DOD AI ethics; GSMA/teleco edge platform; AI trains AI on Cortex-M; Tesla crash.
AI/Edge
The United States Department of Defense (DOD) has adopted ethical principles for using artificial intelligence in warfare that chiefly say the U.S. has to follow the laws, treaties, in use of AI in warfare. Any AI used by DOD has to be responsible, equitable, traceable, reliable and governable. “The Department will design and engineer AI capabilities to fulfill their intended functions while possessing the ability to detect and avoid unintended consequences, and the ability to disengage or deactivate deployed systems that demonstrate unintended behavior,” says a DOD press release. The Joint Artificial Intelligence Center, part of the United States’ Pentagon, is also planning on hiring Alka Patel in September to be its AI ethicist, to help implement those principles, according to NextGov.
The GSMA announced it is bringing together mobile telecos China Unicom, Deutsche Telekom, EE, KDDI, Orange, Singtel, SK Telecom, Telefonica and TIM in 2020 to build interoperable edge platforms called Teleco Edge Clouds. The first Teleco Edge Clouds will launch some European markets but eventually be global. According to a press release, “operators have agreed to work together to develop an Edge Compute architectural framework and reference platform.” The GSMA will support this initiative via the Operator Platform Project, an effort to develop a platform for one access point into all the operators’ networks for edge computing and other services, including storage. A GSMA white paper describes the idea. Edge computing company MobiledgeX is working with GSMA on the project, according to an RCR Wireless story.
Synopsys delivered DesignWare HBM2E IP in TSMC’s N7 process. Aimed at high-performance computing SoCs, the IP provides up to 409 GBps aggregate memory bandwidth with low-power consumption and latency PHY IP.
SI2 launched a special interest group (SIG) for AI and machine learning (ML) in electronic design automation. The group will focus on the growing needs and opportunities in artificial intelligence and machine learning for electronic design automation.
Startup company Cartesiam announced that its AI edge product for resource-constrained embedded systems — an AI development systems that automatically provides the user with a suitable AI algorithm — is now generally available. To keep it simple, Cartesiam focuses its NanoEdge AI Studio on Arm Cortex M microprocessors. Cartesiam says the learning and inferencing happens on the edge chip. The idea is to give embedded systems developers who know nothing about AI an easy way to add AI to edge devices without too much fuss.
MIT named tiny AI as one of its 10 breakthrough technologies. Tiny AI is generic term for any AI that is inferenced on the edge using lighter weight algorithms done on resource constrained devices, such as an IoT device. MIT also named quantum computing as becoming viable, thanks to Google’s quantum efforts.
Data center, cloud
Cadence announced verification IP for the PHY layer. Cadence says its VIP for PHY covers multiple protocols and will enable testing and optimizing of PHY designs, which should shave some development time off data center, artificial intelligence (AI), machine learning (ML) and mobile application designs.
RapterH, ANSYS’ new tool for SoCs and high-performance computing designers wrestling with electromagnetics, combines the company’s RapterX and HFSS tools into one analysis tool. The tool helps designers overcome unwanted interference in their designs by simulating electromagnetic phenomena on advanced nanometer silicon designs across multi-chip 3D-ICs, silicon interposers and advanced packaging, says the company in its press release.
Intel launched its second generation Xeon processor for data centers. The new Xeon Gold processors deliver an average of 1.36-times higher performance and 1.42-times better performance-per-dollar1 compared with the 1st Gen Intel Xeon Gold processors, says the company in a press release.
Internet of Things
Industry organization SEMI held the co-located Flexible Hybrid Electronics Conference and Exhibition (FLEX) and MEMS & Sensors Technical Congress (MSTC) this week in San Jose, California. The 2020 FLEXI Awards went to Epicore Biosystems for R&D Achievement in its research of a soft, skin-interfaced bioelectronics platform with high commercial potential; Product Innovation award to PragmatIC for product design and ingenuity in developing its ConnectIC family; and Technology Champion & Leadership in Education to NextFlex for FlexPro, a program that immerses working professionals into the development, manufacturability, and applications that can benefit from FHE technology.
Silicon Labs expanded its green power IoT Zigbee portfolio with Series 2 SoCs that the company says are optimized for ultra-Low-power applications.
Arm reports that its partners shipped record 6.4 billion Arm-based chips, including a record 4.2 billion Cortex-M processors in the fourth quarter of 2019. The Cortex-M processors are used for embedded and IoT applications.
Intel announced its 10nm 5G chip, Atom P5900, for wireless 5G basestations. Intel is supporting the infrastructure side of 5G.
Security
UltraSoC won a best in show security award for its Bus Sentinel hardware cybersecurity IP. The award was presented by Embedded Computing at embedded world 2020. UltraSoC also introduced CAN Sentinel, which is a layer of security for the CAN bus, targeted specifically at improving cybersecurity for the automotive industry.
Automotive
Tesla received nine recommendations from the United States NTSB (National Transportation Safety Board), which says that a fiery 2018 crash was caused by over-reliance on collision avoidance and Autopilot systems, mixed with driver’s distraction.
Globalfoundries says embedded magnetoresistive non-volatile memory (eMRAM) is now entering production on the company’s 22nm FD-SOI (22FDX) platform. The FDX eMRAM supports AEC-Q100 quality grade 2 designs, with development in process to support an AEC-Q100 quality grade 1 solution next year.
GM‘s Cruise now has a permit for its self-driving cars to take passengers on California roads as long as a safety driver is still present in the vehicle.
People
Philipp A. Hartmann, who served as chair of the Accellera‘s SystemC Language Working Group (LWG), will receive the 2020 Accellera Technical Excellence Award recognizing outstanding achievements of an individual among Accellera’s working group members.The ceremony is next week at DVCon in San Jose, California.
Meyya Meyyappan, chief scientist for exploration technology at the Center for Nanotechnology, NASA Ames Research Center, won a SEMI FLEXI award for his technical leadership in printed and flexible electronics, outreach activities and tireless volunteering to promote the field. FlexTech Chairman Michael McCreary won a FLEXI for his dedication and involvement in the commercialization of FHE, unwavering efforts to forge an industry vision, and contributions to technical developments.
“Brewer Science has successfully launched a zero-defect program to significantly improve the quality of materials, enabling the industry to advance next-generation technologies, including EUV,” said Brewer Science’s Srikanth Kommu, executive director of the Semiconductor Business Unit, in a press release.
Leave a Reply