AI (Artificial Inelegance): A computer trying to intelligently solve problems
ML (Machine Learning): A type of AI which uses mathematical algorithm m’s to intelligently solve a problem
Sentiment Analyses: Determining the sentiment of something. In this case text
Artificial Neural Network (ANN): A type of ML which uses interconnected units (mathematical functions) to learn features of data. These units are organized in layers.
Training: Takeing data with a known desired output and running it through a Neural Network prediction. Then taking difference between the prediction and desired value and changing the behavior of the network to get a better prediction.
CNN (Convocational Neural Network): A type of neural network that picks up well on patterns by special layers in the network. These layers are convolution layers and pooling layers.
FFNN (Forward Feed Neural Network): A neural network that only moves data forward through a network. A given layers only has the output of the last layer.
RNN (Recurrent Neural Network): A neural network that dose not just push data forward but remembers data. It can move data in any direction. Often this data is used so that a given node has both the original data and the output of the last layer.
API (Application Programing Interface): A way for two programs to interchange information. It is like email between programs.
Gate: A math function that is a smaller part of the process of a node. This is just a connivent way to describe the functions diffrent parts of the math the make up the node as a whole.
Layer: A collection of units that are not connected to each other but are all in the same position relative to other layers
Node: A individual math function with a input, output, and some trainable variable
Model: A term used to refer to a AI program or system
GRU (Gradient Recurrent Neural Network): A type of RNN that uses only 3 gates to decide what to do with data. Like LSTM, GRU fixes the Vanishing Gradient problem. GRU is a very good at efficiently learning sequential data.
Vanishing Gradient: A issue with the basic RNN model in which the model is unable to recall information past 5 time steps.
LSTM (Long Short Term Memory): A type of RNN that uses 5 gates it fixes the Vanishing Gradient problem. It is older and less efferent then GRU.
MongoDB: A type of database
DataBase: A place data is stored
NLP (Natural Langauge Processing): A sub-group of AI the involves processing text to extract information from it.
KNN (K-Nearest Naibor): A basic ML model that takes data and plots it in space. When it makes a prediction on new data it plots that data and uses its proximity to other points to make a geuss:
SVM (Support Vector Machine): A early ML method that ploted point in a space then drew a line between all the points. The line was between the two classes of points. If a point is on the left of the line it is Correct and the right Incorrect.
Python: A programing language used to give a computer instructions
TenosorFlow: A tool developed by Google that makes developing AI models far easier