Compression Method for Deep Diagonal State Space Model Based on $H^2$ Optimal Reduction
This addresses deployment challenges for resource-constrained devices by incrementally improving compression methods for SSM-based models.
The paper tackles the problem of large parameter sizes in deep learning models with linear state space models (SSMs) by proposing an efficient compression method using $H^2$ model order reduction from control theory, achieving a reduction to 1/32 of parameters without performance loss on the LRA benchmark.
Deep learning models incorporating linear SSMs have gained attention for capturing long-range dependencies in sequential data. However, their large parameter sizes pose challenges for deployment on resource-constrained devices. In this study, we propose an efficient parameter reduction method for these models by applying $H^{2}$ model order reduction techniques from control theory to their linear SSM components. In experiments, the LRA benchmark results show that the model compression based on our proposed method outperforms an existing method using the Balanced Truncation, while successfully reducing the number of parameters in the SSMs to $1/32$ without sacrificing the performance of the original models.