Skip to main content

ELSA (Autoencoder)

Description​

The ELSA (Efficient Latent Sparse Autoencoder) policy implements a scalable shallow linear autoencoder designed for implicit feedback collaborative filtering. It learns item-item relationships by reconstructing user interaction vectors. To improve scalability over models like EASE, it uses a factorized hidden layer structure (low-rank plus sparse).

Policy Type: elsa Supports: embedding_policy, scoring_policy

Configuration Example​

embedding_policy_elsa.yaml
policy_configs:
  embedding_policy: # Or scoring_policy
    policy_type: elsa
    # Training Hyperparameters
    batch_size: 512         # Samples per training batch
    n_epochs: 20            # Number of training epochs
    lr: 0.1                 # Learning rate
    # Model Hyperparameters
    factors: 10             # Rank (number of latent factors) for the low-rank part

Reference​