arXiv subjects

Tharindu Ranasinghe

Publications and source records attributed to Tharindu Ranasinghe.

63 records · Page 4Linked to original sources

arXiv2020

BRUMS at SemEval-2020 Task 12 : Transformer based Multilingual Offensive Language Identification in Social Media

In this paper, we describe the team \textit{BRUMS} entry to OffensEval 2: Multilingual Offensive Language Identification in Social Media in SemEval-2020. The OffensEval organizers provided participants with annotated datasets containing posts from social media in Arabic, Danish, English, Greek and Turkish. We present a multilingual deep learning model to identify offensive language in social media. Overall, the approach achieves acceptable evaluation scores, while maintaining flexibility between languages.

cs.CL↗

arXiv2020

RGCL at SemEval-2020 Task 6: Neural Approaches to Definition Extraction

This paper presents the RGCL team submission to SemEval 2020 Task 6: DeftEval, subtasks 1 and 2. The system classifies definitions at the sentence and token levels. It utilises state-of-the-art neural network architectures, which have some task-specific adaptations, including an automatically extended training set. Overall, the approach achieves acceptable evaluation scores, while maintaining flexibility in architecture selection.

cs.CL↗

arXiv2020

TransQuest at WMT2020: Sentence-Level Direct Assessment

This paper presents the team TransQuest's participation in Sentence-Level Direct Assessment shared task in WMT 2020. We introduce a simple QE framework based on cross-lingual transformers, and we use it to implement and evaluate two different neural architectures. The proposed methods achieve state-of-the-art results surpassing the results obtained by OpenKiwi, the baseline used in the shared task. We further fine tune the QE framework by performing ensemble and data augmentation. Our approach is the winning solution in all of the language pairs according to the WMT 2020 official results.

cs.CL↗

arXiv2020

Multilingual Offensive Language Identification with Cross-lingual Embeddings

Offensive content is pervasive in social media and a reason for concern to companies and government organizations. Several studies have been recently published investigating methods to detect the various forms of such content (e.g. hate speech, cyberbulling, and cyberaggression). The clear majority of these studies deal with English partially because most annotated datasets available contain English data. In this paper, we take advantage of English data available by applying cross-lingual contextual word embeddings and transfer learning to make predictions in languages with less resources. We project predictions on comparable data in Bengali, Hindi, and Spanish and we report results of 0.8415 F1 macro for Bengali, 0.8568 F1 macro for Hindi, and 0.7513 F1 macro for Spanish. Finally, we show that our approach compares favorably to the best systems submitted to recent shared tasks on these three languages, confirming the robustness of cross-lingual contextual embeddings and transfer learning for this task.

cs.CL↗

arXiv2020

InfoMiner at WNUT-2020 Task 2: Transformer-based Covid-19 Informative Tweet Extraction

Identifying informative tweets is an important step when building information extraction systems based on social media. WNUT-2020 Task 2 was organised to recognise informative tweets from noise tweets. In this paper, we present our approach to tackle the task objective using transformers. Overall, our approach achieves 10th place in the final rankings scoring 0.9004 F1 score for the test set.

cs.CL↗

arXiv2020

Intelligent Translation Memory Matching and Retrieval with Sentence Encoders

Matching and retrieving previously translated segments from a Translation Memory is the key functionality in Translation Memories systems. However this matching and retrieving process is still limited to algorithms based on edit distance which we have identified as a major drawback in Translation Memories systems. In this paper we introduce sentence encoders to improve the matching and retrieving process in Translation Memories systems - an effective and efficient solution to replace edit distance based algorithms.

cs.CL↗

arXiv2020

Offensive Language Identification in Greek

As offensive language has become a rising issue for online communities and social media platforms, researchers have been investigating ways of coping with abusive content and developing systems to detect its different types: cyberbullying, hate speech, aggression, etc. With a few notable exceptions, most research on this topic so far has dealt with English. This is mostly due to the availability of language resources for English. To address this shortcoming, this paper presents the first Greek annotated dataset for offensive language identification: the Offensive Greek Tweet Dataset (OGTD). OGTD is a manually annotated dataset containing 4,779 posts from Twitter annotated as offensive and not offensive. Along with a detailed description of the dataset, we evaluate several computational models trained and tested on this data.

cs.CL↗

arXiv2019

Concept Discovery through Information Extraction in Restaurant Domain

Concept identification is a crucial step in understanding and building a knowledge base for any particular domain. However, it is not a simple task in very large domains such as restaurants and hotel. In this paper, a novel approach of identifying a concept hierarchy and classifying unseen words into identified concepts related to restaurant domain is presented. Sorting, identifying, classifying of domain-related words manually is tedious and therefore, the proposed process is automated to a great extent. Word embedding, hierarchical clustering, classification algorithms are effectively used to obtain concepts related to the restaurant domain. Further, this approach can also be extended to create a semi-automatic ontology on restaurant domain.

cs.CL↗

arXiv2019

User Profile Feature-Based Approach to Address the Cold Start Problem in Collaborative Filtering for Personalized Movie Recommendation

A huge amount of user generated content related to movies is created with the popularization of web 2.0. With these continues exponential growth of data, there is an inevitable need for recommender systems as people find it difficult to make informed and timely decisions. Movie recommendation systems assist users to find the next interest or the best recommendation. In this proposed approach the authors apply the relationship of user feature-scores derived from user-item interaction via ratings to optimize the prediction algorithm's input parameters used in the recommender system to improve the accuracy of predictions when there are less past user records. This addresses a major drawback in collaborative filtering, the cold start problem by showing an improvement of 8.4% compared to the base collaborative filtering algorithm. The user-feature generation and evaluation of the system is carried out using the 'MovieLens 100k dataset'. The proposed system can be generalized to other domains as well.

cs.IR↗