{"abstract":[{"text":"Connectionist temporal classification (CTC) is a sequence-level loss that has been successfully applied to train recurrent neural network (RNN) models for automatic speech recognition. However, one major weakness of CTC is the conditional independence assumption that makes it difficult for the model to learn label dependencies. In this paper, we propose stimulated CTC, which uses stimulated learning to help CTC models learn label dependencies implicitly by using an auxiliary RNN to generate the appropriate stimuli. This stimuli comes in the form of an additional stimulation loss term which encourages the model to learn said label dependencies. The auxiliary network is only used during training and the inference model has the same structure as a standard CTC model. The proposed stimulated CTC model achieves about 35% relative character error rate improvements on a synthetic gesture keyboard recognition task and over 30% relative word error rate improvements on the Librispeech automatic speech recognition tasks over a baseline model trained with CTC only.","lang":"eng"}],"file_date_updated":"2020-02-06T07:24:26Z","oa":"1","status":"public","title":"Improving CTC Using Stimulated Learning for Sequence Modeling","_id":"15812","ddc":["000"],"year":"2019","type":"conference","publication":"ICASSP 2019, Brighton, UK","department":[{"_id":"54"}],"citation":{"short":"J. Heymann, B.L. Khe Chai Sim, in: ICASSP 2019, Brighton, UK, 2019.","apa":"Heymann, J., &#38; Khe Chai Sim, B. L. (2019). Improving CTC Using Stimulated Learning for Sequence Modeling. In <i>ICASSP 2019, Brighton, UK</i>.","mla":"Heymann, Jahn, and Bo Li Khe Chai Sim. “Improving CTC Using Stimulated Learning for Sequence Modeling.” <i>ICASSP 2019, Brighton, UK</i>, 2019.","ieee":"J. Heymann and B. L. Khe Chai Sim, “Improving CTC Using Stimulated Learning for Sequence Modeling,” in <i>ICASSP 2019, Brighton, UK</i>, 2019.","chicago":"Heymann, Jahn, and Bo Li Khe Chai Sim. “Improving CTC Using Stimulated Learning for Sequence Modeling.” In <i>ICASSP 2019, Brighton, UK</i>, 2019.","ama":"Heymann J, Khe Chai Sim BL. Improving CTC Using Stimulated Learning for Sequence Modeling. In: <i>ICASSP 2019, Brighton, UK</i>. ; 2019.","bibtex":"@inproceedings{Heymann_Khe Chai Sim_2019, title={Improving CTC Using Stimulated Learning for Sequence Modeling}, booktitle={ICASSP 2019, Brighton, UK}, author={Heymann, Jahn and Khe Chai Sim, Bo Li}, year={2019} }"},"author":[{"full_name":"Heymann, Jahn","last_name":"Heymann","first_name":"Jahn"},{"last_name":"Khe Chai Sim","full_name":"Khe Chai Sim, Bo Li","first_name":"Bo Li"}],"user_id":"59789","language":[{"iso":"eng"}],"date_created":"2020-02-06T07:22:47Z","date_updated":"2022-01-06T06:52:35Z","project":[{"name":"Computing Resources Provided by the Paderborn Center for Parallel Computing","_id":"52"}],"has_accepted_license":"1","file":[{"date_created":"2020-02-06T07:24:26Z","creator":"huesera","file_size":239665,"access_level":"open_access","file_id":"15813","date_updated":"2020-02-06T07:24:26Z","relation":"main_file","file_name":"ICASSP_2019_Heymann_1_Paper.pdf","content_type":"application/pdf"}]}