Relational inductive biases, deep learning, and graph networks

Relational inductive biases, deep learning, and graph networks Battaglia et al., arXiv’18

Earlier this week we saw the argument that causal reasoning (where most of the interesting questions lie!) requires more than just associational machine learning. Structural causal models have at their core a graph of entities and relationships between them. Today we’ll be looking at a position paper with a wide team of authors from DeepMind, Google Brain, MIT, and the University of Edinburgh, which also makes the case for graph networks as a foundational building block of the next generation of AI. In other words, bringing back and re-integrating some of the techniques from the AI toolbox that were prevalent when resources were more limited.

We argue that combinatorial generalization must be a top priority for AI to achieve human-like abilities, and that structured representation and computations are key to realizing this objective… We explore how using relational inductive biases within deep learning architectures can facilitate learning about entities, relations, and the rules for composing them.

Relational reasoning and structured approaches

Human’s represent complex systems as compositions of entities and their interactions. We use hierarchies to abstract away fine-grained differences, manage part-whole associations and other more Continue reading

VMware NSX-T Data Center in Evaluation for Common Criteria EAL4+ Certification

VMware NSX-T Data Center 2.x is now under evaluation for Common Criteria certification at Evaluation Assurance Level 4+ with BSI, Germany’s Federal Office for Information Security. Common Criteria is an internationally recognized standard (ISO-15408) that defines, validates, and assures security features and capabilities of IT security products. To see the evaluation status for VMware NSX-T 2.x, visit the German BSI certification website and reference certificate # BSI-DSZ-CC-1099.

VMware NSX-T was introduced to help organizations meet the stringent security demands of containerized workloads, multi-hypervisor, and multi-cloud. And this latest milestone for NSX-T 2.x reinforces VMware’s continuing commitment to deliver secure software to our customers. During the Common Criteria certification process, VMware NSX-T will undergo a thorough and rigorous evaluation methodology, with testing performed by a commercial Common Criteria Evaluation Facility under the oversight of the Certification Body. The Common Criteria certification acts as a seal of assurance for the federal government, its agencies, contractors and other organizations and assures that the product complies with strict security requirements specified within the designated level.

Within the VMware NSX portfolio, we have a long history of investing in certification efforts. For example, VMware NSX Data Center for vSphere 6.x also Continue reading

Kernel of Truth episode 7: data center networking in APAC and EMEA

Subscribe to Kernel of Truth on iTunes, Google Play, Spotify, Cast Box and Sticher!

Click here for our previous episode.

We wanted to give this podcast a bit of international flair, so we invited some overseas guests into the recording booth. I’m joined by Attilla de Groot (Sales Engineer for EMEA) and Sutharsan Sivapalan (Sales Engineer for APAC), who filled me in on the networking customers, trends and challenges that are cropping up in their respective regions. There are definitely differences between these two ends of the world, but you’d be surprised how much these regions have in common despite the distance.

Tweet any questions, feedback or topics you want us to discuss at @cumulusnetworks and use the hashtag #KernelOfTruth — let us know if you like what you’re hearing!

Guest bios

Sutharsan Sivapalan: CCIE #40322 (Data Center), is a Senior Systems Engineer covering the US West and Asia-Pacific regions for Cumulus Networks. Prior to joining Cumulus, Sutharsan spent 6 years at Cisco designing and troubleshooting some of the most complex networks in the world, as a member of their Technical Services organisation. In that role, he supported the entire Data Centre portfolio, including UCS, the Nexus Continue reading

Arm Upstart Ampere Starts Chipping Away At Intel Xeons

When this is all done, Intel might have wished it had kept Renee James as president and chief executive officer, because Ampere, an Arm server chip startup that James has been running since this spring, wants a big piece of the Xeon datacenter business and it has the financial backing to start a price war that others can win and only Intel can lose. …

Arm Upstart Ampere Starts Chipping Away At Intel Xeons was written by Timothy Prickett Morgan at .

