Machine Learning
Data-driven approaches to design intelligent algorithms.
MERL has a long history of research activity in machine learning, including the development of various boosting algorithms and contributing to the theory and practice of highly scalable collaborative filtering. Our recent work has focused on deep learning and reinforcement learning, with application to a wide range of applications including automotive, robotics, factory automation, transportation, as well as building and home systems.
Quick Links
-
Researchers
Toshiaki
Koike-Akino
Ye
Wang
Jonathan
Le Roux
Ankush
Chakrabarty
Anoop
Cherian
Gordon
Wichern
Michael J.
Jones
Tim K.
Marks
Stefano
Di Cairano
Philip V.
Orlik
Kieran
Parsons
Pu
(Perry)
WangChristopher R.
Laughman
Daniel N.
Nikovski
Devesh K.
Jha
Diego
Romeres
Chiori
Hori
Bingnan
Wang
Suhas
Lohit
Jing
Liu
Yebin
Wang
Matthew
Brand
Hassan
Mansour
Petros T.
Boufounos
François
Germain
Moitreya
Chatterjee
Kuan-Chuan
Peng
Arvind
Raghunathan
Abraham P.
Vinod
Jianlin
Guo
Vedang M.
Deshpande
Siddarth
Jain
Scott A.
Bortoff
Pedro
Miraldo
Yoshiki
Masuyama
Saviz
Mowlavi
Hongtao
Qiao
William S.
Yerazunis
Radu
Corcodel
Chungwei
Lin
Dehong
Liu
Joshua
Rapp
Hongbo
Sun
Wataru
Tsujita
Wael H.
Ali
Yanting
Ma
Jinyun
Zhang
Purnanand
Elango
Abraham
Goldsmith
Alexander
Schperberg
Avishai
Weiss
Kenji
Inomata
-
Awards
-
AWARD Mitsubishi Electric Team Wins Awards at GalFer Contest Date: June 23, 2025
Awarded to: Bingnan Wang, Tatsuya Yamamoto, Yusuke Sakamoto, Siyuan Sun, Toshiaki Koike-Akino, and Ye Wang
MERL Contacts: Toshiaki Koike-Akino; Bingnan Wang; Ye Wang
Research Areas: Machine Learning, Multi-Physical Modeling, OptimizationBrief- The MELSUR (Mitsubishi Electric SURrogate) team, consisting of a group of MERL and Mitsubishi Electric researchers, ranked first in two out of three categories in the GalFer Contest.
The GalFer (Galileo Ferraris) contest aims to compare the accuracy and efficiency of data-driven methodologies for the multi-physics simulation of traction electric machines. A total of 26 teams worldwide participated in the contest, which consists of three categories. The MELSUR team, including MERL staff Bingnan Wang, Toshiaki Koike-Akino, Ye Wang, MERL intern Siyuan Sun, Mitsubishi Electric researchers Tatsuya Yamamoto and Yusuke Sakamoto, ranked first for the category of "Novelty" and "Interpolation". The results were announced during an award ceremony at the COMPUMAG 2025 conference in Naples, Italy.
- The MELSUR (Mitsubishi Electric SURrogate) team, consisting of a group of MERL and Mitsubishi Electric researchers, ranked first in two out of three categories in the GalFer Contest.
-
AWARD MERL work receives IEEE Transactions on Automation Science and Engineering Best New Application Paper Award from IEEE Robotics and Automation Society Date: May 19, 2025
Awarded to: Yehan Ma, Yebin Wang, Stefano Di Cairano, Toshiaki Koike-Akino, Jianlin Guo, Philip Orlik, Xinping Guan and Chenyang Lu
MERL Contacts: Stefano Di Cairano; Jianlin Guo; Toshiaki Koike-Akino; Philip V. Orlik; Yebin Wang
Research Areas: Communications, Control, Machine LearningBrief- The paper “Smart Actuation for End-Edge Industrial Control Systems”, co-authored by MERL intern Yehan Ma, MERL researchers Yebin Wang, Stefano Di Cairano, Toshiaki Koike-Akino, Jianlin Guo, and Philip Orlik, and academic collaborators Xinping Guan and Chenyang Lu, was recognized as the Best New Application Paper of the IEEE Transactions on Automation Science and Engineering (T-ASE), for "a new industrial automation solution that ensures safety operation through coordinated co-design of edge model predictive control and local actuation".
