A sequence is an ordered series of observations, such as daily weather or hourly scooter rentals. For an RNN to process numerical data over time, represent each observation as a numerical vector. section 8.1.1
An RNN can produce an output for each sequence element. For forecasting, use it to predict the next element, then feed that prediction back in to forecast later steps. section 8.1.1
Table 8.1 The many inputs and outputs of RNNs
|
Type
|
Description
|
Applications
|
|---|---|---|
|
One to many
|
One input tensor used to generate a sequence of output tensors
|
Generate chat messages, answer questions, describe images
|
|
Many to one
|
A sequence of input tensors gathered up into a single output tensor
|
Classify or tag text according to its language, intent, or other characteristics
|
|
Many to many
|
A sequence of input tensors used to generate a sequence of output tensors
|
Translate, tag, or anonymize the tokens within a sequence of tokens, answer questions, participate in a conversation
|
RNNs can map a single input to a sequence, a sequence to one output, or one sequence to another. section 8.1
