Diego Romeres
- Phone: 617-621-7561
- Email:
-
Position:
Research / Technical Staff
Principal Research Scientist -
Education:
Ph.D., University of Padova, 2017 -
Research Areas:
External Links:
Diego's Quick Links
-
Biography
Diego's research interests are in machine learning, system identification and robotic applications. At MERL he is currently working on applying nonparametric machine learning techniques for the control of robotic platforms. His Ph.D. thesis is about the combination of nonparametric data-driven models and physics-based models in gaussian processes for robot dynamics learning.
-
Recent News & Events
-
NEWS New robotics benchmark system Date: November 16, 2020
MERL Contacts: Devesh Jha; Daniel Nikovski; Diego Romeres; Alan Sullivan; Jeroen van Baar
Research Areas: Artificial Intelligence, Machine Learning, RoboticsBrief- MERL researchers, in collaboration with researchers from MELCO and the Department of Brain and Cognitive Science at MIT, have released simulation software Circular Maze Environment (CME). This system could be used as a new benchmark for evaluating different control and robot learning algorithms. The control objective in this system is to tip and the tilt the maze so as to drive one (or multiple) marble(s) to the innermost ring of the circular maze. Although the system is very intuitive for humans to control, it is very challenging for artificial intelligence agents to learn efficiently. It poses several challenges for both model-based as well as model-free methods, due to its non-smooth dynamics, long planning horizon, and non-linear dynamics. The released Python package provides the simulation environment for the circular maze, where movement of multiple marbles could be simulated simultaneously. The package also provides a trajectory optimization algorithm to design a model-based controller in simulation.
-
NEWS Diego Romeres serves on the Programme Committee for the Conference on Innovative Applications of Artificial Intelligence, 2021. Date: February 4, 2021
Where: N/A
MERL Contact: Diego Romeres
Research Areas: Artificial Intelligence, Data Analytics, Machine LearningBrief- Dr. Diego Romeres, Principal Research Scientist in the Data Analytics group, will serve on the Programme Committee for the Thirty-Third Annual Conference on Innovative Applications of Artificial Intelligence (IAAI), 2021.
See All News & Events for Diego -
-
Internships with Diego
-
DA1533: Machine Learning for Robotic Manipulation
MERL is looking for a self-motivated and qualified candidate to work on robotic manipulation projects. The ideal candidate is a PhD student and should have experience and records in multiple of the following areas. Machine learning techniques for modeling and control such as Gaussian Processes and Neural Networks. Knowledge of standard Reinforcement Learning algorithms. Experience in working with robotic systems and familiarity with one physics engine simulator like Mujoco, pyBullet, pyDrake. Proficiency in Python is required. The successful candidate will be expected to develop, in collaboration with MERL employees, state of the art algorithms to solve complex robotic manipulation tasks that will lead to a scientific publication. Typical internship length is 3-4 months. This internship is preferred to be onsite at MERL, but may be done remotely where you live if the COVID pandemic makes it necessary.
-
-
MERL Publications
- , "Towards Human-Level Learning of Complex Physical Puzzles", arXiv, December 2020.BibTeX
- @article{Ota2020dec,
- author = {Ota, Kei and Jha, Devesh and Romeres, Diego and van Baar, Jeroen and Smith, Kevin and Semistsu, Takayuki and Oiki, Tomoaki and Sullivan, Alan and Nikovski, Daniel N. and Tenanbaum, Joshua},
- title = {Towards Human-Level Learning of Complex Physical Puzzles},
- journal = {arXiv},
- year = 2020,
- month = dec
- }
- , "Model-based Policy Search for Partially Measurable Systems", Advances in Neural Information Processing Systems (NeurIPS), December 2020.BibTeX TR2020-174 PDF
- @inproceedings{Romeres2020dec2,
- author = {Romeres, Diego and Amadio, Fabio and Dalla Libera, Alberto and Nikovski, Daniel N. and Carli, Ruggero},
- title = {Model-based Policy Search for Partially Measurable Systems},
- booktitle = {Advances in Neural Information Processing Systems (NeurIPS)},
- year = 2020,
- month = dec,
- url = {https://www.merl.com/publications/TR2020-174}
- }
- , "Feedback Linearization Robot Control based on Gaussian Process Inverse Dynamics Model", Conferenza Italiana di Robotica e Macchine Intelligenti, December 2020.BibTeX TR2020-173 PDF
- @inproceedings{Romeres2020dec,
- author = {Romeres, Diego and Dalla Libera, Alberto and Amadio, Fabio and Carli, Ruggero},
- title = {Feedback Linearization Robot Control based on Gaussian Process Inverse Dynamics Model},
- booktitle = {Conferenza Italiana di Robotica e Macchine Intelligenti},
- year = 2020,
- month = dec,
- url = {https://www.merl.com/publications/TR2020-173}
- }
- , "Tactile-RL for Insertion: Generalization to Objects of Unknown Geometry", arXiv, November 2020.BibTeX
- @article{Dong2020nov,
- author = {Dong, Siyuan and Jha, Devesh and Romeres, Diego and Kim, Sangwoon and Nikovski, Daniel N. and Rodriguez, Alberto},
- title = {Tactile-RL for Insertion: Generalization to Objects of Unknown Geometry},
- journal = {arXiv},
- year = 2020,
- month = nov
- }
- , "Understanding Multi-Modal Perception Using Behavioral Cloning for Peg-In-a-Hole Insertion Tasks", Robotics: Science and Systems, July 2020.BibTeX TR2020-110 PDF
- @inproceedings{Romeres2020jul,
- author = {Romeres, Diego and Liu, Yifang and Jha, Devesh and Nikovski, Daniel N.},
- title = {Understanding Multi-Modal Perception Using Behavioral Cloning for Peg-In-a-Hole Insertion Tasks},
- booktitle = {Robotics: Science and Systems},
- year = 2020,
- month = jul,
- url = {https://www.merl.com/publications/TR2020-110}
- }
- , "Towards Human-Level Learning of Complex Physical Puzzles", arXiv, December 2020.
