STAMP

Reference:

Qiao Liu et al. “STAMP: Short-Term Attention/Memory Priority Model for Session-based Recommendation.” in KDD 2018.

class recbole.model.sequential_recommender.stamp.STAMP(config, dataset)

Bases: SequentialRecommender

STAMP is capable of capturing users’ general interests from the long-term memory of a session context, whilst taking into account users’ current interests from the short-term memory of the last-clicks.

Note

According to the test results, we made a little modification to the score function mentioned in the paper, and did not use the final sigmoid activation function.

calculate_loss(interaction)

Calculate the training loss for a batch data.

Parameters:

interaction (Interaction) – Interaction class of the batch.

Returns:

Training loss, shape: []

Return type:

torch.Tensor

count_alpha(context, aspect, output)

This is a function that count the attention weights

Parameters:

• context (torch.FloatTensor) – Item list embedding matrix, shape of [batch_size, time_steps, emb]

• aspect (torch.FloatTensor) – The embedding matrix of the last click item, shape of [batch_size, emb]

• output (torch.FloatTensor) – The average of the context, shape of [batch_size, emb]

Returns:

attention weights, shape of [batch_size, time_steps]

Return type:

torch.Tensor

forward(item_seq, item_seq_len)

Defines the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

full_sort_predict(interaction)

full sort prediction function. Given users, calculate the scores between users and all candidate items.

Parameters:

interaction (Interaction) – Interaction class of the batch.

Returns:

Predicted scores for given users and all candidate items, shape: [n_batch_users * n_candidate_items]

Return type:

torch.Tensor

predict(interaction)

Predict the scores between users and items.

Parameters:

interaction (Interaction) – Interaction class of the batch.

Returns:

Predicted scores for given users and items, shape: [batch_size]

Return type:

torch.Tensor

training: bool