Pranaydeep Singh


How Language-Dependent is Emotion Detection? Evidence from Multilingual BERT
Luna De Bruyne | Pranaydeep Singh | Orphee De Clercq | Els Lefever | Veronique Hoste
Proceedings of the The 2nd Workshop on Multi-lingual Representation Learning (MRL)

As emotion analysis in text has gained a lot of attention in the field of natural language processing, differences in emotion expression across languages could have consequences for how emotion detection models work. We evaluate the language-dependence of an mBERT-based emotion detection model by comparing language identification performance before and after fine-tuning on emotion detection, and performing (adjusted) zero-shot experiments to assess whether emotion detection models rely on language-specific information. When dealing with typologically dissimilar languages, we found evidence for the language-dependence of emotion detection.

Combining Language Models and Linguistic Information to Label Entities in Memes
Pranaydeep Singh | Aaron Maladry | Els Lefever
Proceedings of the Workshop on Combating Online Hostile Posts in Regional Languages during Emergency Situations

This paper describes the system we developed for the shared task ‘Hero, Villain and Victim: Dissecting harmful memes for Semantic role labelling of entities’ organised in the framework of the Second Workshop on Combating Online Hostile Posts in Regional Languages during Emergency Situation (Constraint 2022). We present an ensemble approach combining transformer-based models and linguistic information, such as the presence of irony and implicit sentiment associated to the target named entities. The ensemble system obtains promising classification scores, resulting in a third place finish in the competition.

Investigating the Quality of Static Anchor Embeddings from Transformers for Under-Resourced Languages
Pranaydeep Singh | Orphee De Clercq | Els Lefever
Proceedings of the 1st Annual Meeting of the ELRA/ISCA Special Interest Group on Under-Resourced Languages

This paper reports on experiments for cross-lingual transfer using the anchor-based approach of Schuster et al. (2019) for English and a low-resourced language, namely Hindi. For the sake of comparison, we also evaluate the approach on three very different higher-resourced languages, viz. Dutch, Russian and Chinese. Initially designed for ELMo embeddings, we analyze the approach for the more recent BERT family of transformers for a variety of tasks, both mono and cross-lingual. The results largely prove that like most other cross-lingual transfer approaches, the static anchor approach is underwhelming for the low-resource language, while performing adequately for the higher resourced ones. We attempt to provide insights into both the quality of the anchors, and the performance for low-shot cross-lingual transfer to better understand this performance gap. We make the extracted anchors and the modified train and test sets available for future research at

When the Student Becomes the Master: Learning Better and Smaller Monolingual Models from mBERT
Pranaydeep Singh | Els Lefever
Proceedings of the 29th International Conference on Computational Linguistics

In this research, we present pilot experiments to distil monolingual models from a jointly trained model for 102 languages (mBERT). We demonstrate that it is possible for the target language to outperform the original model, even with a basic distillation setup. We evaluate our methodology for 6 languages with varying amounts of resources and language families.

SentEMO: A Multilingual Adaptive Platform for Aspect-based Sentiment and Emotion Analysis
Ellen De Geyndt | Orphee De Clercq | Cynthia Van Hee | Els Lefever | Pranaydeep Singh | Olivier Parent | Veronique Hoste
Proceedings of the 12th Workshop on Computational Approaches to Subjectivity, Sentiment & Social Media Analysis

In this paper, we present the SentEMO platform, a tool that provides aspect-based sentiment analysis and emotion detection of unstructured text data such as reviews, emails and customer care conversations. Currently, models have been trained for five domains and one general domain and are implemented in a pipeline approach, where the output of one model serves as the input for the next. The results are presented in three dashboards, allowing companies to gain more insights into what stakeholders think of their products and services. The SentEMO platform is available at


A Pilot Study for BERT Language Modelling and Morphological Analysis for Ancient and Medieval Greek
Pranaydeep Singh | Gorik Rutten | Els Lefever
Proceedings of the 5th Joint SIGHUM Workshop on Computational Linguistics for Cultural Heritage, Social Sciences, Humanities and Literature

This paper presents a pilot study to automatic linguistic preprocessing of Ancient and Byzantine Greek, and morphological analysis more specifically. To this end, a novel subword-based BERT language model was trained on the basis of a varied corpus of Modern, Ancient and Post-classical Greek texts. Consequently, the obtained BERT embeddings were incorporated to train a fine-grained Part-of-Speech tagger for Ancient and Byzantine Greek. In addition, a corpus of Greek Epigrams was manually annotated and the resulting gold standard was used to evaluate the performance of the morphological analyser on Byzantine Greek. The experimental results show very good perplexity scores (4.9) for the BERT language model and state-of-the-art performance for the fine-grained Part-of-Speech tagger for in-domain data (treebanks containing a mixture of Classical and Medieval Greek), as well as for the newly created Byzantine Greek gold standard data set. The language models and associated code are made available for use at