The award recognizes the best application paper published in T-ASE over the previous calendar year, for the significance of new applications, technical merit, originality, potential impact on the field, and clarity of presentation.
- The paper “Smart Actuation for End-Edge Industrial Control Systems”, co-authored by MERL intern Yehan Ma, MERL researchers Yebin Wang, Stefano Di Cairano, Toshiaki Koike-Akino, Jianlin Guo, and Philip Orlik, and academic collaborators Xinping Guan and Chenyang Lu, was recognized as the Best New Application Paper of the IEEE Transactions on Automation Science and Engineering (T-ASE), for "a new industrial automation solution that ensures safety operation through coordinated co-design of edge model predictive control and local actuation".
-
AWARD MERL Wins Awards at NeurIPS LLM Privacy Challenge Date: December 15, 2024
Awarded to: Jing Liu, Ye Wang, Toshiaki Koike-Akino, Tsunato Nakai, Kento Oonishi, Takuya Higashi
MERL Contacts: Toshiaki Koike-Akino; Jing Liu; Ye Wang
Research Areas: Artificial Intelligence, Machine Learning, Information SecurityBrief- The Mitsubishi Electric Privacy Enhancing Technologies (MEL-PETs) team, consisting of a collaboration of MERL and Mitsubishi Electric researchers, won awards at the NeurIPS 2024 Large Language Model (LLM) Privacy Challenge. In the Blue Team track of the challenge, we won the 3rd Place Award, and in the Red Team track, we won the Special Award for Practical Attack.
See All Awards for Machine Learning -
-
News & Events
-
NEWS MERL contributes to 2025 European Control Conference Date: June 24, 2025 - June 27, 2025
Where: Thessaloniki
MERL Contacts: Stefano Di Cairano; Daniel N. Nikovski; Diego Romeres; Yebin Wang
Research Areas: Control, Dynamical Systems, Machine Learning, Multi-Physical Modeling, Optimization, RoboticsBrief- MERL researchers contributed to both the technical program and workshop organization at the 2025 European Control Conference (ECC), held in Thessaloniki, Greece, from June 24 to 27. ECC is one of the premier conferences in the field of control.
In the main conference, MERL researchers presented four papers covering a range of topics, including: Representation learning, Motion planning for tractor-trailers, Motion planning for mobile manipulators, Learning high-dimensional dynamical systems, Model learning for robotics.
Additionally, MERL co-organized a workshop with the University of Padua titled “Reinforcement Learning for Robotic Control: Recent Developments and Open Challenges.” MERL researcher Diego Romeres also delivered an invited talk titled “Human-Robot Collaborative Assembly” in that workshop.
- MERL researchers contributed to both the technical program and workshop organization at the 2025 European Control Conference (ECC), held in Thessaloniki, Greece, from June 24 to 27. ECC is one of the premier conferences in the field of control.
-
NEWS Toshiaki Koike-Akino to give a tutorial talk at ISIT 2025 Quantum Hackathon Date: June 22, 2025
Where: IEEE International Symposium on Information Theory (ISIT)
MERL Contact: Toshiaki Koike-Akino
Research Areas: Artificial Intelligence, Communications, Data Analytics, Machine Learning, Optimization, Signal Processing, Human-Computer Interaction, Information SecurityBrief- Toshiaki Koike-Akino is invited to present a tutorial talk at IEEE ISIT 2025 Quantum Hackathon, to be held at Ann Arbor, Michigan, USA. The talk, entitled "Emerging Quantum AI Technology", will discuss the recent trends, challenges, and applications of quantum artificial intelligence (QAI) technologies.
The ISIT 2025 Quantum Hackathon invites participants to explore the intersection of quantum computing and information theory. Participants will work with quantum simulators, available quantum hardware, and state-of-the-art development kits to create innovative solutions that connect quantum advancements with challenges in communication and signal processing.
The IEEE International Symposium on Information Theory (ISIT) is the flagship conference of the IEEE Information Theory Society. The symposium centers around the presentation in all of the areas of information theory, including source and channel coding, communication theory and systems, cryptography and security, detection and estimation, networks, pattern recognition and learning, statistics, stochastic processes and complexity, and signal processing.
