This paper presents the models submitted by Ghmerti team for subtasks A and B of the OffensEval shared task at SemEval 2019. OffensEval addresses the problem of identifying and categorizing offensive language in social media in three subtasks; whether or not a content is offensive (subtask A), whether it is targeted (subtask B) towards an individual, a group, or other entities (subtask C). The proposed approach includes character-level Convolutional Neural Network, word-level Recurrent Neural Network, and some preprocessing. The performance achieved by the proposed model for subtask A is 77.93% macro-averaged F1-score.
Communicating through social platforms has become one of the principal means of personal communications and interactions. Unfortunately, healthy communication is often interfered by offensive language that can have damaging effects on the users. A key to fight offensive language on social media is the existence of an automatic offensive language detection system. This paper presents the results and the main findings of SemEval-2020, Task 12 OffensEval Sub-task A Zampieri et al. (2020), on Identifying and categorising Offensive Language in Social Media. The task was based on the Arabic OffensEval dataset Mubarak et al. (2020). In this paper, we describe the system submitted by WideBot AI Lab for the shared task which ranked 10th out of 52 participants with Macro-F1 86.9% on the golden dataset under CodaLab username "yasserotiefy". We experimented with various models and the best model is a linear SVM in which we use a combination of both character and word n-grams. We also introduced a neural network approach that enhanced the predictive ability of our system that includes CNN, highway network, Bi-LSTM, and attention layers.
SemEval-2020 Task 12 was OffenseEval: Multilingual Offensive Language Identification in Social Media (Zampieri et al., 2020). The task was subdivided into multiple languages and datasets were provided for each one. The task was further divided into three sub-tasks: offensive language identification, automatic categorization of offense types, and offense target identification. I have participated in the task-C, that is, offense target identification. For preparing the proposed system, I have made use of Deep Learning networks like LSTMs and frameworks like Keras which combine the bag of words model with automatically generated sequence based features and manually extracted features from the given dataset. My system on training on 25% of the whole dataset achieves macro averaged f1 score of 47.763%.
This article introduces the Wanca 2017 corpus of texts crawled from the internet from which the sentences in rare Uralic languages for the use of the Uralic Language Identification (ULI) 2020 shared task were collected. We describe the ULI dataset and how it was constructed using the Wanca 2017 corpus and texts in different languages from the Leipzig corpora collection. We also provide baseline language identification experiments conducted using the ULI 2020 dataset.
In this paper, we present the results of the SemEval-2020 Task 9 on Sentiment Analysis of Code-Mixed Tweets (SentiMix 2020). We also release and describe our Hinglish (Hindi-English) and Spanglish (Spanish-English) corpora annotated with word-level language identification and sentence-level sentiment labels. These corpora are comprised of 20K and 19K examples, respectively. The sentiment labels are - Positive, Negative, and Neutral. SentiMix attracted 89 submissions in total including 61 teams that participated in the Hinglish contest and 28 submitted systems to the Spanglish competition. The best performance achieved was 75.0% F1 score for Hinglish and 80.6% F1 for Spanglish. We observe that BERT-like models and ensemble methods are the most common and successful approaches among the participants.
This paper describes our approach to the task of identifying offensive languages in a multilingual setting. We investigate two data augmentation strategies: using additional semi-supervised labels with different thresholds and cross-lingual transfer with data selection. Leveraging the semi-supervised dataset resulted in performance improvements compared to the baseline trained solely with the manually-annotated dataset. We propose a new metric, Translation Embedding Distance, to measure the transferability of instances for cross-lingual data selection. We also introduce various preprocessing steps tailored for social media text along with methods to fine-tune the pre-trained multilingual BERT (mBERT) for offensive language identification. Our multilingual systems achieved competitive results in Greek, Danish, and Turkish at OffensEval 2020.
