Lifeboat Foundation: Safeguarding Humanity
Toggle light / dark theme

Blog

Category: robotics/AI – Page 1,622

Open source platform enables research on privacy-preserving machine learning

The biggest benchmarking data set to date for a machine learning technique designed with data privacy in mind has been released open source by researchers at the University of Michigan.

Called federated learning, the approach trains learning models on end-user devices, like smartphones and laptops, rather than requiring the transfer of private data to central servers.

“By training in-situ on data where it is generated, we can train on larger real-world data,” explained Fan Lai, U-M doctoral student in computer science and engineering, who presents the FedScale training environment at the International Conference on Machine Learning this week.

Energy Harvester Produces Power from Local Environment, Eliminating Batteries in Wireless Sensors

Advances in low power technology are making it easier to create wireless sensor networks in a wide range of applications, from remote sensing to HVAC monitoring, asset tracking and industrial automation. The problem is that even wireless sensors require batteries that must be regularly replaced—a costly and cumbersome maintenance project. A better wireless power solution would be to harvest ambient mechanical, thermal or electromagnetic energy from the sensor’s local environment.

Typically, harvestable ambient power is on the order of tens of microwatts, so energy harvesting requires careful power management in order to successfully capture microwatts of ambient power and store it in a useable energy reservoir. One common form of ambient energy is mechanical vibration energy, which can be caused by motors running in a factory, airflow across a fan blade or even by a moving vehicle. A piezoelectric transducer can be used to convert these forms of vibration energy into electrical energy, which in turn can be used to power circuitry.

Else Labs Announces Pro Kitchen Focused Oliver Fleet As It Pauses Rollout of Home Cooking Robot

Else Labs, the company behind the countertop home cooking robot called Oliver, announced today the launch of Oliver Fleet, a commercial kitchen reimagining of its original core product.

The new Fleet solution is a respin of its original standalone Oliver home cooking robot into a solution that allows multiple units to be used and managed simultaneously in professional kitchen environments to automate cooking tasks. According to company CEO Khalid Aboujassoum, while the Oliver Fleet units look the same from the outside as the original consumer unit, they’ve been built to withstand the more rugged requirements of the professional kitchen.

“It might look like the household unit from the outside, but the guts of the Oliver Fleet are different,” Aboujassoum said. “The Fleet units are designed for back-to-back cooking, for that harsh environment in the commercial kitchen compared to the household.”

Google fires engineer who claimed that company’s AI has become sentient

Last month, there were a lot of waves in the AI community when a senior Google engineer, Blake Lemoine, alleged that the company’s AI has become sentient. The claim was made about Google’s Language Model for Dialogue Applications (LaMDA) that can engage in a free-flowing way about a seemingly endless number of topics, an ability that unlocks more natural ways of interacting with technology.

Initially, Lemoine was put on paid administrative leave, but it appears that Google has now fired him.

Harmonizing human-robot interactions for a ‘new and weird’ world of work

Robots have always found it a challenge to work with people and vice versa. Two people on the cutting edge of improving that relationship joined us for TC Sessions: Robotics to talk about the present and future of human-robot interaction: Veo Robotics co-founder Clara Vu and Robust.ai founder Rod Brooks (formerly of iRobot and Rethink Robotics).

Part of the HRI challenge is that although we already have robotic systems that are highly capable, the worlds they operate in are still very narrowly defined. Clara said that as we move from “automation to autonomy” (a phrase she stressed she didn’t invent) we’re adding both capabilities and new levels of complexity.

“We’re moving … from robotic systems that do exactly what they were told to do or can perceive a very specific very low-level thing, to systems that have a little bit more autonomy and understanding,” she said. “The system that my company builds would not have been possible five years ago, because the sensors that we’re using and the processors that we’re using to crunch that data just didn’t exist. So as we do have better sensors and more processing capabilities, we’re able to, as you said, understand a little bit more about the world that we’re in and sort of move the level of robotic performance up a notch.”

/* */