- Toshiaki Koike-Akino is invited to present a tutorial talk at IEEE ISIT 2025 Quantum Hackathon, to be held at Ann Arbor, Michigan, USA. The talk, entitled "Emerging Quantum AI Technology", will discuss the recent trends, challenges, and applications of quantum artificial intelligence (QAI) technologies.
See All News & Events for Machine Learning -
-
Research Highlights
-
PS-NeuS: A Probability-guided Sampler for Neural Implicit Surface Rendering -
Quantum AI Technology -
TI2V-Zero: Zero-Shot Image Conditioning for Text-to-Video Diffusion Models -
Gear-NeRF: Free-Viewpoint Rendering and Tracking with Motion-Aware Spatio-Temporal Sampling -
Steered Diffusion -
Sustainable AI -
Edge-Assisted Internet of Vehicles for Smart Mobility -
Robust Machine Learning -
mmWave Beam-SNR Fingerprinting (mmBSF) -
Video Anomaly Detection -
Biosignal Processing for Human-Machine Interaction -
MERL Shopping Dataset -
Task-aware Unified Source Separation - Audio Examples
-
-
Internships
-
ST0105: Internship - Surrogate Modeling for Sound Propagation
MERL is seeking a motivated and qualified individual to work on fast surrogate models for sound emission and propagation from complex vibrating structures, with applications in HVAC noise reduction. The ideal candidate will be a PhD student in engineering or related fields with a solid background in frequency-domain acoustic modeling and numerical techniques for partial differential equations (PDEs). Preferred skills include knowledge of the boundary element method (BEM), data-driven modeling, and physics-informed machine learning. Publication of the results obtained during the internship is expected. The duration is expected to be at least 3 months with a flexible start date.
-
OR0163: Reinforcement/Imitation Learning for Robotic Assembly
MERL is looking for a highly motivated individual to work on learning algorithms for generalist robotic manipulation agent. In this internship we will be working on complex, assembly tasks. The research will develop novel learning-based algorithms for assembly tasks such as tight-tolerance insertion. The ideal candidate should have experience in either one or multiple of the following topics: Deep Reinforcement learning, Diffusion Models/Policies and Force Control for Manipulation. Senior PhD students in machine learning and engineering with a focus on Reinforcement/Imitation Learning and Manipulation are encouraged to apply. Prior experience working with physics engines like Mujoco, Isaac Gym, etc. is required. Prior experience working with physical robots as well vision and tactile sensors is required. The developed algorithm needs to be implemented on the real system. A successful internship will result in submission of results to peer-reviewed conference and journals. Good coding skills in Python and state-of-the-art RL environments (e.g., RL Bench) is required. The expected duration of internship is 3-4 months with expected start date in Sept 2025. This internship is preferred to be onsite at MERL.
Required Specific Experience
- Experience working with physical robot and sensors like cameras, tactile and F/T is a must.
- Prior experience with learning algorithms is a must.
-
MS0156: Internship - Stochastic Model Predictive Control with Generative Models for Smart Building Control
MERL is looking for a research intern to develop efficient transformer-informed stochastic MPC to control net-zero energy buildings. This is an exciting opportunity to make a real impact in the field of cutting-edge deep learning and predictive control on a real system. Publication of results produced during the internship is desired. The expected duration of the internship is 3-6 months with a flexible start date.
The Ideal Candidate Will Have:
- Significant hands-on experience with stochastic MPC
- Publications in SMPC are a strong plus
- Fluency in Python and PyTorch
- Understanding of probabilistic time-series prediction
- Experience with convex programming
- Convex formulations of MPC/SMPC problems are a strong plus
- Completed their MS, or >50% of their PhD program
Mitsubishi Electric Research Labs, Inc. "MERL" provides equal employment opportunities (EEO) to all employees and applicants for employment without regard to race, color, religion, sex, national origin, age, disability or genetics. In addition to federal law requirements, MERL complies with applicable state and local laws governing nondiscrimination in employment in every location in which the company has facilities. This policy applies to all terms and conditions of employment, including recruiting, hiring, placement, promotion, termination, layoff, recall, transfer, leaves of absence, compensation and training.
