[{"doi":"10.1109/ijcnn.2019.8851954","publication_identifier":{"isbn":["9781728119854"]},"user_id":"2DF688A6-F248-11E8-B48F-1D18A9856A87","date_updated":"2021-01-12T08:11:19Z","language":[{"iso":"eng"}],"scopus_import":1,"external_id":{"arxiv":["1809.03864"]},"conference":{"location":"Budapest, Hungary","end_date":"2019-07-19","start_date":"2019-07-14","name":"IJCNN: International Joint Conference on Neural Networks"},"title":"Response characterization for auditing cell dynamics in long short-term memory networks","citation":{"short":"R. Hasani, A. Amini, M. Lechner, F. Naser, R. Grosu, D. Rus, in:, Proceedings of the International Joint Conference on Neural Networks, IEEE, 2019.","ama":"Hasani R, Amini A, Lechner M, Naser F, Grosu R, Rus D. Response characterization for auditing cell dynamics in long short-term memory networks. In: <i>Proceedings of the International Joint Conference on Neural Networks</i>. IEEE; 2019. doi:<a href=\"https://doi.org/10.1109/ijcnn.2019.8851954\">10.1109/ijcnn.2019.8851954</a>","apa":"Hasani, R., Amini, A., Lechner, M., Naser, F., Grosu, R., &#38; Rus, D. (2019). Response characterization for auditing cell dynamics in long short-term memory networks. In <i>Proceedings of the International Joint Conference on Neural Networks</i>. Budapest, Hungary: IEEE. <a href=\"https://doi.org/10.1109/ijcnn.2019.8851954\">https://doi.org/10.1109/ijcnn.2019.8851954</a>","chicago":"Hasani, Ramin, Alexander Amini, Mathias Lechner, Felix Naser, Radu Grosu, and Daniela Rus. “Response Characterization for Auditing Cell Dynamics in Long Short-Term Memory Networks.” In <i>Proceedings of the International Joint Conference on Neural Networks</i>. IEEE, 2019. <a href=\"https://doi.org/10.1109/ijcnn.2019.8851954\">https://doi.org/10.1109/ijcnn.2019.8851954</a>.","ieee":"R. Hasani, A. Amini, M. Lechner, F. Naser, R. Grosu, and D. Rus, “Response characterization for auditing cell dynamics in long short-term memory networks,” in <i>Proceedings of the International Joint Conference on Neural Networks</i>, Budapest, Hungary, 2019.","ista":"Hasani R, Amini A, Lechner M, Naser F, Grosu R, Rus D. 2019. Response characterization for auditing cell dynamics in long short-term memory networks. Proceedings of the International Joint Conference on Neural Networks. IJCNN: International Joint Conference on Neural Networks, 8851954.","mla":"Hasani, Ramin, et al. “Response Characterization for Auditing Cell Dynamics in Long Short-Term Memory Networks.” <i>Proceedings of the International Joint Conference on Neural Networks</i>, 8851954, IEEE, 2019, doi:<a href=\"https://doi.org/10.1109/ijcnn.2019.8851954\">10.1109/ijcnn.2019.8851954</a>."},"oa_version":"Preprint","year":"2019","type":"conference","author":[{"last_name":"Hasani","first_name":"Ramin","full_name":"Hasani, Ramin"},{"last_name":"Amini","full_name":"Amini, Alexander","first_name":"Alexander"},{"last_name":"Lechner","full_name":"Lechner, Mathias","first_name":"Mathias","id":"3DC22916-F248-11E8-B48F-1D18A9856A87"},{"full_name":"Naser, Felix","first_name":"Felix","last_name":"Naser"},{"first_name":"Radu","full_name":"Grosu, Radu","last_name":"Grosu"},{"last_name":"Rus","full_name":"Rus, Daniela","first_name":"Daniela"}],"day":"30","main_file_link":[{"open_access":"1","url":"https://arxiv.org/abs/1809.03864"}],"oa":1,"publisher":"IEEE","publication_status":"published","status":"public","department":[{"_id":"ToHe"}],"quality_controlled":"1","publication":"Proceedings of the International Joint Conference on Neural Networks","arxiv":1,"_id":"6985","date_created":"2019-11-04T15:59:58Z","abstract":[{"lang":"eng","text":"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."}],"date_published":"2019-09-30T00:00:00Z","article_number":"8851954","month":"09"}]