Online training and updating of factorization machines using an alternating least squares optimization

摘要

Techniques are disclosed for training factoring machines (FMs) using streaming mode alternating least squares (ALS) optimization. One methodology that implements the techniques according to one embodiment includes receiving a data point that includes a feature vector and an associated target value. The feature vector includes a user identification, an item identification, and a context. The target value identifies an opinion of the user relative to the item. The method further includes applying an FM to the feature vector to generate an estimate of the target value, and updating parameters of the FM to train the FM. The parameter update is based on the application of a streaming mode ALS optimization on: the data point; the estimated value of the target value; and on an updated sum of intermediately calculated terms generated by applying the streaming mode ALS optimization to previously received data points associated with previous parameter updates of the FM.