Text document summarization is critical for managing today's vast textual data. This paper presents an approach to text document summarization that does not rely on word embedding techniques. Instead, our method follows a step-by-step process, including sentence segmentation, sentence embedding, K-means clustering, and summary generation. The input text is segmented into individual sentences using an NLP tool such as NLTK's sentence tokenizer. Next, we extract contextual embeddings for each sentence using the Sentence Transformer method. These embeddings capture the meaning of each sentence within the context of the surrounding text. The sentence embeddings are then subjected to K-means clustering. This step enables the creation of clusters that represent semantically related sentences. To generate the summary, depending on how far each sentence is from the cluster centroid, we choose one sentence from each cluster. The sentence with the lowest distance from the centroid is chosen, and the selected sentences are ordered as they appeared in the original text. We implemented the summarizer and evaluated its performance on the DUC 2007 dataset, a collection of news articles with manually crafted summaries by human experts. The results demonstrate that our summarizer produces informative and concise summaries, surpassing a baseline approach that solely extracts top-ranked sentences from the input text. Our work contributes to text document summarization by presenting an alternative approach that does not rely on word embedding techniques. By leveraging sentence segmentation, contextual embeddings, K-means clustering, and centroid-based selection, our method offers a viable solution for generating high-quality summaries. Further research can explore enhancements to our approach and its application in various domains where text summarization is essential.

Hate speech is harmful content that directly attacks or promotes hatred against members of groups or individuals based on actual or perceived aspects of identity, such as racism, religion, or sexual orientation. This can affect social life on social media platforms as hateful content shared through social media can harm both individuals and communities. As the prevalence of hate speech increases online, the demand for automated detection as an NLP task is increasing. In this work, the proposed method is using transformer-based model to detect hate speech in social media, like twitter, Facebook, WhatsApp, Instagram, etc. The proposed model is independent of languages and has been tested on Italian, English, German, Bengali. The Gold standard datasets were collected from renowned researcher Zeerak Talat, Sara Tonelli, Melanie Siegel, and Rezaul Karim. The success rate of the proposed model for hate speech detection is higher than the existing baseline and state-of-the-art models with accuracy in Bengali dataset is 89%, in English: 91%, in German dataset 91% and in Italian dataset it is 77%. The proposed algorithm shows substantial improvement to the benchmark method.