Striving for educational equity, one word at a time.
Despite the advancements made, education in the United States and around the world still faces an inequality in the amount of opportunities provided to students. Financial status is still one of the most important factors in the quality of education. Far too many students lack access to a great education.
We have personally experienced this as first and second-generation immigrants whose parents struggled with learning English. With the ongoing COVID-19 pandemic and transition by many schools to distance learning, the inequality will only widen as students get less opportunities to talk to teachers in person. Receiving meaningful direct feedback on writing that’s more than simply spelling is critical to a better understanding of the language and overall success.
Thus, we decided to create NeuroScribe, a free tool that will help address this inequality and help anyone improve their English! Even for those seasoned in the ways of writing, this tool can be used to find that perfect word that was at the tip of your tongue.
AI Integration, How We Built It
NeuroScribe uses a machine learning model to analyze the inputted text, checking whether each word makes sense within the context of the sentence.
More specifically, Neuroscribe uses Natural Language Processing (NLP) with TensorFlow in Python to analyze the text. A bidirectional LSTM (Long Short Term Memory) model then sifts through the input and looks for unreasonable words within the text, as well as suggested corrections.
As next steps, we would like to train the model on a larger or more specific dataset such as business emails to provide more tailored suggestions. We would also like to make the model give more detailed feedback, making it more helpful and furthering our goal of educational equity.