This paper presents the different models submitted by the LT@Helsinki team for the SemEval 2020 Shared Task 12. Our team participated in sub-tasks A and C; titled offensive language identification and offense target identification, respectively. In both cases we used the so-called Bidirectional Encoder Representation from Transformer (BERT), a model pre-trained by Google and fine-tuned by us on the OLID and SOLID datasets. The results show that offensive tweet classification is one of several language-based tasks where BERT can achieve state-of-the-art results.
State-of-the-art spoken language identification (LID) systems, which are based on end-to-end deep neural networks, have shown remarkable success not only in discriminating between distant languages but also between closely-related languages or even different spoken varieties of the same language. However, it is still unclear to what extent neural LID models generalize to speech samples with different acoustic conditions due to domain shift. In this paper, we present a set of experiments to investigate the impact of domain mismatch on the performance of neural LID systems for a subset of six Slavic languages across two domains (read speech and radio broadcast) and examine two low-level signal descriptors (spectral and cepstral features) for this task. Our experiments show that (1) out-of-domain speech samples severely hinder the performance of neural LID models, and (2) while both spectral and cepstral features show comparable performance within-domain, spectral features show more robustness under domain mismatch. Moreover, we apply unsupervised domain adaptation to minimize the discrepancy between the two domains in our study. We achieve relative accuracy improvements that range from 9% to 77% depending on the diversity of acoustic conditions in the source domain.
This paper describes SalamNET, an Arabic offensive language detection system that has been submitted to SemEval 2020 shared task 12: Multilingual Offensive Language Identification in Social Media. Our approach focuses on applying multiple deep learning models and conducting in depth error analysis of results to provide system implications for future development considerations. To pursue our goal, a Recurrent Neural Network (RNN), a Gated Recurrent Unit (GRU), and Long-Short Term Memory (LSTM) models with different design architectures have been developed and evaluated. The SalamNET, a Bi-directional Gated Recurrent Unit (Bi-GRU) based model, reports a macro-F1 score of 0.83.
This paper describes the Duluth systems that participated in SemEval--2020 Task 12, Multilingual Offensive Language Identification in Social Media (OffensEval--2020). We participated in the three English language tasks. Our systems provide a simple Machine Learning baseline using logistic regression. We trained our models on the distantly supervised training data made available by the task organizers and used no other resources. As might be expected we did not rank highly in the comparative evaluation: 79th of 85 in Task A, 34th of 43 in Task B, and 24th of 39 in Task C. We carried out a qualitative analysis of our results and found that the class labels in the gold standard data are somewhat noisy. We hypothesize that the extremely high accuracy (> 90%) of the top ranked systems may reflect methods that learn the training data very well but may not generalize to the task of identifying offensive language in English. This analysis includes examples of tweets that despite being mildly redacted are still offensive.
In this paper, we describe our approach to utilize pre-trained BERT models with Convolutional Neural Networks for sub-task A of the Multilingual Offensive Language Identification shared task (OffensEval 2020), which is a part of the SemEval 2020. We show that combining CNN with BERT is better than using BERT on its own, and we emphasize the importance of utilizing pre-trained language models for downstream tasks. Our system, ranked 4th with macro averaged F1-Score of 0.897 in Arabic, 4th with score of 0.843 in Greek, and 3rd with score of 0.814 in Turkish. Additionally, we present ArabicBERT, a set of pre-trained transformer language models for Arabic that we share with the community.
This paper presents six document classification models using the latest transformer encoders and a high-performing ensemble model for a task of offensive language identification in social media. For the individual models, deep transformer layers are applied to perform multi-head attentions. For the ensemble model, the utterance representations taken from those individual models are concatenated and fed into a linear decoder to make the final decisions. Our ensemble model outperforms the individual models and shows up to 8.6% improvement over the individual models on the development set. On the test set, it achieves macro-F1 of 90.9% and becomes one of the high performing systems among 85 participants in the sub-task A of this shared task. Our analysis shows that although the ensemble model significantly improves the accuracy on the development set, the improvement is not as evident on the test set.
