A voice-based personal assistant built for students, by students to ease and streamline their experience.
Here we aim to guide them through every stage of the learning process.
This figure explains the system architecture of the web application:
Right from information retrieval to evaluation, we have them covered.
It has different functionalities right from voice based text and image retrieval, a chatbot to a text summarizer and an automatic question generator.
Made by amalgamating different concepts of NLP and Machine Learning together.
The question generator model takes a text as input and outputs a series of question and answer pairs. The answers are sentences and phrases extracted from the input text using spaCy. Named entities are used for multiple-choice answers. The wrong answers will be other entities of the same type found in the text. The questions are generated by concatenating the extracted answer with the full text (up to a maximum of 512 tokens) as context. The concatenated string is then encoded and fed into the question generator model having t5-base. The pretrained model was finetuned as a sequence-to-sequence model on a dataset made up several well-known QA datasets The datasets were restructured by concatenating the answer and context fields into the previously mentioned format. The concatenated answer and context was then used as an input for training, and the question field became the targets.
In our web application, we have used Abstractive Summarization. In that we have developed, we made use of the Text To Text Transfer Transformer i.e. the T5 tokenizer by Google. The python library used here is PyTorch and transformers. From transformers, we import the t5 text tokenizer. It achieves state-of-the-art results on multiple NLP tasks like summarization using a text-to-text transformer trained on a large text corpus. This makes it a very powerful and reliable tool to carry out text summarization. In this project, we will be taking the input as a pdf file to carry out the summarization.
This feature acts as a search engine for a technical word. Our web application gives output as links of some most viewed websites which include some information of the provided word. This also works similarly for images, that is it searches the web and gives the output for links of images as per the given input. We are using the custom search API of Google.
And thus using state-of-the-art techniques, we have strived to create a product that truly brings value and acts as a one stop destination for all their needs.