MERL expressly prohibits any form of workplace harassment based on race, color, religion, gender, sexual orientation, gender identity or expression, national origin, age, genetic information, disability, or veteran status. Improper interference with the ability of MERL’s employees to perform their job duties may result in discipline up to and including discharge.
Working at MERL requires full authorization to work in the U.S and access to technology, software and other information that is subject to governmental access control restrictions, due to export controls. Employment is conditioned on continued full authorization to work in the U.S and the availability of government authorization for the release of these items, which might include without limitation, obtaining an export license or other documentation. MERL may delay commencement of employment, rescind an offer of employment, terminate employment, and/or modify job responsibilities, compensation, benefits, and/or access to MERL facilities and information systems, as MERL deems appropriate, to ensure practical compliance with applicable employment law and government access control restrictions.
- Significant hands-on experience with stochastic MPC
See All Internships for Machine Learning -
-
Openings
-
CA0093: Research Scientist - Control for Autonomous Systems
-
EA0042: Research Scientist - Control & Learning
-
CI0130: Postdoctoral Research Fellow - Artificial General Intelligence (AGI)
See All Openings at MERL -
-
Recent Publications
- "Aligning Multimodal Representations through an Information Bottleneck", International Conference on Machine Learning (ICML), July 2025.BibTeX TR2025-109 PDF
- @inproceedings{Almudévar2025jul,
- author = {Almudévar, Antonio and Hernández-Lobato, José, M and Khurana, Sameer and Marxer, Ricard and Ortega, Alfonso},
- title = {{Aligning Multimodal Representations through an Information Bottleneck}},
- booktitle = {International Conference on Machine Learning (ICML)},
- year = 2025,
- month = jul,
- url = {https://www.merl.com/publications/TR2025-109}
- }
, - "Navigating the Trade-offs and Synergies of Economic and Environmental Sustainability Using Process Systems Engineering", American Control Conference (ACC), July 2025.BibTeX TR2025-106 PDF
- @inproceedings{Zhang2025jul2,
- author = {Zhang, Qi and Avraamidou, Styliani and Paulson, Joel A. and Thakkar, Vyom and Wang, Zhenyu and Chiang, Leo and Braun, Birgit and Rathi, Tushar and Chakrabarty, Ankush and Sorouifar, Farshud and Tang, Wei-Ting and Guertin, France and Munoz, Paola and Sampat, Apoorva},
- title = {{Navigating the Trade-offs and Synergies of Economic and Environmental Sustainability Using Process Systems Engineering}},
- booktitle = {American Control Conference (ACC)},
- year = 2025,
- month = jul,
- url = {https://www.merl.com/publications/TR2025-106}
- }
, - "Policy Optimization for PDE Control with a Warm Start", American Control Conference (ACC), July 2025.BibTeX TR2025-105 PDF
- @inproceedings{Zhang2025jul,
- author = {Zhang, Xiangyuan and Mowlavi, Saviz and Benosman, Mouhacine and Basar, Tamer},
- title = {{Policy Optimization for PDE Control with a Warm Start}},
- booktitle = {American Control Conference (ACC)},
- year = 2025,
- month = jul,
- url = {https://www.merl.com/publications/TR2025-105}
- }
, - "Safe Interactive Motion Planning by Differentiable Optimal Control and Online Preference Learning", American Control Conference (ACC), July 2025.BibTeX TR2025-104 PDF
- @inproceedings{ChavezArmijos2025jul,
- author = {Chavez Armijos, Andres and Berntorp, Karl and {Di Cairano}, Stefano},
- title = {{Safe Interactive Motion Planning by Differentiable Optimal Control and Online Preference Learning}},
- booktitle = {American Control Conference (ACC)},
- year = 2025,
- month = jul,
- url = {https://www.merl.com/publications/TR2025-104}
- }
, - "Machine Learning-Powered Radio Frequency Sensing: A Review", IEEE Sensors Journal, June 2025.BibTeX TR2025-099 PDF