LT3 at SemEval-2021 Task 6: Using Multi-Modal Compact Bilinear Pooling to Combine Visual and Textual Understanding in Memes
Pranaydeep Singh | Els Lefever
Proceedings of the 15th International Workshop on Semantic Evaluation (SemEval-2021)

Internet memes have become ubiquitous in social media networks today. Due to their popularity, they are also a widely used mode of expression to spread disinformation online. As memes consist of a mixture of text and image, they require a multi-modal approach for automatic analysis. In this paper, we describe our contribution to the SemEval-2021 Detection of Persuasian Techniques in Texts and Images Task. We propose a Multi-Modal learning system, which incorporates “memebeddings”, viz. joint text and vision features by combining them with compact bilinear pooling, to automatically identify rhetorical and psychological disinformation techniques. The experimental results show that the proposed system constantly outperforms the competition’s baseline, and achieves the 2nd best Macro F1-score and 14th best Micro F1-score out of all participants.


Sentiment Analysis for Hinglish Code-mixed Tweets by means of Cross-lingual Word Embeddings
Pranaydeep Singh | Els Lefever
Proceedings of the The 4th Workshop on Computational Approaches to Code Switching

This paper investigates the use of unsupervised cross-lingual embeddings for solving the problem of code-mixed social media text understanding. We specifically investigate the use of these embeddings for a sentiment analysis task for Hinglish Tweets, viz. English combined with (transliterated) Hindi. In a first step, baseline models, initialized with monolingual embeddings obtained from large collections of tweets in English and code-mixed Hinglish, were trained. In a second step, two systems using cross-lingual embeddings were researched, being (1) a supervised classifier and (2) a transfer learning approach trained on English sentiment data and evaluated on code-mixed data. We demonstrate that incorporating cross-lingual embeddings improves the results (F1-score of 0.635 versus a monolingual baseline of 0.616), without any parallel data required to train the cross-lingual embeddings. In addition, the results show that the cross-lingual embeddings not only improve the results in a fully supervised setting, but they can also be used as a base for distant supervision, by training a sentiment model in one of the source languages and evaluating on the other language projected in the same space. The transfer learning experiments result in an F1-score of 0.556, which is almost on par with the supervised settings and speak to the robustness of the cross-lingual embeddings approach.

LT3 at SemEval-2020 Task 8: Multi-Modal Multi-Task Learning for Memotion Analysis
Pranaydeep Singh | Nina Bauwelinck | Els Lefever
Proceedings of the Fourteenth Workshop on Semantic Evaluation

Internet memes have become a very popular mode of expression on social media networks today. Their multi-modal nature, caused by a mixture of text and image, makes them a very challenging research object for automatic analysis. In this paper, we describe our contribution to the SemEval-2020 Memotion Analysis Task. We propose a Multi-Modal Multi-Task learning system, which incorporates “memebeddings”, viz. joint text and vision features, to learn and optimize for all three Memotion subtasks simultaneously. The experimental results show that the proposed system constantly outperforms the competition’s baseline, and the system setup with continual learning (where tasks are trained sequentially) obtains the best classification F1-scores.

LT3 at SemEval-2020 Task 9: Cross-lingual Embeddings for Sentiment Analysis of Hinglish Social Media Text
Pranaydeep Singh | Els Lefever
Proceedings of the Fourteenth Workshop on Semantic Evaluation

This paper describes our contribution to the SemEval-2020 Task 9 on Sentiment Analysis for Code-mixed Social Media Text. We investigated two approaches to solve the task of Hinglish sentiment analysis. The first approach uses cross-lingual embeddings resulting from projecting Hinglish and pre-trained English FastText word embeddings in the same space. The second approach incorporates pre-trained English embeddings that are incrementally retrained with a set of Hinglish tweets. The results show that the second approach performs best, with an F1-score of 70.52% on the held-out test data.

Identifying Cognates in English-Dutch and French-Dutch by means of Orthographic Information and Cross-lingual Word Embeddings
Els Lefever | Sofie Labat | Pranaydeep Singh
Proceedings of the Twelfth Language Resources and Evaluation Conference

This paper investigates the validity of combining more traditional orthographic information with cross-lingual word embeddings to identify cognate pairs in English-Dutch and French-Dutch. In a first step, lists of potential cognate pairs in English-Dutch and French-Dutch are manually labelled. The resulting gold standard is used to train and evaluate a multi-layer perceptron that can distinguish cognates from non-cognates. Fifteen orthographic features capture string similarities between source and target words, while the cosine similarity between their word embeddings represents the semantic relation between these words. By adding domain-specific information to pretrained fastText embeddings, we are able to obtain good embeddings for words that did not yet have a pretrained embedding (e.g. Dutch compound nouns). These embeddings are then aligned in a cross-lingual vector space by exploiting their structural similarity (cf. adversarial learning). Our results indicate that although the classifier already achieves good results on the basis of orthographic information, the performance further improves by including semantic information in the form of cross-lingual word embeddings.