-
Other Publications
- , "On-line bayesian system identification", Control Conference (ECC), 2016 European, 2016, pp. 1359-1364.BibTeX
- @Inproceedings{romeres2016line,
- author = {Romeres, Diego and Prando, Giulia and Pillonetto, Gianluigi and Chiuso, Alessandro},
- title = {On-line bayesian system identification},
- booktitle = {Control Conference (ECC), 2016 European},
- year = 2016,
- pages = {1359--1364},
- organization = {IEEE}
- }
- , "Online semi-parametric learning for inverse dynamics modeling", Decision and Control (CDC), 2016 IEEE 55th Conference on, 2016, pp. 2945-2950.BibTeX
- @Inproceedings{romeres2016online,
- author = {Romeres, Diego and Zorzi, Mattia and Camoriano, Raffaello and Chiuso, Alessandro},
- title = {Online semi-parametric learning for inverse dynamics modeling},
- booktitle = {Decision and Control (CDC), 2016 IEEE 55th Conference on},
- year = 2016,
- pages = {2945--2950},
- organization = {IEEE}
- }
- , "Online semi-parametric learning for inverse dynamics modeling", Decision and Control (CDC), 2016 IEEE 55th Conference on, 2016, pp. 2945-2950.BibTeX
- @Inproceedings{romeres2016onlinesemiparametric,
- author = {Romeres, Diego and Zorzi, Mattia and Camoriano, Raffaello and Chiuso, Alessandro},
- title = {Online semi-parametric learning for inverse dynamics modeling},
- booktitle = {Decision and Control (CDC), 2016 IEEE 55th Conference on},
- year = 2016,
- pages = {2945--2950},
- organization = {IEEE}
- }
- , "Classical vs. Bayesian methods for linear system identification: Point estimators and confidence sets", Control Conference (ECC), 2016 European, 2016, pp. 1365-1370.BibTeX
- @Inproceedings{tprando2016classical,
- author = {Prando, Giulia and Romeres, Diego and Pillonetto, Gianluigi and Chiuso, Alessandro},
- title = {Classical vs. Bayesian methods for linear system identification: Point estimators and confidence sets},
- booktitle = {Control Conference (ECC), 2016 European},
- year = 2016,
- pages = {1365--1370},
- organization = {IEEE}
- }
- , "Online identification of time-varying systems: A Bayesian approach", Decision and Control (CDC), 2016 IEEE 55th Conference on, 2016, pp. 3775-3780.BibTeX
- @Inproceedings{tprando2016online,
- author = {Prando, Giulia and Romeres, Diego and Chiuso, Alessandro},
- title = {Online identification of time-varying systems: A Bayesian approach},
- booktitle = {Decision and Control (CDC), 2016 IEEE 55th Conference on},
- year = 2016,
- pages = {3775--3780},
- organization = {IEEE}
- }
- , "Region of attraction of power systems", IFAC Proceedings Volumes, Vol. 46, No. 27, pp. 49-54, 2013.BibTeX
- @Article{munz2013region,
- author = {Muenz, Ulrich and Romeres, Diego},
- title = {Region of attraction of power systems},
- journal = {IFAC Proceedings Volumes},
- year = 2013,
- volume = 46,
- number = 27,
- pages = {49--54},
- publisher = {Elsevier}
- }
- , "Novel results on slow coherency in consensus and power networks", Control Conference (ECC), 2013 European, 2013, pp. 742-747.BibTeX
- @Inproceedings{romeres2013novel,
- author = {Romeres, Diego and Doerfler, Florian and Bullo, Francesco},
- title = {Novel results on slow coherency in consensus and power networks},
- booktitle = {Control Conference (ECC), 2013 European},
- year = 2013,
- pages = {742--747},
- organization = {IEEE}
- }
- , "Distributed multi-hop reactive power compensation in smart micro-grids subject to saturation constraints", Decision and Control (CDC), 2012 IEEE 51st Annual Conference on, 2012, pp. 1118-1123.BibTeX
- @Inproceedings{bolognani2012distributed,
- author = {Bolognani, Saverio and Carron, Andrea and Di Vittorio, Alberto and Romeres, Diego and Schenato, Luca and Zampieri, Sandro},
- title = {Distributed multi-hop reactive power compensation in smart micro-grids subject to saturation constraints},
- booktitle = {Decision and Control (CDC), 2012 IEEE 51st Annual Conference on},
- year = 2012,
- pages = {1118--1123},
- organization = {IEEE}
- }
- , "On-line bayesian system identification", Control Conference (ECC), 2016 European, 2016, pp. 1359-1364.
-
Software Downloads
-
Videos
