Overview

for any AI enthusiast’s person, it is a hassle to go to a different website and search for the latest AI news and updates. AI-News-by-AI is a podcast created by AI for AI enthusiasts. More than 51% of the US population has listened to a podcast in 2019. The rising internet era has given lots of choices to audiences but only 35% of those listening to the whole podcast. The main reason for this is because of non-interesting and lengthy content for audiences. This AI-News-By-AI is the solution to that problem by using AI to create a podcast.

The system starts by searching the preferred AI-related news and updates on the internet. The system then curates the content and sends it to the cleaner. The cleaner removes the unwanted ads, texts, emoji, and links from the content and sends it to the neural net summarizer. The pre-trained seq2seq model summarizes the given content to a much smaller extractive summary. The summarized content is then converted to the speech files using Wave Net and then automatically broadcasted to Apples’ iTunes Podcast.

AI-News-By-AI uses the latest AI technology to create the AI-Podcasts for AI enthusiast by itself and broadcast the latest podcast every morning to the Apples’ iTunes Podcast. The podcasts are summarized with the relative and interesting AI content around the world giving audiences the better experience.

Current Team Members:

Dr. Vahid Behzadan,
Bibek Upadhayay

Tools and Datasets:

We started our project with Python 3 for the programming. We are using various libraries for curating news from different websites. Below given example of audio is the conversion of feeds directly from AI NEWS. The text are then converted to audio files using Google WaveNet neural network API which are after then converted into Podcast Episodes.

Project Link: AI NEWS by AI