Machine Learning' and Earining
This particular type of AI model mimics the structure of nerve cells being connected to each other in the brain, with an input layer, one or more “hidden” layers of cells, and an output layer.
Neural Network
These APIs typically are designed to access specific named resources (think a user account, a photo, a comment on a file) through URLs.
REST (Representational State Transfer) APIs:
This is a type of database used specifically to support analysis for the purpose of making better business decisions (rather than to implement a product for customers)
Data warehouse
Loosing this type of authentication method might make you lose your identity
Password
This model training technique involves given a set of inputs, a set of labeled outputs, and the goal is to learn how to take inputs and choose the right label
Supervised Learning
These APIs use many different endpoints to access specific resources and use a single endpoint that accepts queries written in the same language as this API, to fetch data
GraphQL API
A centralized repository storing both structured and unstructured data in its native formats, alongside metadata tags and unique identifiers.
Data Lake
This problem can occur commonly when training models where the model isn't big enough, the training algorithm is too simple, or if you don't have enough training data
Underfitting
This is an older format for making APIs, and it is distinguished by the use of XML as the message format for requests and responses, which are typically sent over HTTPS
SOAP (Simple Object Access Protocol) APIs
Software that is meant to address data lake challenges – complexity, sluggish time to value, and overall messiness.
Any one of these four mechanisms can be used to implement authorization - who can access what resource and perform which actions. Name at least one.
Access control lists or permissions or roles or policies
This is a generic version of a large language model designed to work across many uses cases and intended to be accessed by an API
Foundation models
Unlike HTTP-based APIs, these APIs provide a persistent connection between the client and server, enabling real-time data transfer without having to re-establish a connection each time.
WebSocket
Data lakes can be tremendously cheaper than databases when working with high scales which leads to this advantage of a data lake approach
Resource optimization
Without proper authentication and authorization your company may be exposed to this type of risk
Security
The unit of measurement of context widows, a cap on the amount of text models can process
Tokens
This is mechanism for an API provider to provide updates or notifications without having to maintain a persistent connection
Webhooks
The solution to the challenges of data lakes and it allows traditional analytics, data science and machine learning to coexist in the same system, all in an open format.
Lakehouse
Name a third-party vendor used for authentication
Auth0