Papers by Orphee De Clercq
An Exploratory Study into Automated Précis Grading (2020.lrec-1)
Copied to clipboard
| Challenge: | Existing systems for writing evaluation focus on argumentative texts . a précis is a written text that provides a coherent summary of main points . |
| Approach: | They propose to use a corpus of English précis texts to train a machine learning model . they find it is able to predict the grade of précis texts with only a moderate error margin . |
| Outcome: | The proposed model predicts the grade of précis texts with only a moderate error margin. |
Aspect-Based Emotion Analysis and Multimodal Coreference: A Case Study of Customer Comments on Adidas Instagram Posts (2022.lrec-1)
Copied to clipboard
| Challenge: | Aspect-based sentiment analysis of user-generated content has been relatively unexplored in recent years. |
| Approach: | They present a multimodal dataset for Aspect-Based Emotion Analysis (ABEA) they take the first steps in investigating the utility of multimodal coreference resolution in an ABEA framework. |
| Outcome: | The proposed dataset consists of 4,900 comments on 175 images and is annotated with aspect and emotion categories and the emotional dimensions of valence and arousal. |
An Emotional Mess! Deciding on a Framework for Building a Dutch Emotion-Annotated Corpus (2020.lrec-1)
Copied to clipboard
| Challenge: | Existing frameworks for emotion recognition are limited and do not allow for categorical versus dimensional oppositions. |
| Approach: | They propose to use the emotions joy, love, anger, sadness and fear as well as dimensional models to annotate texts from different domains and topics. |
| Outcome: | The proposed frameworks are well-suited to annotate texts from different domains and topics, but the connotation of the labels strongly depends on the origin of the texts. |
Benefits of Data Augmentation for NMT-based Text Normalization of User-Generated Content (D19-55)
Copied to clipboard
| Challenge: | Social media texts are considered important language resources for several NLP tasks, but their use of non-standard words makes it difficult to process and analyze UGC. |
| Approach: | They propose to use a Neural Machine Translation approach to normalize lexical variants to their canonical forms to overcome performance drop in UGC. |
| Outcome: | The proposed approach overcomes a data bottleneck in Dutch, a low-resource language. |
EnerGIZAr: Leveraging GIZA++ for Effective Tokenizer Initialization (2025.findings-acl)
Copied to clipboard
| Challenge: | Continual pre-training has long been considered the default strategy for adapting models to non-English languages, but struggles with initializing new embeddings, especially for non-Latin scripts. |
| Approach: | They propose a method that leverages statistical word alignment techniques to improve continual pre-training by leveraging word alignment matrix between source and target tokens. |
| Outcome: | The proposed method outperforms existing methods on key NLP tasks including POS tagging, Sentiment Analysis, NLI, and NER in Hindi, Basque, Arabic and Korean. |
Lost in Activations: A Neuron-level Analysis of Encoders for Cross-Lingual Emotion Detection (2026.eacl-short)
Copied to clipboard
| Challenge: | XLM-R models for multilingual emotion classification are still lacking in understanding of their internal decision-making mechanisms. |
| Approach: | They propose to use neuron-level activation analysis to study the inter-language differences between neurons. |
| Outcome: | The proposed model consistently encodes emotion-related concepts across languages, but others show strong monolingual specialization. |