Seshat

Inspiration

After the recent pandemic began, most of the students switched to online learning mode due to travel restriction. While online learning has several challenges, it also brings several opportunities in this area. That caught my attention and i decided to build something useful for students as well as teachers.

What it does

Sesaht generates deep analysis on given text input like below:

• Identifying key topics from the input, that helps to understand what this is about.
• Generates short summary of long input, it ease the process of understanding any topics quickly and remembering it.
• Identifies and highlights the key terms from the input like person, location, organization etc to quickly get the useful information in very short time.
• Stores all the previously generated results to quickly review it.

How we built it

The Web-App is built with Python Streamlit library and it calls the Below mentioned Modzy APIs to generate results:

• Text Topic Modelling
• Text Summarization
• Named Entity Recognition
• Text to Speech Conversion

Challenges we ran into

Problem formation and which APIs of Modzy can be useful in this project, deciding this part was very challenging. And Understanding of Modzy APIs and it's responses was little challenging but they have provided great documentation so it helped a lot.

Accomplishments that we're proud of

Journey from Problem definition to building working POC has been a great, learned a lot of things during this.

What's next for Seshat

I have below tasks planned:

• Integration of search in audio files, that way students can find topic discussed during online lectures or from recorded lecture and jump to that part.
• Generating Q&A from the Input Corpus to practice.

Built With

• modzy
• python
• streamlit