Maria Mihaela Trusca

Vladimir Araujo, Maria Mihaela Trusca, Rodrigo Tufiño et al.

In recent years, significant advancements in pre-trained language models have driven the creation of numerous non-English language variants, with a particular emphasis on encoder-only and decoder-only architectures. While Spanish language models based on BERT and GPT have demonstrated proficiency in natural language understanding and generation, there remains a noticeable scarcity of encoder-decoder models explicitly designed for sequence-to-sequence tasks, which aim to map input sequences to generate output sequences conditionally. This paper breaks new ground by introducing the implementation and evaluation of renowned encoder-decoder architectures exclusively pre-trained on Spanish corpora. Specifically, we present Spanish versions of BART, T5, and BERT2BERT-style models and subject them to a comprehensive assessment across various sequence-to-sequence tasks, including summarization, question answering, split-and-rephrase, dialogue, and translation. Our findings underscore the competitive performance of all models, with the BART- and T5-based models emerging as top performers across all tasks. We have made all models publicly available to the research community to foster future explorations and advancements in Spanish NLP: https://github.com/vgaraujov/Seq2Seq-Spanish-PLMs.

Maria Mihaela Trusca, Wolf Nuyts, Jonathan Thomm et al.

Current diffusion models create photorealistic images given a text prompt as input but struggle to correctly bind attributes mentioned in the text to the right objects in the image. This is evidenced by our novel image-graph alignment model called EPViT (Edge Prediction Vision Transformer) for the evaluation of image-text alignment. To alleviate the above problem, we propose focused cross-attention (FCA) that controls the visual attention maps by syntactic constraints found in the input sentence. Additionally, the syntax structure of the prompt helps to disentangle the multimodal CLIP embeddings that are commonly used in T2I generation. The resulting DisCLIP embeddings and FCA are easily integrated in state-of-the-art diffusion models without additional training of these models. We show substantial improvements in T2I generation and especially its attribute-object binding on several datasets.\footnote{Code and data will be made available upon acceptance.

Maria Mihaela Trusca, Mingxiao Li, Marie-Francine Moens

Text-based image editing is typically approached as a static task that involves operations such as inserting, deleting, or modifying elements of an input image based on human instructions. Given the static nature of this task, in this paper, we aim to make this task dynamic by incorporating actions. By doing this, we intend to modify the positions or postures of objects in the image to depict different actions while maintaining the visual properties of the objects. To implement this challenging task, we propose a new model that is sensitive to action text instructions by learning to recognize contrastive action discrepancies. The model training is done on new datasets defined by extracting frames from videos that show the visual scenes before and after an action. We show substantial improvements in image editing using action-based text instructions and high reasoning capabilities that allow our model to use the input image as a starting scene for an action while generating a new image that shows the final scene of the action.

Maria Mihaela Trusca, Liesbeth Allein

Interpretations of a single sentence can vary, particularly when its context is lost. This paper aims to simulate how readers perceive content with varying toxicity levels by generating diverse interpretations of out-of-context sentences. By modeling toxicity, we can anticipate misunderstandings and reveal hidden toxic meanings. Our proposed decoding strategy explicitly controls toxicity in the set of generated interpretations by (i) aligning interpretation toxicity with the input, (ii) relaxing toxicity constraints for more toxic input sentences, and (iii) promoting diversity in toxicity levels within the set of generated interpretations. Experimental results show that our method improves alignment with human-written interpretations in both syntax and semantics while reducing model prediction uncertainty.

Maria Mihaela Trusca, Tinne Tuytelaars, Marie-Francine Moens

Text-based semantic image editing assumes the manipulation of an image using a natural language instruction. Although recent works are capable of generating creative and qualitative images, the problem is still mostly approached as a black box sensitive to generating unexpected outputs. Therefore, we propose a novel model to enhance the text-based control of an image editor by explicitly reasoning about which parts of the image to alter or preserve. It relies on word alignments between a description of the original source image and the instruction that reflects the needed updates, and the input image. The proposed Diffusion Masking with word Alignments (DM-Align) allows the editing of an image in a transparent and explainable way. It is evaluated on a subset of the Bison dataset and a self-defined dataset dubbed Dream. When comparing to state-of-the-art baselines, quantitative and qualitative results show that DM-Align has superior performance in image editing conditioned on language instructions, well preserves the background of the image and can better cope with long text instructions.

