Factored Temporal Sigmoid Belief Networks for Sequence Learning
This work addresses sequence learning problems for domains requiring robust classification and synthesis, but it is incremental as it builds on existing TSBN methods.
The paper tackled learning temporal dependencies in multiple sequences by developing a factored Temporal Sigmoid Belief Network with a three-way weight tensor to incorporate side information, achieving state-of-the-art predictive and classification performance on sequential data.
Deep conditional generative models are developed to simultaneously learn the temporal dependencies of multiple sequences. The model is designed by introducing a three-way weight tensor to capture the multiplicative interactions between side information and sequences. The proposed model builds on the Temporal Sigmoid Belief Network (TSBN), a sequential stack of Sigmoid Belief Networks (SBNs). The transition matrices are further factored to reduce the number of parameters and improve generalization. When side information is not available, a general framework for semi-supervised learning based on the proposed model is constituted, allowing robust sequence classification. Experimental results show that the proposed approach achieves state-of-the-art predictive and classification performance on sequential data, and has the capacity to synthesize sequences, with controlled style transitioning and blending.