According to the psychological literature, implicit motives allow for the characterization of behavior, subsequent success, and long-term development. Contrary to personality traits, implicit motives are often deemed to be rather stable personality characteristics. Normally, implicit motives are obtained by Operant Motives, unconscious intrinsic desires measured by the Operant Motive Test (OMT). The OMT test requires participants to write freely descriptions associated with a set of provided images and questions. In this work, we explore different recent machine learning techniques and various text representation techniques for facing the problem of the OMT classification task. We focused on advanced language representations (e.g, BERT, XLM, and DistilBERT) and deep Supervised Autoencoders for solving the OMT task. We performed an exhaustive analysis and compared their performance against fully connected neural networks and traditional support vector classifiers. Our comparative study highlights the importance of BERT which outperforms the traditional machine learning techniques by a relative improvement of 7.9%. In addition, we performed an analysis of how the BERT attention mechanism is being modified. Our findings indicate that the writing style features acquire higher importance at the moment of accurately identifying the different OMT categories. This is the first time that a study to determine the performance of different transformer-based architectures in the OMT task is performed. Similarly, our work propose, for the first time, using deep supervised autoencoders in the OMT classification task. Our experiments demonstrate that transformer-based methods exhibit the best empirical results, obtaining a relative improvement of 7.9% over the competitive baseline suggested as part of the GermEval 2020 challenge. Additionally, we show that features associated with the writing style are more important than content-based words. Some of these findings show strong connections to previously reported behavioral research on the implicit psychometrics theory.

This paper describes our participation in the shared evaluation campaign of MexA3T 2020. Our main goal was to evaluate a Supervised Autoencoder (SAE) learning algorithm in text classification tasks. For our experiments, we used three different sets of features as inputs, namely classic word n-grams, char n-grams, and Spanish BERT encodings. Our results indicate that SAE is adequate for longer and more formal written texts. Accordingly, our approach obtained the best performance (F = 85.66%) in the fake-news classification task.

In this paper, we describe the participation of the Idiap Research Institute at GermEval 2020 shared task on the Classification and Regression of Cognitive and Motivational style from Text, specifically on subtask 2, Classification of the Operant Motive Test (OMT). Generally speaking, GermEval 2020 aims at encouraging the Natural Language Understanding (NLU) research community in proposing novel methodologies for assessing the connection between freely written texts and its cognitive and motivational styles. For evaluating this task, organizers provided a large dataset containing textual descriptions, in German language, generated by more than 14,000 participants. Our participation aims at evaluating the impact of advanced language representation, e.g., Bert, XLM, and DistilBERT in combination with some traditional machine learning algorithms. Our best configuration was able to obtain an F1 macro of 69.8% on the test partition, which represents a relative improvement of 7.4% in comparison to the proposed baseline.

Language detection is considered a difficult task especially for similar languages, varieties, and dialects. With the growing number of on-line content in different languages, the need for reliable and robust language detection tools also increased. In this work, we use supervised autoencoders with a bayesian optimizer for language detection and highlight its effi-ciency in detecting similar languages with dialect variance in comparison to other state-of-the-art techniques. We evaluated our approach on multiple datasets (Ling10, Discriminating between Similar Language (DSL), and Indo-Aryan Language Identification (ILI)). Obtained results demonstrate that SAE is highly effective in detecting languages, up to a 100% accuracy in the Ling10. Similarly, we obtain a competitive performance in identifying similar languages, and dialects, 92%, and 85% for DSL and ILI datasets respectively.