Fine-Tuning BERT Models for Intent Recognition Using a Frequency Cut-Off Strategy for Domain-Specific Vocabulary Extension Fernández Martínez, Fernando Griol Barres, David Callejas Carrión, Zoraida Topic classification Intent detection Conversational systems Recurrent networks Attentive RNN Attentive LSTM Transformer models Transfer learning The work leading to these results was supported by the Spanish Ministry of Science and Innovation through the R& D&i projects GOMINOLA (PID2020-118112RB-C21 and PID2020118112RB-C22, funded by MCIN/AEI/10.13039/501100011033), CAVIAR (TEC2017-84593-C2-1-R, funded by MCIN/ AEI/10.13039/501100011033/FEDER "Una manera de hacer Europa"), and AMICPoC (PDC2021-120846-C42, funded by MCIN/AEI/10.13039/501100011033 and by "the European Union "NextGenerationEU/PRTR"). This research also received funding from the European Union's Horizon2020 research and innovation program under grant agreement No 823907 (http://menhirproject.eu, accessed on 2 February 2022). Furthermore, R.K.'s research was supported by the Spanish Ministry of Education (FPI grant PRE2018-083225). Intent recognition is a key component of any task-oriented conversational system. The intent recognizer can be used first to classify the user’s utterance into one of several predefined classes (intents) that help to understand the user’s current goal. Then, the most adequate response can be provided accordingly. Intent recognizers also often appear as a form of joint models for performing the natural language understanding and dialog management tasks together as a single process, thus simplifying the set of problems that a conversational system must solve. This happens to be especially true for frequently asked question (FAQ) conversational systems. In this work, we first present an exploratory analysis in which different deep learning (DL) models for intent detection and classification were evaluated. In particular, we experimentally compare and analyze conventional recurrent neural networks (RNN) and state-of-the-art transformer models. Our experiments confirmed that best performance is achieved by using transformers. Specifically, best performance was achieved by fine-tuning the so-called BETO model (a Spanish pretrained bidirectional encoder representations from transformers (BERT) model from the Universidad de Chile) in our intent detection task. Then, as the main contribution of the paper, we analyze the effect of inserting unseen domain words to extend the vocabulary of the model as part of the fine-tuning or domain-adaptation process. Particularly, a very simple word frequency cut-off strategy is experimentally shown to be a suitable method for driving the vocabulary learning decisions over unseen words. The results of our analysis show that the proposed method helps to effectively extend the original vocabulary of the pretrained models. We validated our approach with a selection of the corpus acquired with the Hispabot-Covid19 system obtaining satisfactory results. 2022-03-08T07:22:40Z 2022-03-08T07:22:40Z 2022-02-03 journal article Fernández-Martínez, F... [et al.]. Fine-Tuning BERT Models for Intent Recognition Using a Frequency Cut-Off Strategy for Domain-Specific Vocabulary Extension. Appl. Sci. 2022, 12, 1610. [https://doi.org/10.3390/app12031610] http://hdl.handle.net/10481/73197 10.3390/app12031610 eng info:eu-repo/grantAgreement/EC/H2020/823907 http://creativecommons.org/licenses/by/3.0/es/ open access Atribución 3.0 España MDPI