JACOB DAVID CARLS

Developer | B.Sc.(Psych) M.Sc.(AppLing)

Projects



<> Text Segmentation NLP

Languages: PHP server-side & html, css, js client-side

Here is a rough version of a web application I am working on: Text Segmentation Tool . This is an initial version - the tool has already been deployed to the web-server in its current state.

Note: Right now the tool is not using any AI.

Originally it was written in Python. I rewrote the functionality in JavaScript and had even the main functions rendered in the browser – I thought that was probably not a good idea and transferred the logic of the program to run in PHP on a server. The program allows the abstraction of structural and semantic features from text input. The application can return various structures such as number of words, sentences, consonants, nouns etc. On a semantic level, it allows the user to filter out semantically important information through the removal of “meaningless” language structures such as articles, prepositions and conjunctions. In its current state, there are just a few functions available as it is still work in progress.



<> Sentiment Analysis

I am planning to include Sentiment Analysis as a function in the Text Segmentation Tool. This project involves the use of PyTorch in applying the softmax function to convert model outputs into probabilities. The tool utilizes the BERT model from the Transformers library provided by Hugging Face, allowing text classification into ‘positive’ or ‘negative’ sentiment categories. My webserver is currently running PHP, so I have to find a way to do this using some PHP libraries.