Qeexo’s Director of Engineer, Dr. Rajen Bhatt, gave an excellent presentation of Qeexo AutoML at this year’s ARM AIOT Conference: https://youtu.be/1pI9dpgB0W4

Using Qeexo AutoML, companies can launch machine learning on embedded devices and begin analyzing data in a few hours

Oct 7, 2019; Mountain View–Qeexo today announces the launch of its AutoML product, a one-click, fully automated platform that allows customers to rapidly build machine learning solutions for Edge devices using sensor data. Qeexo has selected the Arm® Cortex™-M0-M4 class MCUs as the first hardware targets to be supported by Qeexo AutoML. At launch, Qeexo AutoML will support STMicroelectronics’s SensorTile.box, a compact multi-sensor module which includes the Cortex-M4 MCU and will continue to augment support for other hardware platforms.

Machine learning is moving to embedded processors on edge devices, improving privacy, latency, and availability. However, given limited computation power, memory size, and battery life, building machine learning solutions for edge devices is challenging. Achieving commercial-grade performance requires a team of difficult-to-hire machine learning engineers who devote their time to: preprocess data, extract features, select models, optimize hyperparameters, validate results, and deploy models to target. Even for experts, this is a lengthy, error prone, and repetitive process.

With its one-click, fully automated workflow, Qeexo AutoML greatly simplifies the machine-learning-solution development process and eliminates room for errors. All the complicated machine learning tasks are automated by Qeexo AutoML. Machine learning engineers can now focus their time on mission-critical R&D instead of performing tedious, repetitive steps. In addition, Qeexo AutoML eliminates the need for companies to invest in expensive, in-house machine learning teams, resulting in huge time and cost savings.

“Thousands of companies are collecting vast amounts of data at the edge. These companies want to leverage machine learning but don’t have the necessary tools or the technical staff,” said Sang Won Lee, CEO of Qeexo. “With Qeexo AutoML, companies can iterate through prototypes and projects to produce production-ready models with a fraction of the time and resources previously required. We chose to provide support first to Arm-based MCUs due to Arm’s dedication to building a world-class ecosystem and its global leadership in the edge markets.”

“Machine learning is solving complex problems that have often required significant performance,” said Dennis Laudick, vice president of Marketing, Machine Learning Group, Arm. “Qeexo’s optimizations will bring new machine learning capabilities to an even broader range of devices, and targeting Arm-based MCUs means their technology will benefit a rich ecosystem serving nearly all industries.”
“An automated machine learning tool like Qeexo AutoML extends the reach of our products while providing tremendous value to our joint customers,” said Miguel Castro, Head of Marketing for the STMicroelectronics AI Solutions Group. “Qeexo’s choice of the ST SensorTile.box evaluation kit, which embeds our advanced STM32 microcontroller, STNRG Bluetooth, and 8 sensors to be the first hardware target to support on Qeexo AutoML highlights the importance and usefulness of the module.”

Qeexo AutoML is based on the same machine learning platform that Qeexo developed as the basis for its FingerSense, EarSense, and TouchTools products, which are commercialized on over 210 million consumer devices worldwide.

About Qeexo
Qeexo is the first company to automate end-to-end machine learning for embedded Edge devices (Cortex M0-M4 class). Our one-click, fully-automated, Qeexo AutoML platform allows customers to leverage sensor data to rapidly build machine learning solutions for highly constrained environments with applications in mobile, IoT, wearables, automotive, and more.

Delivering high performance, solutions built with Qeexo AutoML are optimized to have ultra-low latency, ultra-low power consumption, and an incredibly small memory footprint. As billions of sensors collect data on every device imaginable, Qeexo can equip them with machine learning to discover knowledge, make predictions, and generate actionable insights.

Spun out of Carnegie Mellon University, Qeexo is venture-backed and headquartered in Mountain View, CA, with offices in Pittsburgh, Shanghai, and Beijing. To learn more, visit https://automl.qeexo.com.

December 06, 2018; Mountain View, CA – Qeexo, developer of machine learning and AI solutions for sensor data, today announced the launch of Qeexo Machine Learning Platform for embedded products and applications. The company also announced that its EarSense product has been named a CES® 2019 Innovation Awards Honoree.

Qeexo Embedded Machine Learning is a lightweight, general-purpose platform that can perform inferencing locally, on an embedded edge device, in real-time and without relying on the cloud. The solution is built upon Qeexo’s proprietary Machine Learning Platform, which already powers over 170 million smartphones and tablets worldwide with Qeexo’s FingerSense and EarSense products.

“Qeexo Embedded Machine Learning can help any company make sense of the constant streams of data their products and equipment are already gathering or could be gathering,” said Sang Won Lee, CEO of Qeexo. “As silicon chips continue to become more powerful and less expensive, we believe that the trend is for machine learning to move towards the edge.”

Qeexo Embedded Machine Learning can add intelligence to products and processes in any industry. For example, in an industrial setting, Qeexo-powered sensors can be set up in factories to monitor and analyze processes, equipment, and products of interest, allowing machinery to function longer and more optimally. In automotive, sensors equipped with Qeexo Embedded Machine Learning can relay current road and automobile conditions for real-time response or predictive maintenance of the car itself. In smart home and IoT, edge devices can be augmented with more useful and convenient functions at a low added cost.

