MoDiPO: text-to-motion alignment via AI-feedback-driven Direct Preference Optimization
Authors:
Massimiliano Pappa,
Luca Collorone,
Giovanni Ficarra,
Indro Spinelli,
Fabio Galasso
Abstract:
Diffusion Models have revolutionized the field of human motion generation by offering exceptional generation quality and fine-grained controllability through natural language conditioning. Their inherent stochasticity, that is the ability to generate various outputs from a single input, is key to their success. However, this diversity should not be unrestricted, as it may lead to unlikely generati…
▽ More
Diffusion Models have revolutionized the field of human motion generation by offering exceptional generation quality and fine-grained controllability through natural language conditioning. Their inherent stochasticity, that is the ability to generate various outputs from a single input, is key to their success. However, this diversity should not be unrestricted, as it may lead to unlikely generations. Instead, it should be confined within the boundaries of text-aligned and realistic generations. To address this issue, we propose MoDiPO (Motion Diffusion DPO), a novel methodology that leverages Direct Preference Optimization (DPO) to align text-to-motion models. We streamline the laborious and expensive process of gathering human preferences needed in DPO by leveraging AI feedback instead. This enables us to experiment with novel DPO strategies, using both online and offline generated motion-preference pairs. To foster future research we contribute with a motion-preference dataset which we dub Pick-a-Move. We demonstrate, both qualitatively and quantitatively, that our proposed method yields significantly more realistic motions. In particular, MoDiPO substantially improves Frechet Inception Distance (FID) while retaining the same RPrecision and Multi-Modality performances.
△ Less
Submitted 6 May, 2024;
originally announced May 2024.
AI-based Data Preparation and Data Analytics in Healthcare: The Case of Diabetes
Authors:
Marianna Maranghi,
Aris Anagnostopoulos,
Irene Cannistraci,
Ioannis Chatzigiannakis,
Federico Croce,
Giulia Di Teodoro,
Michele Gentile,
Giorgio Grani,
Maurizio Lenzerini,
Stefano Leonardi,
Andrea Mastropietro,
Laura Palagi,
Massimiliano Pappa,
Riccardo Rosati,
Riccardo Valentini,
Paola Velardi
Abstract:
The Associazione Medici Diabetologi (AMD) collects and manages one of the largest worldwide-available collections of diabetic patient records, also known as the AMD database. This paper presents the initial results of an ongoing project whose focus is the application of Artificial Intelligence and Machine Learning techniques for conceptualizing, cleaning, and analyzing such an important and valuab…
▽ More
The Associazione Medici Diabetologi (AMD) collects and manages one of the largest worldwide-available collections of diabetic patient records, also known as the AMD database. This paper presents the initial results of an ongoing project whose focus is the application of Artificial Intelligence and Machine Learning techniques for conceptualizing, cleaning, and analyzing such an important and valuable dataset, with the goal of providing predictive insights to better support diabetologists in their diagnostic and therapeutic choices.
△ Less
Submitted 20 July, 2022; v1 submitted 13 June, 2022;
originally announced June 2022.