- @article{Santra2025jun,
- author = {Santra, Avik and Wang, Pu and Shaker, George and Mysore, Bhavani Shankar and Dolmans, Guido and Chen, Yan and Shariati, Negin and Pandharipande, Ashish},
- title = {{Machine Learning-Powered Radio Frequency Sensing: A Review}},
- journal = {IEEE Sensors Journal},
- year = 2025,
- month = jun,
- url = {https://www.merl.com/publications/TR2025-099}
- }
, - "A Hierarchical Approach for Tractor-trailer Motion Planning Using Graph Search and Reinforcement Learning", European Control Conference (ECC), June 2025.BibTeX TR2025-093 PDF
- @inproceedings{Ma2025jun,
- author = {Ma, Haitong and Zhang, Tianpeng and Li, Na and {Di Cairano}, Stefano and Wang, Yebin},
- title = {{A Hierarchical Approach for Tractor-trailer Motion Planning Using Graph Search and Reinforcement Learning}},
- booktitle = {European Control Conference (ECC)},
- year = 2025,
- month = jun,
- url = {https://www.merl.com/publications/TR2025-093}
- }
, - "Quantum Diffusion Models for Few-Shot Learning", ICAD, June 2025.BibTeX TR2025-095 PDF
- @inproceedings{Wang2025jun2,
- author = {Wang, Ruhan and Wang, Ye and Liu, Jing and Koike-Akino, Toshiaki},
- title = {{Quantum Diffusion Models for Few-Shot Learning}},
- booktitle = {ICAD},
- year = 2025,
- month = jun,
- url = {https://www.merl.com/publications/TR2025-095}
- }
, - "Image-based Deep Learning Models for Electric Motors", International Conference on the Computation of Electromagnetic Fields (COMPUMAG), June 2025.BibTeX TR2025-088 PDF
- @inproceedings{Sun2025jun,
- author = {Sun, Siyuan and Wang, Ye and Koike-Akino, Toshiaki and Yamamoto, Tatsuya and Sakamoto, Yusuke and Wang, Bingnan},
- title = {{Image-based Deep Learning Models for Electric Motors}},
- booktitle = {International Conference on the Computation of Electromagnetic Fields (COMPUMAG)},
- year = 2025,
- month = jun,
- url = {https://www.merl.com/publications/TR2025-088}
- }
,
- "Aligning Multimodal Representations through an Information Bottleneck", International Conference on Machine Learning (ICML), July 2025.
-
Videos
-
Software & Data Downloads
-
MEL-PETs Joint-Context Attack for LLM Privacy Challenge -
Long-Tailed Online Anomaly Detection dataset -
Generalization in Deep RL with a Robust Adaptation Module -
MEL-PETs Defense for LLM Privacy Challenge -
Learned Born Operator for Reflection Tomographic Imaging -
Group Representation Networks -
Stabilizing Subject Transfer in EEG Classification with Divergence Estimation -
Retrieval-Augmented Neural Field for HRTF Upsampling and Personalization -
ComplexVAD Dataset -
Self-Monitored Inference-Time INtervention for Generative Music Transformers -
Radar dEtection TRansformer -
Millimeter-wave Multi-View Radar Dataset -
Gear Extensions of Neural Radiance Fields -
Long-Tailed Anomaly Detection Dataset -
Target-Speaker SEParation -
Pixel-Grounded Prototypical Part Networks -
Steered Diffusion -
BAyesian Network for adaptive SAmple Consensus -
Meta-Learning State Space Models -
Explainable Video Anomaly Localization -
Simple Multimodal Algorithmic Reasoning Task Dataset -
Partial Group Convolutional Neural Networks -
SOurce-free Cross-modal KnowledgE Transfer -
Audio-Visual-Language Embodied Navigation in 3D Environments -
Nonparametric Score Estimators -
3D MOrphable STyleGAN -
Instance Segmentation GAN -
Audio Visual Scene-Graph Segmentor -
Generalized One-class Discriminative Subspaces -
Hierarchical Musical Instrument Separation -
Generating Visual Dynamics from Sound and Context -
Adversarially-Contrastive Optimal Transport -
Online Feature Extractor Network -
MotionNet -
FoldingNet++ -
Quasi-Newton Trust Region Policy Optimization -
Landmarks’ Location, Uncertainty, and Visibility Likelihood -
Robust Iterative Data Estimation -
Gradient-based Nikaido-Isoda -
Circular Maze Environment -
Discriminative Subspace Pooling -
Kernel Correlation Network -
Fast Resampling on Point Clouds via Graphs -
FoldingNet -
Deep Category-Aware Semantic Edge Detection -
MERL Shopping Dataset
-