Fetching the paper…

Seq2Tens: An Efficient Representation of Sequences by Low-Rank Tensor Projections · Around