Microsoft-backed robovans to deliver grub in London

British startup Wayve gets supercomputing leg up


Microsoft is pumping supercomputing oomph as well as funds into a British-born autonomous vehicle startup.

On Wednesday Wayve, the upstart in question, confirmed it has struck a deal with Microsoft – not surprising since Redmond has already sunk a chunk of change into the business – to use Azure to train next-gen self-driving machines from data collected from human drivers out on the road. Richard Branson, Meta AI Chief Yann LeCun, and other heavyweights are also early investors alongside the Windows giant.

"Joining forces with Microsoft to design the supercomputing infrastructure needed to accelerate deep learning for autonomous mobility is an opportunity that we are honored to lead," said Alex Kendall, CEO of Wayve.

"Deep learning systems thrive on data, and we've put an immense amount of effort into understanding what it takes to get these systems on the road. We are excited by the opportunities that this collaboration will create as we push deep learning to new levels of scale."

Wayve has also partnered with UK retailers Asda and Ocado Group, and shipping company DPD, to trial its self-driving system in delivery vans in London. The upstart reckons data generated during the trial will provide Wayve with fleet-scale information to further improve its autonomous vehicle technology for last-mile delivery.

autonomous car

British motorists will be allowed to watch TV in self-driving vehicles

READ MORE

Wavye seems to rely on human drivers in camera-laden cars demonstrating how to follow the laws of the road; data from these journeys are used to teach autonomous vehicle control systems how to drive in a general sense, it is claimed. The biz argued this allows cars powered by its kit to get around cities without the need for highly mapped streets, as the computer system has learned to obey traffic laws and navigate through neighborhoods like a person.

Wayve began working with Microsoft in 2020 when its computing needs had grown beyond the 12-HDD RAID array the company had originally built, and the SSD-based cache server that replaced it. "With the scale of our ambition to build our Fleet Learning platform to train on diverse and enormous driving data, it became immediately clear we need to move to the cloud," the biz wrote in a blog post that year. 

The amount of data produced by such tests and the computing resources necessary to process it, however, are quickly hitting those aforementioned "upper bounds of what's possible with commercially available cloud and edge computing services," Wayve said. 

Wayve trains its AV models using a constant loop of processing and parsing data, training models, re-training to eliminate failures, deploying models, and updating them based on fresh data. Wayve's Kendall told Reuters that the company's cars generate a terabyte of data every minute. 

The startup said it wants infrastructure able to scale up to "training models with trillions of parameters and exabyte-scale image data from real-world driving and simulation." ®


Other stories you might like

  • Supply chain blamed amid claims of Azure capacity issues
    Microsoft says it'll move to 'restrict trials and internal workloads to prioritize growth of existing customers'

    Microsoft's Azure cloud is having difficulty providing enough capacity to meet demand, according to some customers, with certain regions said to refusing new subscriptions for services.

    Azure comprises over 200 datacenters globally spread across 60 regions, but reports suggest that over two dozen of these are operating with limited capacity, and that the cloud and IT giant is being forced to prioritize resources in order to serve existing customers.

    According to technology news site The Information, capacity issues are affecting Azure datacenters in Washington State in the US as well as across Europe and Asia, and it claims that server capacity is expected to remain limited until early next year, citing a Microsoft insider.

    Continue reading
  • FabricScape: Microsoft warns of vuln in Service Fabric
    Not trying to spin this as a Linux security hole, surely?

    Microsoft is flagging up a security hole in its Service Fabric technology when using containerized Linux workloads, and urged customers to upgrade their clusters to the most recent release.

    The flaw is tracked as CVE-2022-30137, an elevation-of-privilege vulnerability in Microsoft's Service Fabric. An attacker would need read/write access to the cluster as well as the ability to execute code within a Linux container granted access to the Service Fabric runtime in order to wreak havoc.

    Through a compromised container, for instance, a miscreant could gain control of the resource's host Service Fabric node and potentially the entire cluster.

    Continue reading
  • Cerebras sets record for 'largest AI model' on a single chip
    Plus: Yandex releases 100-billion-parameter language model for free, and more

    In brief US hardware startup Cerebras claims to have trained the largest AI model on a single device powered by the world's largest Wafer Scale Engine 2 chip the size of a plate.

    "Using the Cerebras Software Platform (CSoft), our customers can easily train state-of-the-art GPT language models (such as GPT-3 and GPT-J) with up to 20 billion parameters on a single CS-2 system," the company claimed this week. "Running on a single CS-2, these models take minutes to set up and users can quickly move between models with just a few keystrokes."

    The CS-2 packs a whopping 850,000 cores, and has 40GB of on-chip memory capable of reaching 20 PB/sec memory bandwidth. The specs on other types of AI accelerators and GPUs pale in comparison, meaning machine learning engineers have to train huge AI models with billions of parameters across more servers.

    Continue reading
  • Meta: We need 5x more GPUs to combat TikTok, stat
    And 30% fewer new engineers this year

    Comment Facebook parent Meta has reportedly said it needs to increase its fleet of datacenter GPUs fivefold to help it compete against short-form video app and perennial security concern TikTok.

    The oft-controversial tech giant needs these hardware accelerators in its servers by the end of the year to power its so-called discovery engine that will become the center of future social media efforts, according to an internal memo seen by Reuters that was written by Meta Chief Product Officer Chris Cox.

    Separately, CEO Mark Zuckerberg told Meta staff on Thursday in a weekly Q&A the biz had planned to hire 10,000 engineers this year, and this has now been cut to between 6,000 and 7,000 in the shadow of an economic downturn. He also said some open positions would be removed, and pressure will be placed on the performance of those staying at the corporation.

    Continue reading
  • Is computer vision the cure for school shootings? Likely not
    Gun-detecting AI outfits want to help while root causes need tackling

    Comment More than 250 mass shootings have occurred in the US so far this year, and AI advocates think they have the solution. Not gun control, but better tech, unsurprisingly.

    Machine-learning biz Kogniz announced on Tuesday it was adding a ready-to-deploy gun detection model to its computer-vision platform. The system, we're told, can detect guns seen by security cameras and send notifications to those at risk, notifying police, locking down buildings, and performing other security tasks. 

    In addition to spotting firearms, Kogniz uses its other computer-vision modules to notice unusual behavior, such as children sprinting down hallways or someone climbing in through a window, which could indicate an active shooter.

    Continue reading
  • Amazon can't channel the dead, but its deepfake voices take a close second
    Megacorp shows Alexa speaking like kid's deceased grandma

    In the latest episode of Black Mirror, a vast megacorp sells AI software that learns to mimic the voice of a deceased woman whose husband sits weeping over a smart speaker, listening to her dulcet tones.

    Only joking – it's Amazon, and this is real life. The experimental feature of the company's virtual assistant, Alexa, was announced at an Amazon conference in Las Vegas on Wednesday.

    Rohit Prasad, head scientist for Alexa AI, described the tech as a means to build trust between human and machine, enabling Alexa to "make the memories last" when "so many of us have lost someone we love" during the pandemic.

    Continue reading

Biting the hand that feeds IT © 1998–2022