The age of personalised AI is right here. Samsung Electronics provides secure, tailor-made AI experiences backed by person insights throughout its merchandise — from cellular gadgets and TVs to residence home equipment — by means of steady innovation in software program and platform know-how.
Just as efficiency metrics and bodily measurements are important to designing personalized exercise routines for skilled athletes, information performs an important position in creating personalised AI. As such, Samsung is prioritizing information intelligence analysis to rework huge quantities of knowledge into significant and intuitive person experiences.
Samsung Newsroom takes a better take a look at how Samsung’s superior information analysis and modern connectivity options are pioneering the period of personalised AI.
The Power of Data Intelligence in Personalized AI
A very personalised AI expertise begins with understanding the person. Data intelligence identifies a person’s wants at particular moments based mostly not solely on historic gadget utilization information but additionally on the person’s current context. This permits acceptable companies to be delivered in a well timed method. However, there are challenges — for instance, customers don’t hand in refined experiences of their wants and needs; the matching of anticipated wants and potential companies have to be carried out in actual time; and most of all, private data should always be stored secure.
As Samsung reaches customers throughout practically all every day touchpoints — together with work, leisure, family chores, well being and safety — the corporate has the distinctive alternative to realize deep insights into person behaviors and preferences.
In order to harness this potential, Samsung has strengthened its information know-how capabilities by researching person information collected yearly from a whole lot of tens of millions of gadgets and companies — all in strict compliance with inside and exterior laws. As AI fashions develop a greater understanding of customers by means of information assortment, processing and evaluation, they will leverage retrieval-augmented era (RAG) to ship extra personalised companies for various contexts. By utilizing applied sciences to attach dependable, logic-based information bases to massive language fashions (LLMs), these AI fashions entry the newest, most dependable data and discover essentially the most appropriate solutions.
As Samsung focuses on applied sciences that join information to refine personalised AI experiences, a key query is how can such huge and complicated information be managed extra successfully and tailor-made exactly to particular person wants? The reply, partially, lies in information graphs.
▲ Yunsu Lee, VP, Head of Data Intelligence Team, Samsung Research explains information intelligence.
Enhancing AI Experiences With Knowledge Graphs
A information graph is a information illustration know-how that organizes associated data in an interconnected graph format, making it simply interpretable for each people and computer systems. By defining particular person entities throughout the information and linking them, their relationship could be represented as a “triple” within the type of “subject — predicate — object.” For instance, “the capital of South Korea is Seoul” could be represented within the following graph, “South Korea — capital — Seoul.” Various varieties of information comparable to individuals, objects, occasions and locations could be structured in a extremely refined method by connecting quite a few entities on this triple format.
Unlike conventional relational databases that fill tables with data, information graphs successfully retailer information and shortly retrieve data or new inferences as wanted.
▲ (Right) Knowledge graphs can retailer information extra effectively than relational databases (left).
Samsung is actively advancing analysis in information graphs. While earlier types of information utilization have been restricted to particular person apps, this know-how permits at no cost information search and utilization throughout app or…
