Skip to content

Code for the paper on neural title generation. It uses encoder based LSTM architectures and some heuristics to predict catchy titles for scientific papers

Notifications You must be signed in to change notification settings

jain-harshil/Neural-Title-Generation

Repository files navigation

Neural Title Generation for Scientific Papers

This repository contains all code for the paper TiZen: Neural Title Generation for Scientific Papers

Images of TiZen Dashboard:

Example 1

Actual Title - You Only Look Once: Unified, Real-Time Object Detection

Example 2

Actual Title - SEAGLE: Sparsity-Driven Image Reconstruction under Multiple Scattering

Example 3

Actual Title - ExEC: Elastic Extensible Edge Cloud

Example 4

Actual Title - Weakly-Supervised Deep Learning for Domain Invariant Sentiment Classification

Table showing comparison of actual titles and titles generated by TiZen along with Catchiness Scores

A comparison of actual and predicted titles by TiZen is illustrated in the table below:

About

Code for the paper on neural title generation. It uses encoder based LSTM architectures and some heuristics to predict catchy titles for scientific papers

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published