@inproceedings{6985,
  abstract     = {In this paper, we introduce a novel method to interpret recurrent neural networks (RNNs), particularly long short-term memory networks (LSTMs) at the cellular level. We propose a systematic pipeline for interpreting individual hidden state dynamics within the network using response characterization methods. The ranked contribution of individual cells to the network's output is computed by analyzing a set of interpretable metrics of their decoupled step and sinusoidal responses. As a result, our method is able to uniquely identify neurons with insightful dynamics, quantify relationships between dynamical properties and test accuracy through ablation analysis, and interpret the impact of network capacity on a network's dynamical distribution. Finally, we demonstrate the generalizability and scalability of our method by evaluating a series of different benchmark sequential datasets.},
  author       = {Hasani, Ramin and Amini, Alexander and Lechner, Mathias and Naser, Felix and Grosu, Radu and Rus, Daniela},
  booktitle    = {Proceedings of the International Joint Conference on Neural Networks},
  isbn         = {9781728119854},
  location     = {Budapest, Hungary},
  publisher    = {IEEE},
  title        = {{Response characterization for auditing cell dynamics in long short-term memory networks}},
  doi          = {10.1109/ijcnn.2019.8851954},
  year         = {2019},
}