Monolingual and Cross-Lingual Intent Detection without Training Data in Target Languages

Abstract

Due to recent DNN advancements, many NLP problems can be effectively solved using transformer-based models and supervised data. Unfortunately, such data is not available in some languages. This research is based on assumptions that (1) training data can be obtained by the machine translating it from another language; (2) there are cross-lingual solutions that work without the training data in the target language. Consequently, in this research, we use the English dataset and solve the intent detection problem for five target languages (German, French, Lithuanian, Latvian, and Portuguese). When seeking the most accurate solutions, we investigate BERT-based word and sentence transformers together with eager learning classifiers (CNN, BERT fine-tuning, FFNN) and lazy learning approach (Cosine similarity as the memory-based method). We offer and evaluate several strategies to overcome the data scarcity problem with machine translation, cross-lingual models, and a combination of the previous two. The experimental investigation revealed the robustness of sentence transformers under various cross-lingual conditions. The accuracy equal to ~0.842 is achieved with the English dataset with completely monolingual models is considered our top-line. However, cross-lingual approaches demonstrate similar accuracy levels reaching ~0.831, ~0.829, ~0.853, ~0.831, and ~0.813 on German, French, Lithuanian, Latvian, and Portuguese languages.

Towards Media Monitoring: Detecting Known and Emerging Topics through Multilingual and Crosslingual Text Classification

“…Such an approach allows for leveraging high-quality, abundant resources available in English to benefit less-resourced languages. However, it is important to acknowledge that crosslingual models often experience a drop in accuracy compared to their monolingual counterparts, a phenomenon well documented in the literature [20]. This accuracy dip is partly due to the intrinsic variability and complexity of language translation and contextual usage across languages.…”

A comparative analysis of text representation, classification and clustering methods over real project proposals

“…This study presents the application of word embedding methods to effectively use rich information in the abstract and title of project proposals in Turkish. Neural network-based word embedding methods (e.g., FastText, BERT) are used in numerous NLP tasks successfully (Romanov and Khusainova, 2019; Kapočiūtė-Dzikienė et al ., 2021; Kalyan et al ., 2021). However, a project proposal grouping research based on such next-generation representation approaches has not yet been reported to the best of our knowledge.…”

“…Neural network technologies for word embedding have recently shown remarkable results in NLP tasks (Romanov and Khusainova, 2019; Kapočiūtė-Dzikienė et al ., 2021; Kalyan et al ., 2021). However, a project proposal grouping study based on high-performance neural network-based textual feature extraction techniques has not yet been reported to the best of our knowledge.…”

Triple Channel Feature Fusion Few-Shot Intent Recognition With Orthogonality Constrained Multi-Head Attention

“…Aiming at the problems of high computational cost and poor generalization ability of traditional intention recognition models, the BERT-FNN intention recognition model was proposed by Zheng et al [10]. Two BERT-based vectorization types, word embedding and sentence embedding, were used by Jurgita et al [11] to solve the problem of cross-domain intent recognition without training data. The performance of the model is improved to a great extent through the above channel fusion technique.…”