Bulk learning is sometimes called batch learning, and sequential learning is called online learning.
Batch learning is often introduced as a learning procedure in general explanations of machine learning without the word "batch".
It is a procedure of "collecting data, learning and creating a model", after which predictions and judgments start.
Onlinel learning, as the name suggests, is a method that updates the model each time new data is obtained and always uses the latest model for prediction and judgment.
The features of sequential learning are as follows.
When "big data" began to attract attention in the world, the method of processing it became an issue. Batch learning requires the temporary use of a large amount of computer memory, making it impossible to process big data.
One of several solutions was incremental learning.
In online learning, the previous learning results, that is, model parameters, are modified using new data. At the time of this calculation, there is no need to refer to all past data again, and only the parameters of the model made with past data are referred, so the algorithm does not use much memory.
In batch learning, once training is completed, the model will basically continue to be used. If you want to recreate the model and improve its accuracy because the amount of data has increased, you need to prepare the data from the past and redo the learning process.
As mentioned above, batch learning uses a lot of memory, so it is not possible to redo the learning, or even if it is possible, it takes a very long time. On the other hand, online learning can always prepare the latest state of the model.
Averages can also be computed with online learning. From this formula, we can see that it is sufficient to have the n-th data and the n-1th average values ??to calculate the n-th average value.
NEXT 
AI explainability and interpretability (XAI)
