Menu

Google Cloud, machine learning injections

Google Cloud presents the new features of the machine learning platform to Google Cloud Next: Artificial Intelligence will make the difference.

Three steps forward, Artificial Intelligence direction: Google Cloud Next starts with a triptych of announcements related to new opportunities made available to those who want to develop for the corporate world, where the AI can really make the difference between the past and the future, between the fall and the success.

AI is no longer a technological niche: it is the element that makes the difference for companies in every sector.
Fei-Fei Li, Chief Scientist, Google AI

As already announced for the blockchain, Artificial Intelligence is also a kind of utility for Google, an enrichment of the offer that the Google Cloud platform intends to make available so that those who intend to develop will have at their disposal the best possible tools to reach the target. The group’s investment remains on the platform and on its business model, to be enriched through a galaxy of increasingly complete and updated services. “It is our commitment to provide the tools that will allow these companies to radically transform themselves,” explains Google, and then announces what the group intends to set in motion on machine learning:

  • CloudAtomML
    Google presents its solution as a necessary answer to the whole world that intends to have something more than simple pre-trained Cloud Vision API models, but at the same time they do not want to refine the TensorFlow and Cloud ML Engine levels: ” Cloud AutoML allows anyone to extend powerful machine learning models to meet the specific needs of their business, without requiring specialized knowledge of machine learning or coding “. Specifically, AutoML Vision extends, for example, Cloud Vision API “to recognize completely new categories of images. Refine the experience with our users in alpha, today we are happy to announce the public version in beta “; AutoML Natural Language helps to predict specific categories of text in the desired area; AutoML Translation allows you to train a specific translation model.
  • Improvements for Dialogflow Enterprise
    “Through intelligent tools that eliminate repetition and support human qualities such as creativity and problem solving, we are demonstrating that the greatest potential of AI is to complete and enrich human skills, not to replace them”: the expected improvements will make the conversational agents service that will eventually replace people within the contact centers. This is where the new Virtual Agents are born, first to respond to the user and then, if a satisfactory answer is not found, to forward the request to a person offering help in the customer service. “The result is a flexible solution”, Google explains, “which adapts to the specific needs of each call, provides a consistent experience between people and virtual agents, which take on a more or less relevant role depending on the situation, and respects our rules of data governance and privacy, all without an infrastructure, on a platform that can be scaled according to needs “.
  • AI Contact Center
    Google explains that it is working on intelligent Contact Center services able to improve the work, for example, of recruiters: less time lost and the ability to circumvent prejudices or other typically human elements that could lower the quality of the choices made. The group already has partners such as Appian, Chatbase, Cisco, Five9, Genesys, Mitel, Quantiphi, RingCentral, Twilio, UiPath, Upwire and Vonage and with each of them Google also works ethically. Google proves to have learned the lesson after the recent misstep with the presentation of Google Duplex to Google I / O: warn users that to answer the call is a bot – and not a person – allows you to get around the concerns that may arise when the AI begins to approach that threshold in which man is no longer able to understand independently if on the other side of the call there is a real voice or an intelligent system.

Leave a Reply

Your email address will not be published. Required fields are marked *