Ron Hochstenbach, Flavius Frasincar, Maria Mihaela Trusca

The increasing popularity of the Web has subsequently increased the abundance of reviews on products and services. Mining these reviews for expressed sentiment is beneficial for both companies and consumers, as quality can be improved based on this information. In this paper, we consider the state-of-the-art HAABSA++ algorithm for aspect-based sentiment analysis tasked with identifying the sentiment expressed towards a given aspect in review sentences. Specifically, we train the neural network part of this algorithm using an adversarial network, a novel machine learning training method where a generator network tries to fool the classifier network by generating highly realistic new samples, as such increasing robustness. This method, as of yet never in its classical form applied to aspect-based sentiment analysis, is found to be able to considerably improve the out-of-sample accuracy of HAABSA++: for the SemEval 2015 dataset, accuracy was increased from 81.7% to 82.5%, and for the SemEval 2016 task, accuracy increased from 84.4% to 87.3%.

Lisa Meijer, Flavius Frasincar, Maria Mihaela Trusca

Many high performance machine learning models for Aspect-Based Sentiment Classification (ABSC) produce black box models, and therefore barely explain how they classify a certain sentiment value towards an aspect. In this paper, we propose explanation models, that inspect the internal dynamics of a state-of-the-art neural attention model, the LCR-Rot-hop, by using a technique called Diagnostic Classification. Our diagnostic classifier is a simple neural network, which evaluates whether the internal layers of the LCR-Rot-hop model encode useful word information for classification, i.e., the part of speech, the sentiment value, the presence of aspect relation, and the aspect-related sentiment value of words. We conclude that the lower layers in the LCR-Rot-hop model encode the part of speech and the sentiment value, whereas the higher layers represent the presence of a relation with the aspect and the aspect-related sentiment value of words.

Tomas Liesting, Flavius Frasincar, Maria Mihaela Trusca

Data augmentation is a way to increase the diversity of available data by applying constrained transformations on the original data. This strategy has been widely used in image classification but has to the best of our knowledge not yet been used in aspect-based sentiment analysis (ABSA). ABSA is a text analysis technique that determines aspects and their associated sentiment in opinionated text. In this paper, we investigate the effect of data augmentation on a state-of-the-art hybrid approach for aspect-based sentiment analysis (HAABSA). We apply modified versions of easy data augmentation (EDA), backtranslation, and word mixup. We evaluate the proposed techniques on the SemEval 2015 and SemEval 2016 datasets. The best result is obtained with the adjusted version of EDA, which yields a 0.5 percentage point improvement on the SemEval 2016 dataset and 1 percentage point increase on the SemEval 2015 dataset compared to the original HAABSA model.

Gino V. H. Mangnoesing, Maria Mihaela Trusca, Flavius Frasincar

Online reviews are an important source of feedback for understanding customers. In this study, we follow novel approaches that target this absence of actionable insights by classifying reviews as defect reports and requests for improvement. Unlike traditional classification methods based on expert rules, we reduce the manual labour by employing a supervised system that is capable of learning lexico-semantic patterns through genetic programming. Additionally, we experiment with a distantly-supervised SVM that makes use of noisy labels generated by patterns. Using a real-world dataset of app reviews, we show that the automatically learned patterns outperform the manually created ones, to be generated. Also the distantly-supervised SVM models are not far behind the pattern-based solutions, showing the usefulness of this approach when the amount of annotated data is limited.

Maria Mihaela Trusca, Daan Wassenberg, Flavius Frasincar et al.

The Web has become the main platform where people express their opinions about entities of interest and their associated aspects. Aspect-Based Sentiment Analysis (ABSA) aims to automatically compute the sentiment towards these aspects from opinionated text. In this paper we extend the state-of-the-art Hybrid Approach for Aspect-Based Sentiment Analysis (HAABSA) method in two directions. First we replace the non-contextual word embeddings with deep contextual word embeddings in order to better cope with the word semantics in a given text. Second, we use hierarchical attention by adding an extra attention layer to the HAABSA high-level representations in order to increase the method flexibility in modeling the input data. Using two standard datasets (SemEval 2015 and SemEval 2016) we show that the proposed extensions improve the accuracy of the built model for ABSA.

Maria Mihaela Trusca, Gerasimos Spanakis

The tiled convolutional neural network (tiled CNN) has been applied only to computer vision for learning invariances. We adjust its architecture to NLP to improve the extraction of the most salient features for sentiment analysis. Knowing that the major drawback of the tiled CNN in the NLP field is its inflexible filter structure, we propose a novel architecture called hybrid tiled CNN that applies a filter only on the words that appear in the similar contexts and on their neighbor words (a necessary step for preventing the loss of some n-grams). The experiments on the datasets of IMDB movie reviews and SemEval 2017 demonstrate the efficiency of the hybrid tiled CNN that performs better than both CNN and tiled CNN.