English Original:Adobe makes big bets on AI and the public cloud
AdobeEarlier this month, the annual MAX conference was held for Creative Cloud users. Adobe will be in the General Assembly released Photoshop and Premiere Pro applications such as new features. This year, however, Adobe also unveiled Sensei, a new platform for artificial intelligence and machine learning that combines the company's knowledge of photos, video, documentation and marketing data with a unified framework for artificial intelligence and machine learning.
Microsoft and Google are trying to inject "smart" elements into all products. Similarly, Adobe is also trying to make the product more intelligent, whether it is based on machine learning tools and features, or more intelligent traditional data analysis services. Sensei is Adobe's solution in this area. This framework will be applied to all Adobe products, not just Creative Cloud tools, will also include more important Marketing Cloud and Document Cloud services. In addition, Adobe plans to open these services to third-party software developers (ISVs) and independent developers through the application programming interface (API), allowing third parties to integrate the same functionality into their applications.
Currently, Sensei to Adobe Stock and Photoshop applications such as providing a "search by image" feature. (This is similar toGoogle PhotosIn the search function. ) Adobe wants the public to see Sensei as a "unified AI framework."
Adobe CEOAbbe & middot; Parnasis(Abhey Parasnis), said: "At the top, we have been concerned about how the technology and product investment, and ultimately optimize the digital experience, whether it is innovative side of the user experience, or service to marketers products. We are a technology product company, but this should be incorporated into three categories of products that drive innovation: Creative Cloud, Marketing Cloud, and Document Cloud. "Sensei is the most strategic project that Adobe has released so far this week, not only for the moment, but also for the next five to 10 years," he said. "We expect this to have a significant impact." This will help our products, but is more important to the user. & Rdquo;
Palais Nice said, Sensei is a "continuous learning system, from the user behavior to learn. Sensei can learn what users do in our suite of products. "He also stressed that Sensei will Adobe's two aspects, namely, creative tools and analysis tools, together.
Adobe Sensei Videos:Https://www.youtube.com/watch?v=t0QGAQwLHog&feature=youtu.be
From a technical perspective, Palas Nice pointed out that Adobe used a number of open source tools, such as Spark, Torch and TensorFlow, these tools into a unified framework, training its model. He believes that the next 10 years, the most successful intelligent system will have a very deep understanding of the field. For Adobe, this is creativity, document productivity, customer processes, and customer experience. "Unlike any other AI framework, our idea is to provide in-depth service and present the behavior that end-users can consume," Palas said. "He mentioned Photoshop's Face-Aware Liquify, and Document Cloud find similar documents and other functions.
Parasnis admits that companies like Google and Microsoft seem to be rivals for Adobe, but he believes that the company's platform is relatively broad, and Adobe plans to explore in depth. He even said that we will see Sensei and Microsoft infrastructure integration. This is not surprising considering that Adobe recently announced that it has selected Microsoft Azure as a priority public cloud service provider.
On cloud computing, Palas Nice believes that Adobe can use the development of public cloud to drive a lot of innovation. "In the long term, migration to public clouds is inevitable," he said. However, Adobe is still operating a small number of autonomous data centers, these data centers, at least in part based on the openOpenStack platform. The company does not plan to close these data centers, Palas Nis pointed out that many Adobe services, especially on data analysis and Marketing Cloud services, the need for very reliable delay performance, in less than 1 second time Experience. In his view, at least in the next few years, this will not change. "We currently have the most complex cloud computing infrastructure, including public and private clouds and our own data center." & Rdquo;