In this demo, learn how you can explode a Bill of Materials using Graph Shortest Path function, introduced with SQL Server 2019 CTP3.1, to find out which BOMs/assemblies a given product/part belongs to. This information can be useful for reporting or product recall scenarios.
Michael Hansen joins Scott Hanselman to explain what FHIR is and how to get started with FHIR on Azure. Fast Healthcare Interoperability Resources (or, FHIR) is a new standard for representing and exchanging healthcare data. Developed by the HL7 community to address problems with interoperability, pieces of healthcare data are represented as resources in FHIR (i.e, a patient is a resource, observation is a resource, etc.). Resources are healthcare data objects with properties (e.g., a patient has a name) and relationships. The FHIR specification also describes how to exchange these objects using a REST API. A FHIR server is a REST API that enables you to search, retrieve, modify, and delete healthcare data objects. Microsoft developed a first-party FHIR server, which is available as an open source project on GitHub and as a managed service, Azure API for FHIR.
George Hotz explores the current wave of self-driving cars, both the real and the hype, as well as those profiting off the technology in a less-than-ethical way.
Deep learning has had enormous success on perceptual tasks but still struggles in providing a model for inference. Here’s an interesting talk about making neural networks that can reason.
To address this gap, we have been developing networks that support memory, attention, composition, and reasoning. Our MACnet and NSM designs provide a strong prior for explicitly iterative reasoning, enabling them to learn explainable, structured reasoning, as well as achieve good generalization from a modest amount of data. The Neural State Machine (NSM) design also emphasizes the use of a more symbolic form of internal computation, represented as attention over symbols, which have distributed representations. Such designs impose structural priors on the operation of networks and encourage certain kinds of modularity and generalization. We demonstrate the models’ strength, robustness, and data efficiency on the CLEVR dataset for visual reasoning (Johnson et al. 2016), VQA-CP, which emphasizes disentanglement (Agrawal et al. 2018), and our own GQA (Hudson and Manning 2019). Joint work with Drew Hudson.
Lex Fridman interviews Gustav Soderstrom, the Chief Research & Development Officer at Spotify.He leads Product, Design, Data, Technology & Engineering teams. This interview is part of the ongoing Artificial Intelligence podcast.
Two Minute Papers explores the “Learning the Depths of Moving People by Watching Frozen People” paper in this video.