{"id":1336,"date":"2022-03-28T20:28:30","date_gmt":"2022-03-28T20:28:30","guid":{"rendered":"https:\/\/elo-x.eu\/?p=1336"},"modified":"2023-07-24T12:15:50","modified_gmt":"2023-07-24T12:15:50","slug":"flavia-acerbo","status":"publish","type":"post","link":"https:\/\/elo-x.eu\/?p=1336","title":{"rendered":"Flavia Acerbo"},"content":{"rendered":"\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\n\t\t
\n\n\t\t\t\t\t\t\t\t

\n\t\t\t\t\t\t\t\t\n\t\t\t\tFlavia Acerbo \n\t\t\t<\/h2 > \n\t\t\t\t\t<\/div>\n\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t
\n\t\t\t\n\t\t\t\t\t\t<\/span>\n\t\t<\/div>\n\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\tAssociate Research Engineer, PhD Candidate at KU Leuven<\/span><\/span>\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t
\n\nSiemens Digital Industries Software<\/b><\/span>\n\n<\/div>\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\"\"\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t

Flavia Sofia Acerbo is a researcher at Siemens Digital Industries Software, based in Leuven. She obtained her Bachelor\u2019s degree in Computer Engineering in 2017 and her Master\u2019s degree in Mechatronic Engineering in 2019, both cum laude and from Politecnico di Torino. After graduating, she joined Siemens Digital Industries Software, where she is currently working on a PhD project with KU Leuven on human imitation for autonomous driving comfort, which tries to develop methodologies to learn human-like controllers from driving data to improve passenger perception in automated vehicles. Her research interests lie in the intersection between optimal control and machine learning, specifically around learning-based Model Predictive Control and Reinforcement Learning.<\/p>\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t

\n\t\t\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t
\n\t\t\t\n\t\t\t\t\t\t<\/span>\n\t\t<\/div>\n\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t

Project description<\/h2>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t

\u00a0\u00a0<\/span><\/span>The PhD project aims at developing a methodology to systematically include a human-like driving style into Advanced Driver Assistance Systems (ADAS) development, providing both safe and personalized driving. The study relies on demonstrations performed by the considered user and mixed-real virtual design and testing. The approach should be easily adapted to various automated driving tasks with minimal use of domain knowledge, automatically inferring the key parameters that describe the demonstrated driving style from high-dimensional multi-source datasets.<\/p>\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t

\n\t\t\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t
\n\t\t\t\n\t\t\t\t\t\t<\/span>\n\t\t<\/div>\n\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t

Publications<\/h2>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t
<\/a>
<\/div><\/div><\/form>
1.<\/div>

Wang, Renzi; Acerbo, Flavia Sofia; Son, Tong Duy; Patrinos, Panagiotis<\/p>

Imitation Learning from Observations: An Autoregressive Mixture of Experts Approach<\/a> Working paper<\/span> <\/p>

2024<\/span>.<\/p>

Abstract<\/a><\/span> | Links<\/a><\/span> | BibTeX<\/a><\/span><\/p>

@workingpaper{wang2024imitation,
\r\ntitle = {Imitation Learning from Observations: An Autoregressive Mixture of Experts Approach},
\r\nauthor = {Renzi Wang and Flavia Sofia Acerbo and Tong Duy Son and Panagiotis Patrinos},
\r\nurl = {https:\/\/doi.org\/10.48550\/arXiv.2411.08232
\r\n},
\r\nyear  = {2024},
\r\ndate = {2024-11-12},
\r\nabstract = {This paper presents a novel approach to imitation learning from observations, where an autoregressive mixture of experts model is deployed to fit the underlying policy. The parameters of the model are learned via a two-stage framework. By leveraging the existing dynamics knowledge, the first stage of the framework estimates the control input sequences and hence reduces the problem complexity. At the second stage, the policy is learned by solving a regularized maximum-likelihood estimation problem using the estimated control input sequences. We further extend the learning procedure by incorporating a Lyapunov stability constraint to ensure asymptotic stability of the identified model, for accurate multi-step predictions. The effectiveness of the proposed framework is validated using two autonomous driving datasets collected from human demonstrations, demonstrating its practical applicability in modelling complex nonlinear dynamics.},
\r\nkeywords = {},
\r\npubstate = {published},
\r\ntppubtype = {workingpaper}
\r\n}
\r\n<\/pre><\/div>
Close<\/a><\/p><\/div>
This paper presents a novel approach to imitation learning from observations, where an autoregressive mixture of experts model is deployed to fit the underlying policy. The parameters of the model are learned via a two-stage framework. By leveraging the existing dynamics knowledge, the first stage of the framework estimates the control input sequences and hence reduces the problem complexity. At the second stage, the policy is learned by solving a regularized maximum-likelihood estimation problem using the estimated control input sequences. We further extend the learning procedure by incorporating a Lyapunov stability constraint to ensure asymptotic stability of the identified model, for accurate multi-step predictions. The effectiveness of the proposed framework is validated using two autonomous driving datasets collected from human demonstrations, demonstrating its practical applicability in modelling complex nonlinear dynamics.<\/div>
Close<\/a><\/p><\/div>