Title: Music Plagiarism Detection

Olesia Omelchuk, Audio AI Engineer at It-Jim

Abstract:

This talk will explore the topic of plagiarism detection in the music industry. We will begin by outlining different types of music plagiarism, along with real-world cases. To establish a foundation for further discussion, we will review modern approaches to music representation learning, focusing on popular encoder models such as MuQ, MERT, and MusicFM, as well as recent trends in training. Building on this foundation, we will return to the core topic by examining existing solutions for music plagiarism detection: from rule-based systems like MIPPIA to deep learning approaches such as MelodySim. 

Finally, practical findings from our It-Jim R&D Music Lab will be presented, showcasing how combining the strengths of existing approaches can lead to more robust and accurate plagiarism detection. These findings also include an implemented data augmentation pipeline, which is designed to expand typically limited open-source training datasets and improve model generalization.

Bio:

Olesia Omelchuk is an Audio AI Engineer at It-Jim, specializing in audio processing and machine learning applications for music and speech. Last year, she graduated with a Bachelor’s degree in Computer Science from the Ukrainian Catholic University. For her thesis, “Ukrainian Music and Its Place in a Genre Map,” she developed an audio representation learning pipeline designed to capture genre-specific characteristics without relying on manual labels. She also completed a Mitacs research internship in Canada, where she worked on the Document Dewarping problem, resulting in a paper accepted to the IEEE CAI 2026 conference.

About the company:

It‑Jim is a team of engineers and researchers in audio processing, 2D/3D CV, AI‑powered mobile, and NLP. With over ten years of experience, we have delivered hundreds of effective AI solutions in music, medicine, biotech, robotics, retail, arts, and sports. Our scientific expertise and genuine interest in AI have shaped a workplace where exciting R&D and problem‑solving are everyday practice.

Activity supported by:

Cátedra UPF-BMAT en Inteligencia Artificial y Música (TSI-100929-2023-1). Project funded by Secretaría de Estado de Digitalización e Inteligencia Artificial, the European Union-Next Generation EU, and by BMAT Music Innovators, the Music Operating System