AMD plots its move against Intel in the data center

Smelling blood in the water, a revitalized AMD is preparing for a big push against Intel in the data center, hoping to win back the market share it gained and lost a decade ago.AMD is promoting its Epyc processors, with 16 or 32 cores, as a lower TCO, higher performance option than Intel’s Xeon. It argues a single-socket 32-core server is cheaper up front and in the long run than a dual socket setup, which is Intel’s bread and butter.“We’re not saying single socket is for everyone, but at the heart of the market is where 50 percent to 80 percent are 32 cores per server and down, and our top single socket can do it more efficiently with lower costs and licensing. But in some cases some people will want to stay at two-socket,” said Glen Keels, director of product and segment marketing for data center products at AMD.To read this article in full, please click here

AMD plots its move against Intel in the data center

Smelling blood in the water, a revitalized AMD is preparing for a big push against Intel in the data center, hoping to win back the market share it gained and lost a decade ago.AMD is promoting its Epyc processors, with 16 or 32 cores, as a lower TCO, higher performance option than Intel’s Xeon. It argues a single-socket 32-core server is cheaper up front and in the long run than a dual socket setup, which is Intel’s bread and butter.“We’re not saying single socket is for everyone, but at the heart of the market is where 50 percent to 80 percent are 32 cores per server and down, and our top single socket can do it more efficiently with lower costs and licensing. But in some cases some people will want to stay at two-socket,” said Glen Keels, director of product and segment marketing for data center products at AMD.To read this article in full, please click here

IDG Contributor Network: Zero Trust Networking (ZTN): don’t trust anything

John Kindervag, a former analyst from Forrester Research, was the first to introduce the Zero-Trust model back in 2010. The focus then was more on the application layer. However, once I heard that Sorell Slaymaker from Techvision Research was pushing the topic at the network level, I couldn’t resist giving him a call to discuss the generals on Zero Trust Networking (ZTN). During the conversation, he shone a light on numerous known and unknown facts about Zero Trust Networking that could prove useful to anyone. The traditional world of networking started with static domains. The classical network model divided clients and users into two groups – trusted and untrusted. The trusted are those inside the internal network, the untrusted are external to the network, which could be either mobile users or partner networks. To recast the untrusted to become trusted, one would typically use a virtual private network (VPN) to access the internal network.To read this article in full, please click here

IDG Contributor Network: Zero Trust Networking (ZTN): don’t trust anything

John Kindervag, a former analyst from Forrester Research, was the first to introduce the Zero-Trust model back in 2010. The focus then was more on the application layer. However, once I heard that Sorell Slaymaker from Techvision Research was pushing the topic at the network level, I couldn’t resist giving him a call to discuss the generals on Zero Trust Networking (ZTN). During the conversation, he shone a light on numerous known and unknown facts about Zero Trust Networking that could prove useful to anyone. The traditional world of networking started with static domains. The classical network model divided clients and users into two groups – trusted and untrusted. The trusted are those inside the internal network, the untrusted are external to the network, which could be either mobile users or partner networks. To recast the untrusted to become trusted, one would typically use a virtual private network (VPN) to access the internal network.To read this article in full, please click here

Scale Computing, APC partner to offer micro data center in a box

Hyperconverged infrastructure (HCI) vendor Scale Computing and power management specialist APC (formerly American Power Conversion, now owned by Schneider Electric) have partnered to offer a range of turnkey micro data centers for the North American market.The platform combines Scale’s hyperconverged software, HC3 HyperCore, running on top of its own hardware and built on APC’s ready-to-deploy racks for a micro data center. Micro will sell the platform as a single SKU.The pre-packaged platform is entirely turnkey, with automated virtualization, power management resources, and built-in redundancy. This makes it well-suited for remote edge locations, such as cell phone towers, where staff is not immediately available to maintain the equipment.To read this article in full, please click here

1 … 1,481 1,482 1,483 1,484 1,485 … 3,860