TTML: tensor trains for general supervised machine learning
This work provides a new estimator for supervised machine learning that reduces memory usage, but it is incremental as it builds on existing tensor train methods and requires initialization from other estimators.
The authors tackled the problem of developing a general-purpose supervised machine learning estimator by using tensor trains to parametrize discretized functions, resulting in a competitive, fast estimator with lower memory usage than other methods.
This work proposes a novel general-purpose estimator for supervised machine learning (ML) based on tensor trains (TT). The estimator uses TTs to parametrize discretized functions, which are then optimized using Riemannian gradient descent under the form of a tensor completion problem. Since this optimization is sensitive to initialization, it turns out that the use of other ML estimators for initialization is crucial. This results in a competitive, fast ML estimator with lower memory usage than many other ML estimators, like the ones used for the initialization.