AI & Robotics

Despite the proliferation of accelerator chips like Google’s tensor processing unit (TPU) and Intel’s forthcoming Nervana NNP-T, most machine learning practitioners are limited by budget or design to commodity processors. Unfortunately, these processors tend to train sophisticated AI models rather slowly, exacerbating one of the many challenges involved in AI R&D. Hence, Neural Magic. MIT…

The US Postal Service (USPS) delivers an estimated 146 billion pieces of mail annually, including 6 billion packages. In an attempt to process package data more efficiently, USPS is experimenting with AI. Today, NVIDIA announced that it will provide USPS with its AI tech. NVIDIA claims its system will process package data 10-times faster and with higher accuracy. For months, engineering teams…

San Francisco-based HomeLight, which develops AI-driven software and services for home buyers, sellers, and real estate agents, today announced that it has raised $109 million in funding from a raft of investors, bringing its total raised to over $160 million. Zeev Ventures led the round, with participation from Group 11, Menlo Ventures, Crosslink Capital, Stereo Capital, and others. CEO Drew Uher…

Smart home thermostat company Ecobee today rolled out Eco+, a free suite of machine learning features, to all its thermostats with voice support. Eco+ uses machine learning to tailor temperature settings and cut costs. It also helps you further reduce energy consumption and move to clean energy. In July 2018, Ecobee kicked off a pilot called Peak Relief that aims to save energy and money by…