Lean AI’s solution addresses these challenges by introducing a game-changing technology that allows fast an easy deployment, robust to process and environmental variations, and scalable to new camera angles, defects, surfaces and products.
Lean AI’s unsupervised, self-learning technology only requires access to a non-labeled production data, which is available in abundance and requires no manual tagging. This is done by utilizing unsupervised learning processes that are similar to human brain. Moreover, this technology is robust to process variation due its elasticity and online learning capabilities, that will in real-time adapt the model to any process or environmental change. This proprietary signatures architecture requires far less computing power than the traditional deep learning systems and can run on low-processing platforms and will not require a monstrous HW installation.
- Predictive quality
- Easy deployment- unsupervised AI
- No maintenance-
robust to process variation
Technology
We’ve developed the next generation of autonomous Al that enables machines to think. Our unique approach operates in a fundamentally different way from traditional deep learning systems that are mainly deployed in the market today. It is based on a computational paradigm shift, a product of multi-disciplinary experimental research and modelling of the structure and function of the mammal neocortex. This unique unsupervised approach to learning mimics the way the brain processes information. It’s why our Al system is closer to human intelligence than deep learning.
Our technology is robust and generic and applicable within a multitude of signal domains such visual, audio, time series and other domains; visual inspection is only the beginning.
Autonomous AI is a mature and proven technology, protected by more than 250 patents and patent applications, and successfully deployed in other verticals such as automotive, security, fintech and medical.
Proven technological superiority in a range of areas
-
Unsupervised
Unlike Deep learning, Cortica’s AI’s Autonomous AI is capable of learning without manually labeled training sets, especially without defect examples.
-
Generic
Like the human Cortex, it can handle any type of signal (visual, audio, radar, time series, etc.). handle both visual and non-visual signals such as audio and time series.
-
Robust
robustness to environmental conditions such as lighting, angles, occlusions,
-
Adaptive
Keeps learning and continuously improving itself.
Product
Lean.AI develops an open and flexible platform solution for
- Easy Deployment – The features described below drastically reduce the deployment time of AI-based QA systems:
- No need in gathering defective parts for training set
- No Need in data scientist or support – all done by production line operator
- Generic – works of various types of surfaces and parts
- Accurate, reliable and scalable
- No maintenance- doesn’t require constant tuning, uses online learning to adapt itself to variations in the process and environment
- SW solution–doesn’t require purchasing expensive inspection equipment, nor requires a camera from a specific vendor or model. Equipment agnostic, robust and flexible.
- High Throughput–light and flat architecture allows performing the processing without delaying the production line while the learning process is happening in parallel in the background
About
Lean.Al aims to revolutionize and refine the ways in which quality control is achieved. Our vision is to enable ideal quality in complex manufacturing processes through unsupervised self-learning Al technology. providing accurate detection and predictive process control.
Our Mission
Johnson Electric Group (JE) and Cortica Group have embarked on a joint venture to build an Autonomous Inspection System that is revolutionizing the Quality Inspection Market. Leveraging Cortica Group’s Autonomous AI Technology and JE’s vast knowledge and experience in manufacturing processes, the partnership is solving an unsolved problem of today.
Cortica Group has developed Autonomous AI that simulates the natural processes of the mammal cortex. Its unsupervised approach to learning mimics the way the brain processes information: enabling machines to learn, collaborate and interact with the world without human input. The technology is backed by more than 250 patents and patent applications and 10 years of cutting-edge research and is revolutionizing visual intelligence for a smarter and safer world.
The Johnson Electric Group is a global leader in electric motors, actuators, motion subsystems and related electro-mechanical components. It serves a broad range of industries including Automotive, Smart Metering, Medical Devices, Business Equipment, Home Automation, Ventilation, White Goods, Power Tools, and Lawn & Garden Equipment. The Group is headquartered in Hong Kong and employs over 35,000 individuals in 22 countries worldwide. Johnson Electric Holdings Limited is listed on The Stock Exchange of Hong Kong Limited (Stock Code: 179).
Contact
Want to further explore machine learning? Let’s chat.