Tech Showcase: Fake News Detectors and Many More

Fake News Detector

The fast-spreading nature of fraudulent information has emerged as a pressing issue. Distorted facts, cleverly disguised as authentic news, can easily deceive and mislead. In particularly crucial moments, such as political elections or global pandemics, the insidious impact of fake news becomes amplified.

The rapid spread of rumors and deceitful reports can have severe consequences, even endangering human lives. In light of this, it is imperative to promptly detect and combat this phenomenon to prevent the escalation of panic and the misguidance of a vast population. This presents an opportunity for an interesting project.

Translator App

For those interested in entering the field of Natural Language Processing, a great project to kickstart your journey is building a translator app with the assistance of a transformer. A transformer model extracts features from sentences and also determines the significance of each word within a sentence. This powerful tool consists of both encoding and decoding components, both of which are expertly trained end-to-end.

With the help of a transformer, you have the opportunity to create your very own AI translator app. Simply load a pre-trained transformer model into your Python environment and convert your desired text into tokens to be inputted into the model. For this purpose, the GluonNLP library is highly recommended. Additionally, the same library allows you to easily access the train and test datasets required for this exciting AI project.

Instagram Spam Detection

Have you ever been notified of a comment on your Instagram post, only to eagerly grab your phone and find it’s yet another sneaky bot promoting bogus shoes? The comment sections of countless Instagram posts are infiltrated with these machines. Some simply annoy, while others can be outright dangerous, demanding action from you. But fear not – with the help of AI techniques, you can create a powerful spam detection model to distinguish between spam and genuine comments.

While it may be challenging to locate a dataset specifically dedicated to Instagram spam comments, there are ways to gather this crucial information for your analysis. One such method is web scraping, through which you can access unlabelled comments from Instagram using the Python programming language. Alternatively, you could utilize a different dataset for training purposes, such as the YouTube spam collection dataset found on Kaggle.