In this paper, we present various systems submitted by our team problemConquero for SemEval-2020 Shared Task 12 Multilingual Offensive Language Identification in Social Media. We participated in all the three sub-tasks of OffensEval-2020, and our final submissions during the evaluation phase included transformer-based approaches and a soft label-based approach. BERT based fine-tuned models were submitted for each language of sub-task A (offensive tweet identification). RoBERTa based fine-tuned model for sub-task B (automatic categorization of offense types) was submitted. We submitted two models for sub-task C (offense target identification), one using soft labels and the other using BERT based fine-tuned model. Our ranks for sub-task A were Greek-19 out of 37, Turkish-22 out of 46, Danish-26 out of 39, Arabic-39 out of 53, and English-20 out of 85. We achieved a rank of 28 out of 43 for sub-task B. Our best rank for sub-task C was 20 out of 39 using BERT based fine-tuned model.
Discussions on Twitter involve participation from different communities with different dialects and it is often necessary to summarize a large number of posts into a representative sample to provide a synopsis. Yet, any such representative sample should sufficiently portray the underlying dialect diversity to present the voices of different participating communities representing the dialects. Extractive summarization algorithms perform the task of constructing subsets that succinctly capture the topic of any given set of posts. However, we observe that there is dialect bias in the summaries generated by common summarization approaches, i.e., they often return summaries that under-represent certain dialects. The vast majority of existing "fair" summarization approaches require socially salient attribute labels (in this case, dialect) to ensure that the generated summary is fair with respect to the socially salient attribute. Nevertheless, in many applications, these labels do not exist. Furthermore, due to the ever-evolving nature of dialects in social media, it is unreasonable to label or accurately infer the dialect of every social media post. To correct for the dialect bias, we employ a framework that takes an existing text summarization algorithm as a blackbox and, using a small set of dialect-diverse sentences, returns a summary that is relatively more dialect-diverse. Crucially, this approach does not need the posts being summarized to have dialect labels, ensuring that the diversification process is independent of dialect classification/identification models. We show the efficacy of our approach on Twitter datasets containing posts written in dialects used by different social groups defined by race or gender; in all cases, our approach leads to improved dialect diversity compared to standard text summarization approaches.
The task of written language identification involves typically the detection of the languages present in a sample of text. Moreover, a sequence of text may not belong to a single inherent language but also may be mixture of text written in multiple languages. This kind of text is generated in large volumes from social media platforms due to its flexible and user friendly environment. Such text contains very large number of features which are essential for development of statistical, probabilistic as well as other kinds of language models. The large number of features have rich as well as irrelevant and redundant features which have diverse effect over the performance of the learning model. Therefore, feature selection methods are significant in choosing feature that are most relevant for an efficient model. In this article, we basically consider the Hindi-English language identification task as Hindi and English are often two most widely spoken languages of India. We apply different feature selection algorithms across various learning algorithms in order to analyze the effect of the algorithm as well as the number of features on the performance of the task. The methodology focuses on the word level language identification using a novel dataset of 6903 tweets extracted from Twitter. Various n-gram profiles are examined with different feature selection algorithms over many classifiers. Finally, an exhaustive comparative analysis is put forward with respect to the overall experiments conducted for the task.
Due to a drastic improvement in the quality of internet services worldwide, there is an explosion of multilingual content generation and consumption. This is especially prevalent in countries with large multilingual audience, who are increasingly consuming media outside their linguistic familiarity/preference. Hence, there is an increasing need for real-time and fine-grained content analysis services, including language identification, content transcription, and analysis. Accurate and fine-grained spoken language detection is an essential first step for all the subsequent content analysis algorithms. Current techniques in spoken language detection may lack on one of these fronts: accuracy, fine-grained detection, data requirements, manual effort in data collection \& pre-processing. Hence in this work, a real-time language detection approach to detect spoken language from 5 seconds' audio clips with an accuracy of 91.8\% is presented with exiguous data requirements and minimal pre-processing. Novel architectures for Capsule Networks is proposed which operates on spectrogram images of the provided audio snippets. We use previous approaches based on Recurrent Neural Networks and iVectors to present the results. Finally we show a ``Non-Class'' analysis to further stress on why CapsNet architecture works for LID task.
