News category classifier with voice enabled feature and headlines finder

After pressing Record button App will return what the user says into the microphone and will predict the nearest possible category for the right genre of the news among 5 different categories namely Politics, Technology, Entertainment, Business and health, also giving the latest news headlines relevant to the topic

Business Value

Easily marketable AI technology which uses Facebook popular open source RoBERTa algorithm
App is easily extensible by pre-training the model on more news categories.
User can save time by getting needful current news results around the asked topic
Voice controlled feature makes it more fun and interactive expanding its possibilities to be a trending app among users

Implementation Plan

Model was trained on a dataset of 2500 different news topics uniformly distributed among 5 different categories namely Politics – 0, Technology- 1, Entertainment- 2, Business-3 and health-4.

Company
Skills
Used technologies
Complexity
Easy
Effort Needed
1 week
ROI
na
Customer
na
Industry
AI
Partner
na
Application Type
AI enabled Topic Identifier
Tech Stack
Python is the core building technology for the app. Model extraction , computation and pre-training was done within a jupyter environment. GUI was built using tkinter which is a python package with thin object-oriented layer on top of Tcl/Tk.
Learning Path
DataCamp
Github
https://github.com/Codila/SpeechBasedNewsClassifier
X
X