Mizzou Health Sciences 4100 Exam 1

Algorithm

A set of precise instructions that a computer can follow to solve a problem or complete a task (Recepie)

Artifical Intelligence (AI)

A branch of computer science focused on creating intelligent machines that can mimic human cognitive functions like learning, problem-solving, and decision-making.

Machine Learning

A type of AI where computers learn from data without needing explicit programming for every situation.

Artificial Neural Networks (ANNs)

Computational models made up of interconnected nodes (artificial neurons) that process information and transmit signals to each other

What is ANN loosely inspired by?

The human brain!!

Deep Learning

a subfield of machine learning that utilizes artificial neural networks with many layers

Language Models

AI models trained specifically to handle and understand

human language

Generative AI

refers to AI models that can create entirely new data, like

images, text, or audio. These models are trained on large

datasets of existing content and learn to identify the underlying patterns and relationships

If output is incorrect ______ helps identify where the deeplearning tool went wrong

backpropogation

Supervised learning

Tool is trained on a labeled data set so the tool can learn to label new, unseen data in the same way. EX: Diagnostic images, treatment plans

Unsupervised learning

Tool is trained on an UNlabeled data set. Finds patterns and relationships within the data that weren't previously recognized

Structured Prompting + Steps

Turn the AI into a tool designed for a

specific purpose. STEPS:

Role and Goal

Step by Step instruction

Constraits

Personalization

Role and Goal

Tell the AI who it is and how it should behave

Step-by-Step Instructions

Give it very specific instructions on

what it should be doing

Constraints

Clearly identify things the AI should avoid doing in

this task

Personalization

Make sure the tool knows to provide a response

that reflects the needs of each user.

What are the 3 types of machine learning tasks?

Classification

Regression

Clustering

Classification

Known classes/categories

Regression

Numerical data

Clustering

New classes/categories

Gradient descent

works by slowly changing the model's weights to reduce the loss (error) between predicted values and the actual data

Overfitting

when a model fits the training data so closely that it loses the ability to generalize to new data

THINK: Studying only practice questions and not concepts

Activation function

determines whether a neuron in a neural network should "activate" (pass information forward) or not

Weighting

numbers in a neural network that determine how important each input is

Convolutional Neural Networks (CNN)

A type of neural network designed to analyze images and recognize patterns within them

- Look at an image piece by piece instead of all at once. They scan small areas of the image to detect patterns like edges, shapes, textures, and objects.

Recurrent Neural Networks (RNN)

a neural network that processes data in sequence and uses information from previous steps to understand patterns over time

System 1 Thinking

fast, automatic, and intuitive decision making/problem solving

System 2 thinking

slow, deliberate, and analytical decision making/problem solving

Diagnostic Decision Support Systems (DDSS)

computerized tool that analyzes patient data to help healthcare providers identify possible diagnoses and make better clinical decisions

What is the challenge of DDSS?

The output relies on the quality of the input. A lot of input data is incomplete.

Clinical Decision Support Systems (CDSS)

a health information technology tool that analyzes patient data and provides recommendations or alerts to help clinicians make better medical decisions

What do CDSS output predictions rely on?

quantity and quality of input data

Triage

Emergency severity index

• Trying to project risk efficiently and accurately

Predictive analytics

Predict patient outcomes based on current and past patient data

Precision medicine

Diagnosis and treatment decisions tailored specifically to the individual patient

Patient education

Patient specific information, appropriate complexity to support understanding, accessible 24/7