Research PaperResearchia:202611.03064

Song lyrics generation using machine learning techniques

Casey Cannon· Research institute: California State Polytechnic University

Abstract

This research covered various lyric generation using machine learning methods and models such as context-free grammars (CFG), genetic algorithms (GA), the skip-gram model, and long short-term memory (LSTM). Research about the advantages and disadvantages of these models was conducted and analyzed throughout the process. We concluded that the LSTM model outperforms other approaches in terms of generating lyrics that not only sound grammatically correct but also have meanings and are capable of ma...

Submitted: November 3, 2026Subjects: Peer Reviewed; Research

Description / Details

This research covered various lyric generation using machine learning methods and models such as context-free grammars (CFG), genetic algorithms (GA), the skip-gram model, and long short-term memory (LSTM). Research about the advantages and disadvantages of these models was conducted and analyzed throughout the process. We concluded that the LSTM model outperforms other approaches in terms of generating lyrics that not only sound grammatically correct but also have meanings and are capable of maintaining coherence over extended passages. This advancement will be a great leap forward for all industries, particularly the technology industry.

Source: ScienceOpen - ScienceOpen Research PDF: https://www.scienceopen.com/document_file/c4c24520-1396-4324-8db1-c71e40234a8d/ScienceOpen/sor20260002.pdf Original Link: https://doi.org/10.14293/s2199-1006.1.sor.2026.0002.v1

Please sign in to join the discussion.

No comments yet. Be the first to share your thoughts!