Automated evaluation of generated headlines

Printer-friendly versionSend by email
Qualifications Required: 
a) Experience with internet technologies and web applications, b) Good programming skills Java/Python
Qualifications Desired: 
reading/writing research articles, machine learning, natural language processing

This project deals with developing a platform for the automated evaluation of headlines. The platform will contain several metrics, which will give a score for each generated headline given its document and gold-standard headlines (manually created). The platform may include existing metrics (like Rouge), extended metrics (which integrate several existing approaches and based on the existing ones), and novel metrics (which will be elaborated during the project).

Additional Keywords: Headline generation, ROUGE, text similarity, summarization, KL-divergence, VSM, word embedding


© 2018 - Institute of Informatics and Telecommunications | National Centre for Scientific Research "Demokritos"