Bias
lack of transparency and explainability
human expert deskilling
safety and privacy concerns
