The ever-evolving landscape of online social media has initiated a transformative phase in communication, presenting unprecedented opportunities alongside inherent challenges. The pervasive issue of false information, commonly termed fake news, has emerged as a significant concern within these dynamic platforms. This study delves into the domain of Fake News Detection, with a specific focus on Malayalam. Utilizing advanced transformer models like mBERT, ALBERT, and XMLRoBERTa, our research proficiently classifies social media text into original or fake categories. Notably, our proposed model achieved commendable results, securing a rank of 3 in Task 1 with macro F1 scores of 0.84 using mBERT, 0.56 using ALBERT, and 0.84 using XMLRoBERTa. In Task 2, the XMLRoBERTa model excelled with a rank of 12, attaining a macro F1 score of 0.21, while mBERT and BERT achieved scores of 0.16 and 0.11, respectively. This research aims to develop robust systems capable of discerning authentic from deceptive content, a crucial endeavor in maintaining information reliability on social media platforms amid the rampant spread of misinformation.

Social media has become a vital platform for personal communication. Its widespread use as a primary means of public communication offers an exciting opportunity for early detection and management of mental health issues. People often share their emotions on social media, but understanding the true depth of their feelings can be challenging. Depression, a prevalent problem among young people, is of particular concern due to its link with rising suicide rates. Identifying depression levels in social media texts is crucial for timely support and prevention of negative outcomes. However, it’s a complex task because human emotions are dynamic and can change significantly over time. The DepSign-LT-EDI@RANLP 2023 shared task aims to classify social media text into three depression levels: “Not Depressed,” “Moderately Depressed,” and “Severely Depressed.” This overview covers task details, dataset, methodologies used, and results analysis. Roberta-based models emerged as top performers, with the best result achieving an impressive macro F1-score of 0.584 among 31 participating teams.

Depression is a mental fitness disorder from persistent reactions of unhappiness, void, and a deficit of interest in activities. It can influence differing facets of one’s life, containing their hopes, sympathy, and nature. Depression can stem from a sort of determinant, in the way that ancestral willingness, life occurrences, and social circumstances. In current years, the influence of social media on mental fitness has become an increasing concern. Excessive use of social media and the negative facets that guide it, can exacerbate or cause impressions of distress. The nonstop exposure to cautiously curated lives, social comparison, cyberbullying, and the pressure to meet unreal standards can impact an individual’s pride, social connections, and overall well-being. We participated in the shared task at DepSignLT-EDI@RANLP 2023 and have proposed a model that identifies the levels of depression from social media text using the data set shared for the task. Different transformer models like ALBERT and RoBERTa are used by the proposed model for implementing the task. The macro F1 score obtained by ALBERT model and RoBERTa model are 0.258 and 0.143 respectively.

The prevalence of depression is increasing globally, and there is a need for effective screening and detection tools. Social media platforms offer a rich source of data for mental health research. The paper aims to detect the signs of depression of a person from their social media postings wherein people share their feelings and emotions. The task is to create a system that, given social media posts in English, should classify the level of depression as ‘not depressed’, ‘moderately depressed’ or ‘severely depressed’. The paper presents the solution for the Shared Task on Detecting Signs of Depression from Social Media Text at LT-EDI@RANLP 2023. The proposed system aims to develop a machine learning model using machine learning algorithms like SVM, Random forest and Naive Bayes to detect signs of depression from social media text. The model is trained on a dataset of social media posts to detect the level of depression of the individuals as ‘not depressed’, ‘moderately depressed’ or ‘severely depressed’. The dataset is pre-processed to remove duplicates and irrelevant features, and then, feature engineering techniques is used to extract meaningful features from the text data. The model is trained on these features to classify the text into the three categories. The performance of the model is evaluated using metrics such as accuracy, precision, recall, and F1-score. The ensemble model is used to combine these algorithms which gives accuracy of 90.2% and the F1 score is 0.90. The results of the proposed approach could potentially aid in the early detection and prevention of depression for individuals who may be at risk.

Sarcasm refers to the use of words that have different literal and intended meanings. It represents the usage of words that are opposite of what is literally said, especially in order to insult, mock, criticise or irritate someone. These types of statements may be funny or amusing to others but may hurt or annoy the person towards whom it is intended. Identification of sarcastic phrases from social media posts finds its application in different domains like sentiment analysis, opinion mining, author profiling, and harassment detection. We have proposed a model for the shared task iSarcasmEval - Intended Sarcasm Detection in English and Arabic (CITATION) by SemEval-2022 considering the language English based on ELmo embeddings for Subtasks A and C and TF-IDF vectors and Gaussian Naive bayes classifier for Subtask B. The proposed model resulted in a F1 score 0.2012 for sarcastic texts in Subtask A, macro-F1 score of 0.0387 and 0.2794 for Subtasks B and C respectively.

Social media platforms play a major role in our day-to-day life and are considered as a virtual friend by many users, who use the social media to share their feelings all day. Many a time, the content which is shared by users on social media replicate their internal life. Nowadays people love to share their daily life incidents like happy or unhappy moments and their feelings in social media and it makes them feel complete and it has become a habit for many users. Social media provides a new chance to identify the feelings of a person through their posts. The aim of the shared task is to develop a model in which the system is capable of analyzing the grammatical markers related to onset and permanent symptoms of depression. We as a team participated in the shared task Detecting Signs of Depression from Social Media Text at LT-EDI 2022- ACL 2022 and we have proposed a model which predicts depression from English social media posts using the data set shared for the task. The prediction is done based on the labels Moderate, Severe and Not Depressed. We have implemented this using different transformer models like DistilBERT, RoBERTa and ALBERT by which we were able to achieve a Macro F1 score of 0.337, 0.457 and 0.387 respectively. Our code is publicly available in the github

Social media is considered as a platform whereusers express themselves. The rise of social me-dia as one of humanity’s most important publiccommunication platforms presents a potentialprospect for early identification and manage-ment of mental illness. Depression is one suchillness that can lead to a variety of emotionaland physical problems. It is necessary to mea-sure the level of depression from the socialmedia text to treat them and to avoid the nega-tive consequences. Detecting levels of depres-sion is a challenging task since it involves themindset of the people which can change period-ically. The aim of the DepSign-LT-EDI@ACL-2022 shared task is to classify the social me-dia text into three levels of depression namely“Not Depressed”, “Moderately Depressed”, and“Severely Depressed”. This overview presentsa description on the task, the data set, method-ologies used and an analysis on the results ofthe submissions. The models that were submit-ted as a part of the shared task had used a va-riety of technologies from traditional machinelearning algorithms to deep learning models. It could be observed from the result that thetransformer based models have outperformedthe other models. Among the 31 teams whohad submitted their results for the shared task,the best macro F1-score of 0.583 was obtainedusing transformer based model.