Bias is pervasive for NLP models, motivating the development of automatic debiasing techniques. Evaluation of NLP debiasing methods has largely been limited to binary attributes in isolation, e.g., debiasing with respect to binary gender or race, how
ever many corpora involve multiple such attributes, possibly with higher cardinality. In this paper we argue that a truly fair model must consider gerrymandering' groups which comprise not only single attributes, but also intersectional groups. We evaluate a form of bias-constrained model which is new to NLP, as well an extension of the iterative nullspace projection technique which can handle multiple identities.
Existing techniques for mitigating dataset bias often leverage a biased model to identify biased instances. The role of these biased instances is then reduced during the training of the main model to enhance its robustness to out-of-distribution data
. A common core assumption of these techniques is that the main model handles biased instances similarly to the biased model, in that it will resort to biases whenever available. In this paper, we show that this assumption does not hold in general. We carry out a critical investigation on two well-known datasets in the domain, MNLI and FEVER, along with two biased instance detection methods, partial-input and limited-capacity models. Our experiments show that in around a third to a half of instances, the biased model is unable to predict the main model's behavior, highlighted by the significantly different parts of the input on which they base their decisions. Based on a manual validation, we also show that this estimate is highly in line with human interpretation. Our findings suggest that down-weighting of instances detected by bias detection methods, which is a widely-practiced procedure, is an unnecessary waste of training data. We release our code to facilitate reproducibility and future research.
We describe our two NMT systems submitted to the WMT2021 shared task in English-Czech news translation: CUNI-DocTransformer (document-level CUBBITT) and CUNI-Marian-Baselines. We improve the former with a better sentence-segmentation pre-processing a
nd a post-processing for fixing errors in numbers and units. We use the latter for experiments with various backtranslation techniques.
Written communication is of utmost importance to the progress of scientific research. The speed of such development, however, may be affected by the scarcity of reviewers to referee the quality of research articles. In this context, automatic approac
hes that are able to query linguistic segments in written contributions by detecting the presence or absence of common rhetorical patterns have become a necessity. This paper aims to compare supervised machine learning techniques tested to accomplish genre analysis in Introduction sections of software engineering articles. A semi-supervised approach was carried out to augment the number of annotated sentences in SciSents (Avaliable on: ANONYMOUS). Two supervised approaches using SVM and logistic regression were undertaken to assess the F-score for genre analysis in the corpus. A technique based on logistic regression and BERT has been found to perform genre analysis highly satisfactorily with an average of 88.25 on F-score when retrieving patterns at an overall level.
The following system description presents our approach to the detection of persuasion techniques in texts and images. The given task has been framed as a multi-label classification problem with the different techniques serving as class labels. The mu
lti-label classification problem is one in which a list of target variables such as our class labels is associated with every input chunk and assumes that a document can simultaneously and independently be assigned to multiple labels or classes. In order to assign class labels to the given memes, we opted for RoBERTa (A Robustly Optimized BERT Pretraining Approach) as a neural network architecture for token and sequence classification. Starting off with a pre-trained model for language representation we fine-tuned this model on the given classification task with the provided annotated data in supervised training steps. To incorporate image features in the multi-modal setting, we rely on the pre-trained VGG-16 model architecture.
Sentence weighting is a simple and powerful domain adaptation technique. We carry out domain classification for computing sentence weights with 1) language model cross entropy difference 2) a convolutional neural network 3) a Recursive Neural Tensor
Network. We compare these approaches with regard to domain classification accuracy and and study the posterior probability distributions. Then we carry out NMT experiments in the scenario where we have no in-domain parallel corpora and and only very limited in-domain monolingual corpora. Here and we use the domain classifier to reweight the sentences of our out-of-domain training corpus. This leads to improvements of up to 2.1 BLEU for German to English translation.
This paper describes and examines different systems to address Task 6 of SemEval-2021: Detection of Persuasion Techniques In Texts And Images, Subtask 1. The task aims to build a model for identifying rhetorical and psycho- logical techniques (such a
s causal oversimplification, name-calling, smear) in the textual content of a meme which is often used in a disinformation campaign to influence the users. The paper provides an extensive comparison among various machine learning systems as a solution to the task. We elaborate on the pre-processing of the text data in favor of the task and present ways to overcome the class imbalance. The results show that fine-tuning a RoBERTa model gave the best results with an F1-Micro score of 0.51 on the development set.
We developed a system for task 6 sub-task 1 for detecting propaganda in memes. An external dataset and augmentation data-set were used to extend the official competition data-set. Data augmentation techniques were applied on the external data-set and
competition data-set to come up with the augmented data-set. We trained 5 transformers (DeBERTa, and 4 RoBERTa) and ensembled them to make the prediction. We trained 1 RoBERTa model initially on the augmented data-set for a few epochs and then fine-tuned it on the competition data-set which improved the f1-micro up to 0.1 scores. After that, another initial RoBERTa model was trained on the external data-set merged with the augmented data-set for few epochs and fine-tuned it on the competition data-set. Furthermore, we ensembled the initial models with the models after fine-tuning. For the final model in the ensemble, we trained a DeBERTa model on the augmented data-set without fine-tuning it on the competition data-set. Finally, we averaged the output of each model in the ensemble to make the prediction.
Inscribing persuasion techniques in memes is the most impactful way to influence peoples' mindsets. People are more inclined to memes as they are more stimulating and convincing and hence memes are often exploited by tactfully engraving propaganda in
its context with the intent of attaining specific agenda. This paper describes our participation in the three subtasks featured by SemEval 2021 task 6 on the detection of persuasion techniques in texts and images. We utilize a fusion of logistic regression, decision tree, and fine-tuned DistilBERT for tackling subtask 1. As for subtask 2, we propose a system that consolidates a span identification model and a multi-label classification model based on pre-trained BERT. We address the multi-modal multi-label classification of memes defined in subtask 3 by utilizing a ResNet50 based image model, DistilBERT based text model, and a multi-modal architecture based on multikernel CNN+LSTM and MLP model. The outcomes illustrated the competitive performance of our systems.
The objective of subtask 2 of SemEval-2021 Task 6 is to identify techniques used together with the span(s) of text covered by each technique. This paper describes the system and model we developed for the task. We first propose a pipeline system to i
dentify spans, then to classify the technique in the input sequence. But it severely suffers from handling the overlapping in nested span. Then we propose to formulize the task as a question answering task by MRC framework which achieves a better result compared to the pipeline method. Moreover, data augmentation and loss design techniques are also explored to alleviate the problem of data sparse and imbalance. Finally, we attain the 3rd place in the final evaluation phase.
