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.
MERL Contacts: Kyeong Jin (K.J.) Kim; Hongbo Sun
Research Areas: Artificial Intelligence, Communications, Machine Learning, Signal Processing, Information SecurityBrief
Date: May 22, 2019
- MERL researchers, Kyeong Jin Kim, Hongbo Sun, Philip Orlik, along with lead author and former MERL intern Siriramya Bhamidipati were awarded the Smart Grid Symposium Best Paper Award at this year's International Conference on Communications (ICC) held in Shanghai, China. There paper titled "GPS Spoofing Detection and Mitigation in PMUs Using Distributed Multiple Directional Antennas," described a technique to rapidly detect and mitigate GPS timing attacks/errors via hardware (antennas) and signal processing (Kalman Filtering)
MERL Contact: Teng-Yok Lee
Research Areas: Artificial Intelligence, Computer Vision, Data Analytics, Machine LearningBrief
Date: April 23, 2019
- MERL researcher Teng-yok Lee has won the Best Visualization Note Award at the PacificVis 2019 conference held in Bangkok Thailand, from April 23-26, 2019. The paper entitled "Space-Time Slicing: Visualizing Object Detector Performance in Driving Video Sequences" presents a visualization method called Space-Time Slicing to assist a human developer in the development of object detectors for driving applications without requiring labeled data. Space-Time Slicing reveals patterns in the detection data that can suggest the presence of false positives and false negatives.
MERL Contacts: Alan Sullivan; Ziming Zhang
Research Areas: Artificial Intelligence, Computer Vision, Machine LearningBrief
Date: November 16, 2018
- Researchers and developers from MERL, Mitsubishi Electric and Mitsubishi Electric Engineering (MEE) have been recognized with an R&D100 award for the development of a deep learning-based water detector. Automatic detection of water levels in rivers and streams is critical for early warning of flash flooding. Existing systems require a height gauge be placed in the river or stream, something that is costly and sometimes impossible. The new deep learning-based water detector uses only images from a video camera along with 3D measurements of the river valley to determine water levels and warn of potential flooding. The system is robust to lighting and weather conditions working well during the night as well as during fog or rain. Deep learning is a relatively new technique that uses neural networks and AI that are trained from real data to perform human-level recognition tasks. This work is powered by Mitsubishi Electric's Maisart AI technology.
See All Awards for Machine Learning
MERL Contacts: Heejin Ahn; Mouhacine Benosman; Karl Berntorp; Ankush Chakrabarty; Claus Danielson; Stefano Di Cairano; Devesh Jha; Rien Quirynen; Yebin Wang; Avishai Weiss
Research Areas: Control, Machine Learning, OptimizationBrief
Date: July 10, 2019 - July 12, 2019
- At the American Control Conference, MERL presented 8 papers on subjects including model predictive control applications, estimation and motion planning for vehicles, modular control architectures, and adaptation and learning.
Speaker: Prof. Jeff Linderoth, University of Wisconsin-Madison
MERL Host: Arvind Raghunathan
Research Areas: Machine Learning, OptimizationBrief
Date & Time: Tuesday, July 16, 2019; 12:00 PM
- Algorithms to solve mixed integer linear programs have made incredible progress in the past 20 years. Key to these advances has been a mathematical analysis of the structure of the set of feasible solutions. We argue that a similar analysis is required in the case of mixed integer quadratic programs, like those that arise in sparse optimization in machine learning. One such analysis leads to the so-called perspective relaxation, which significantly improves solution performance on separable instances. Extensions of the perspective reformulation can lead to algorithms that are equivalent to some of the most popular, modern, sparsity-inducing non-convex regularizations in variable selection. Based on joint work with Hongbo Dong (Washington State Univ. ), Oktay Gunluk (IBM), and Kun Chen (Univ. Connecticut)
See All News & Events for Machine Learning
DA1344: Learning from Demonstration (LfD) for Robotics
MERL is looking for a highly motivated intern to work on developing algorithms for robot learning using learning from demonstration, imitation learning and/or deep reinforcement learning. Successful candidate will collaborate with MERL researchers to design, analyze, and implement new algorithms, conduct experiments, and prepare results for publication. The candidate should have a strong background in (deep) reinforcement learning, Imitation Learning (or Learning from Demonstrations, LfD), machine learning and robotics. Prior experience of working with robotic systems is required. The candidate should be comfortable implementing the developed algorithms in Python and should have prior experience working with ROS. Prior exposure to deep learning and hands-on experience with packages such as Pytorch and/or Tensorflow is expected. The candidate is expected to be a PhD student in Computer Science, Electrical Engineering, Operations Research, Statistics, Applied Mathematics, or a related field, with relevant publication record. Expected duration of the internship is at least 3 months. The position is expected to be available starting late August or early September. Interested candidates are encouraged to apply with their recent CV with list of related publications and links to GitHub repositories (if any).
CV1269: Machine Learning for Computer Vision
MERL is looking for a self-motivated intern to work on machine learning for computer vision. There are several available topics to choose from. The ideal candidate would be a Ph.D. student with a strong background in machine learning and computer vision. Proficiency in Python programming is necessary. You are expected to collaborate with MERL researchers to develop algorithms and prepare manuscripts for scientific publications. Start date is flexible.
CV1354: Multisensor fusion for scene understanding
MERL is looking for a self-motivated intern to work on multisensor fusion for scene understanding. There are several available topics to choose from. The ideal candidate would be a Ph.D. student with a strong background in computer vision and machine learning. Proficiency in Python programming is necessary. You are expected to collaborate with MERL researchers to develop algorithms and prepare manuscripts for scientific publications. Start date is flexible.
See All Internships for Machine Learning
See All Openings at MERL
- "Steering of Autonomous Vehicles Based on Friction-Adaptive Nonlinear Model-Predictive Control", American Control Conference (ACC), July 2019. ,
- "Data-Driven Control Policies for Partially Known Systems via Kernelized Lipschitz Learning", American Control Conference (ACC), July 2019. ,
- "Recursive Bayesian Inference and Learning of Gaussian-Process State-Space Models", European Control Conference (ECC), June 2019. ,
- "Bayesian Tire-Friction Learning by Gaussian-Process State-Space Models", European Control Conference (ECC), June 2019. ,
- "Approximate Dynamic Programming For Linear Systems with State and Input Constraints", European Control Conference (ECC), June 2019. ,
- "Anomaly Detection for Insertion Tasks in Robotic Assembly Using Gaussian Process Models", European Control Conference (ECC), June 2019. ,
- "Audio-Visual Scene-Aware Dialog", IEEE Conference on Computer Vision and Pattern Recognition (CVPR), June 2019. ,
- "Wide-Area GPS Time Monitoring Against Spoofing Using Belief Propagation", IEEE International Conference on Sensing, Communication and Networking , June 2019. ,