On oktoober 24, Lenovo Group avalikustatud its most comprehensive artificial intelligence product technology to date, and its vision of "AI for All" at the Tech World. See aasta's sündmus, Lenovo's üheksas, fookused on kuidas Lenovo is unleashing the power of AI to drive intelligent transformatsioon in iga aspekt of people's lives and every industry - from the pocket to the cloud, providing dedicated AI devices, infrastructure, solutions and services to industries, enterprises and individuals around the world.
Yang Yuanqing, esimees ja tegevjuht of Lenovo Group, ütles et kunst intelligentsus tehnoloogia on muutunud nii reaalne eest esimene aeg see on tihedalt seotud iga iga inimene ja iga ettevõte. samas rõõmustamine tehnoloogiline progress, me on ka alustatud a sügavam uurimine.
Yang Yuanqing, CEO of Lenovo Group: For the first time, kunst intelligentsus tehnoloogia on muutunud nii reaalne
Mis on the "key but invisible" elements that make generative AI and large language models work? There is no doubt that data, as a product of the first phase of digital transformation, is essential. Algorithm/model-based training and reasoning are key, but at the same time, computing power is indispensable, not only on the cloud, but also with more powerful computing power on the edge side and the end side.
To create panoramic AI, "pocket to cloud" computing power, multiple forms of application are needed, and solutions from different industries are also critical.
Seal on a dilemma siin, kui kasutaja tahab the public big model to vasta sinu küsimused or talk to you in the right and appropriate content that you want, then you have to honestly tell it what you really think, the real records, the data, but then the user's personal data and even the privacy and the business secrets become part of the public information.
Inimesed tahavad the ability to have big models that can vastata questions, but they also want their data to live only on their own devices or within their own business. How can you do "both want - also want"? The answer is obviously not limited to public large models, we can achieve through individual large models and enterprise large models.
Lenovo's large model compression technology enables our own smart terminals and devices to run personal-scale large models. On these AI-enabled terminals and edge devices, local knowledge bases will be built to better understand users. The individual large scale model will use personal data stored on a device or home server for inference. Kui volitatud by the user, the user's personal data is never shared or sent to the public cloud, see enensure personal privacy and data security.
ettevõte suur mudel will kooseksisteerib koos avalik suur mudel ja avalik pilv to vorm a hübriid vorm ja hübriid juurutamine tehis intelligentsus. see hübriid trend on juba esineb in palju stsenaariumid, selline nagu hübriid ja digitaalne kontor Ruumid.