Features of Qeexo’s Machine Learning Platform include:

• Millisecond-Latency: Since Qeexo’s millisecond-latency is faster than human perception, actions triggered by Qeexo’s machine learning feel instantaneous. Traditionally, machine learning could not be used for time-sensitive applications such as touchscreens, since calculations take too long and users would be confused when a touch surface does not respond immediately. Qeexo’s machine learning, as demonstrated in the FingerSense and EarSense products, can immediately differentiate between different types of touches and respond to user inputs without missing a beat.
• Ultra-Low Power Consumption, Memory, and Processing Requirement: Embedded and mobile applications are heavily constrained by processing power, memory, and power consumption. Qeexo’s Embedded Machine Learning is highly optimized, allowing for inferencing right at the edge, which results in a much wider range of possible applications.
• Sensor Data: Qeexo’s Machine Learning Platform uniquely works with data from all types of sensors. According to industry experts, over 1 trillion sensors could be deployed by 2020. Qeexo Embedded Machine Learning can leverage the vast amount data collected by those sensors, to make every device smarter and more convenient to use.

Demos

• Predictive Maintenance via Anomaly Detection: Detect current anomalies to predict likelihood of future system failures
• Environment-Aware Countertop: Know what’s happening on a customer’s counter to customize the kitchen experience
• Intelligent Shipping: Track impacts and shipping conditions
• Interactive Wall: Knock and wipe anywhere on the wall to control lighting

“One common complaint that we hear from companies is that they invest in collecting and storing data but they don’t know how best to utilize this data,” continued Lee. “Often, very primitive analysis is being done, if anything. With advanced machine learning algorithms, Qeexo Embedded Machine Learning can help companies realize the value that they are missing from sensor data.”

As another testament to the strength of Qeexo’s machine learning technology, Qeexo’s EarSense product won the CES 2019® Innovation Awards as an Honoree in the Software and Mobile Apps category. EarSense, which launched on the OPPO Find X in July, uses state-of-the-art AI to replace hardware proximity sensors on smartphones to turn the screen off during phone calls. By eliminating the need for a physical proximity sensor, EarSense allows manufacturers to achieve true full-screen designs.

EarSense will be on display at Qeexo’s booth during CES 2019, from January 8th to 11th, alongside other products powered by Qeexo’s Machine Learning Platform.

About Qeexo
Qeexo develops machine learning solutions that generate actionable insights from sensor data. The company works with leading device OEMs and component manufacturers to power new and exciting products and user experiences on over 170 million devices worldwide. In industries such as mobile, IoT and automotive, there are billions of devices where computation and memory are highly constrained. Qeexo’s proprietary, low-latency, low-power models are engineered to have an incredibly small footprint – ideal for making high-accuracy predictions in these environments.

Spun out of Carnegie Mellon University, Qeexo is venture-backed and headquartered in Mountain View, CA, with offices in Pittsburgh, Shanghai, Shenzhen and Beijing. To learn more, visit https://www.qeexo.com.

Qeexo EarSense uses AI to bring true bezel-less designs to OPPO smartphones

July 12, 2018; Mountain View, CA – Qeexo, developer of lightweight machine learning and AI solutions for sensor data, announces a partnership with smartphone and consumer electronics manufacturer, OPPO, to bring Qeexo’s EarSense technology to the OPPO Find X smartphone. EarSense is Qeexo’s state-of-the-art AI solution that allows for true bezel-less design. It replaces the traditional hardware proximity sensor on smartphones, allowing devices to turn off the screen during phone calls.

Traditional hardware proximity sensors, which allow devices to turn off the screen during phone calls, must be placed at the top of the smartphone, requiring devices to be made with bezels or notches. EarSense is the first software-only AI solution to replace the traditional hardware proximity sensor and was built upon the same AI platform as Qeexo’s FingerSense, which has already been deployed on over 120 million devices. EarSense’s proprietary AI algorithms identify a person’s face and ear when they are near or touching the screen, letting the device turn off the screen during a call – just like a hardware proximity sensor. By eliminating the need for a physical proximity sensor, EarSense finally gives manufacturers the freedom to remove top bezels and notches from smartphones and create more beautiful and imaginative designs.

“The footprint of a traditional proximity sensor makes it very difficult for manufacturers to design around, resulting in the unpleasant notches you see on many smartphones today,” said Sang Won Lee, CEO of Qeexo. “In addition, other existing proximity sensor replacement solutions are incredibly costly and perform poorly. EarSense overcomes these limitations, allowing for bezel-less and notch-less designs with high-performance ear and face detection. We’re thrilled to work with OPPO, a leader in the smartphone industry, to help them realize their design ambitions.”

To view a demo of EarSense, go to: https://www.youtube.com/watch?v=VgbyW_5gdvo.

About Qeexo

Qeexo develops machine learning and AI solutions that generate actionable insights from sensor data. The company works with leading device OEMs and component manufacturers to power beautiful designs and novel user experiences on over 100 million devices worldwide. In industries such as mobile, IoT and automotive, there are billions of devices where computation and memory are highly constrained. Qeexo’s proprietary, low-latency, low-power models are engineered to have an incredibly small footprint – ideal for making high-accuracy predictions in these environments.

Spun out of Carnegie Mellon University, Qeexo is venture-backed and headquartered in Mountain View, CA, with offices in Pittsburgh, Shanghai, Shenzhen and Beijing. To learn more, visit www.qeexo.com.

For further information, please contact:

Qeexo
Lisa Langsdorf
+1 347-645-0484
lisalangsdorf@gmail.com