Code-switching (CS) is a common phenomenon and recognizing CS speech is challenging. But CS speech data is scarce and there' s no common testbed in relevant research. This paper describes the design and main outcomes of the ASRU 2019 Mandarin-English code-switching speech recognition challenge, which aims to improve the ASR performance in Mandarin-English code-switching situation. 500 hours Mandarin speech data and 240 hours Mandarin-English intra-sentencial CS data are released to the participants. Three tracks were set for advancing the AM and LM part in traditional DNN-HMM ASR system, as well as exploring the E2E models' performance. The paper then presents an overview of the results and system performance in the three tracks. It turns out that traditional ASR system benefits from pronunciation lexicon, CS text generating and data augmentation. In E2E track, however, the results highlight the importance of using language identification, building-up a rational set of modeling units and spec-augment. The other details in model training and method comparsion are discussed.
Multilingual ASR technology simplifies model training and deployment, but its accuracy is known to depend on the availability of language information at runtime. Since language identity is seldom known beforehand in real-world scenarios, it must be inferred on-the-fly with minimum latency. Furthermore, in voice-activated smart assistant systems, language identity is also required for downstream processing of ASR output. In this paper, we introduce streaming, end-to-end, bilingual systems that perform both ASR and language identification (LID) using the recurrent neural network transducer (RNN-T) architecture. On the input side, embeddings from pretrained acoustic-only LID classifiers are used to guide RNN-T training and inference, while on the output side, language targets are jointly modeled with ASR targets. The proposed method is applied to two language pairs: English-Spanish as spoken in the United States, and English-Hindi as spoken in India. Experiments show that for English-Spanish, the bilingual joint ASR-LID architecture matches monolingual ASR and acoustic-only LID accuracies. For the more challenging (owing to within-utterance code switching) case of English-Hindi, English ASR and LID metrics show degradation. Overall, in scenarios where users switch dynamically between languages, the proposed architecture offers a promising simplification over running multiple monolingual ASR models and an LID classifier in parallel.
With the growing use of social media and its availability, many instances of the use of offensive language have been observed across multiple languages and domains. This phenomenon has given rise to the growing need to detect the offensive language used in social media cross-lingually. In OffensEval 2020, the organizers have released the \textit{multilingual Offensive Language Identification Dataset} (mOLID), which contains tweets in five different languages, to detect offensive language. In this work, we introduce a cross-lingual inductive approach to identify the offensive language in tweets using the contextual word embedding \textit{XLM-RoBERTa} (XLM-R). We show that our model performs competitively on all five languages, obtaining the fourth position in the English task with an F1-score of $0.919$ and eighth position in the Turkish task with an F1-score of $0.781$. Further experimentation proves that our model works competitively in a zero-shot learning environment, and is extensible to other languages.
Lexical normalization, the translation of non-canonical data to standard language, has shown to improve the performance of manynatural language processing tasks on social media. Yet, using multiple languages in one utterance, also called code-switching (CS), is frequently overlooked by these normalization systems, despite its common use in social media. In this paper, we propose three normalization models specifically designed to handle code-switched data which we evaluate for two language pairs: Indonesian-English (Id-En) and Turkish-German (Tr-De). For the latter, we introduce novel normalization layers and their corresponding language ID and POS tags for the dataset, and evaluate the downstream effect of normalization on POS tagging. Results show that our CS-tailored normalization models outperform Id-En state of the art and Tr-De monolingual models, and lead to 5.4% relative performance increase for POS tagging as compared to unnormalized input.