The emergence of chatbots over the last 50 years has been the primary consequence of the need of a virtual aid. Unlike their biological anthropomorphic counterpart in the form of fellow homo sapiens, chatbots have the ability to instantaneously present themselves at the user's need and convenience. Be it for something as benign as feeling the need of a friend to talk to, to a more dire case such as medical assistance, chatbots are unequivocally ubiquitous in their utility. This paper aims to develop one such chatbot that is capable of not only analyzing human text (and speech in the near future), but also refining the ability to assist them medically through the process of accumulating data from relevant datasets. Although Recurrent Neural Networks (RNNs) are often used to develop chatbots, the constant presence of the vanishing gradient issue brought about by backpropagation, coupled with the cumbersome process of sequentially parsing each word individually has led to the increased usage of Transformer Neural Networks (TNNs) instead, which parses entire sentences at once while simultaneously giving context to it via embeddings, leading to increased parallelization. Two variants of the TNN Bidirectional Encoder Representations from Transformers (BERT), namely KeyBERT and BioBERT, are used for tagging the keywords in each sentence and for contextual vectorization into Q/A pairs for matrix multiplication, respectively. A final layer of GPT-2 (Generative Pre-trained Transformer) is applied to fine-tune the results from the BioBERT into a form that is human readable. The outcome of such an attempt could potentially lessen the need for trips to the nearest physician, and the temporal delay and financial resources required to do so.

Along with the growth of the Internet, social media usage has drastically expanded. As people share their opinions and ideas more frequently on the Internet and through various social media platforms, there has been a notable rise in the number of consumer phrases that contain sentiment data. According to reports, cyberbullying frequently leads to severe emotional and physical suffering, especially in women and young children. In certain instances, it has even been reported that sufferers attempt suicide. The bully may occasionally attempt to destroy any proof they believe to be on their side. Even if the victim gets the evidence, it will still be a long time before they get justice at that point. This work used OCR, NLP, and machine learning to detect cyberbullying in photos in order to design and execute a practical method to recognize cyberbullying from images. Eight classifier techniques are used to compare the accuracy of these algorithms against the BoW Model and the TF-IDF, two key features. These classifiers are used to understand and recognize bullying behaviors. Based on testing the suggested method on the cyberbullying dataset, it was shown that linear SVC after OCR and logistic regression perform better and achieve the best accuracy of 96 percent. This study aid in providing a good outline that shapes the methods for detecting online bullying from a screenshot with design and implementation details.