Making AI tiny
and accurate.
Finally.
Radically more efficient deep learning for ground-breaking AI features at the edge.
Radically more efficient deep learning for ground-breaking AI features at the edge.
We are working on a future in which your doorbell tells you the kids got home safely. Your front door automatically opens while you carry the groceries. And the savannah warns the ranger of poachers. It’s a future that is safer and more thoughtful thanks to intelligent sensors inside all the things around us.
Yesterday,
you had to sit at your PC.
Today, you take your
smartphone with you.
Tomorrow, intelligent sensors
turn computing ambient.
A large number of intelligent chips are required to make computing ambient. Therefore, they have to be small, cheap, and low-power. Our technology works on $1 chips with 1MB of memory. These are so efficient, they can be powered by a coin battery or small solar cell.
Plumerai™ has developed a complete software solution that includes familiar face identification, stranger identification, people detection, vehicle detection, animal detection, advanced motion detection, and more. The AI software is deployed on all major camera SOC and cloud platforms and on millions of cameras in the field. To ensure safe and responsible deployment in the US, EU, and UK, the software is compliant with local privacy laws, including GDPR, CCPA and BIPA.
Traditional smart home cameras generate many unnecessary alerts triggered by moving objects or changes in lighting. Plumerai's AI solution eliminates these false alarms, ensuring users receive only relevant messages. Tag known safe individuals, such as the weekly gardener, to avoid unnecessary notifications. Precise messages like "Stranger on the driveway" or "Rachel came home" keep users informed and engaged. This encourages continued use of the app, enhances the overall experience, and provides peace of mind.
We combine our optimized inference engine with our collection of tiny AI models to provide turnkey software solutions. These are highly accurate and so efficient that they run on nearly every off-the-shelf chip.
We develop our neural networks from scratch. We optimize for small embedded devices with customized model architectures and training strategies, based on our world-class research on model quantization. This results in tiny but highly accurate AI models.
Developers use the Plumerai Inference Engine to make AI models run efficiently on CPUs, with minimal compute and memory. The inference engine is optimized for Arm Cortex-M, Arm Cortex-A, and RISC-V processors.
We collect, label and build our own dataset of over 30M images and videos. Our data pipeline identifies failure cases to ensure that our models are highly reliable and accurate. We hire data actors that go out in the field to capture additional data. This approach guarantees the diversity and relevance of our training data, and ensures an ethical data collection practice.
We work closely with semiconductor companies that offer high performance and low power microcontrollers and SOCs. Using these chips, our software powers security cameras, consumer electronics, smart home cameras, and intelligent sensors.
Processing data on the device is inherently more reliable than a connection with the cloud. Intelligence shouldn’t have to depend on weak WiFi.
Sending data from the sensor to the cloud, processing the data, and sending it back again takes time. Sometimes whole seconds. This latency is problematic for products that need to respond to sensor input in real-time.
Running AI in the cloud comes with significant recurring compute costs. Executing the AI on the device instead saves several dollars per month in additional cloud costs.
Sending sensor data such as audio and video to the cloud increases privacy and security risks. To reduce abuse and give people confidence to let intelligent sensors into their lives, the data should not leave the device.
Ubiquitous connected sensors would overwhelm the network. Plumerai software only uses the network when it has something to report. This keeps bandwidth and mobile data costs low.
The farther we move data, the more energy we use. Sending data to the cloud uses a lot of energy. Processing data on-chip is more efficient by orders of magnitude. If a device needs a battery life of months or years, data needs to be processed locally.
Thank you for submitting the form. We’ll get in touch as soon as possible.
There was an error submitting the form. Please try again later.
We’re partnering with industry experts
