In order to apply machine learning algorithms to conversational AI, we need to build up vector representations of conversations.
Each story corresponds to a tracker which consists of the states of the conversation just before each action was taken.
Featurising a single state works like this: the tracker provides a bag of active_features comprising:
- what the last action was (e.g.
prev_action_listen) - features indicating intents and entities, if this is the first state in a turn, e.g. it's the first action we will take after parsing the user's message. (e.g.
[intent_restaurant_search, entity_cuisine]) - features indicating which slots are currently defined, e.g.
slot_locationif the user previously mentioned the area they're searching for restaurants. - features indicating the results of any API calls stored in slots, e.g.
slot_matches
SingleStateFeaturizer converts all of these states into numeric vectors X, y.
We use the X, y notation that's common for supervised learning, where X is a matrix of shape (num_data_points, time_dimension, num_features), and y is an array of length num_data_points containing the target class labels encoded as one-hot vectors.
The target labels correspond to actions taken by the bot. .. note:: If the domain defines the possible actions: [ActionGreet, ActionGoodbye], two additional default actions are added: [ActionListen, ActionRestart]. Therefore, label 0 indicates default action listen, label 1 default restart, label 2 a greeting and 3 indicates goodbye.
BinarySingleStateFeaturizerconverts all of these features to a binary vectorsX, ywhich just indicates if they're present.e.g.
[0 0 1 1 0 1 ...]
It's often useful to include a bit more history than just the current state in memory. The parameter max_history defines how many states go into defining each row in X.
Hence X has shape (num_states, max_history, num_features). For some algorithms you want a flat feature vector, so you will have to reshape this to (num_states, max_history * num_features).