"the world's fastest knowledge graph"
I note that nowhere is it mentioned anything about accuracy of results. No, the important thing is that it is the fastest because, of course, if you're not the fastest, you're lagging behind.
Samsung announced the acquisition of UK knowledge graph startup Oxford Semantic Technologies on Thursday, to boost its AI smarts and offer more personalized experiences and content on its devices. Oxford Semantic Technologies (OST) specializes in the knowledge graphs – a database tech that stores and organizes data as an …
Oh, goodie.
Another list of products I don't want.
Is my life really so much poorer without major corporations being aware of my 'proclivities'? I don't think so... I recently bought a paraglider. Guess how many adverts I saw for it? Correct: none.
And the Ad slingers wonder why we hate adverts?
It happened to me as well after buying an Anvil (from Amazon). For three years after that I was pestered with ads for Anvils. One lasts more than a lifetime but the ad slingers seem to think that I could make use of more than one or three or six...
"One lasts more than a lifetime but the ad slingers seem to think that I could make use of more than one or three or six..."
Admit it your real name is Mr. Wile E. Coyote, and you know you use anvils like I use sugar in Tea !!!
[s]ACME[/s]Amazon knows all !!!
:)
I feel I might perhaps like to keep my procivities to myself. ;) Certainly not have them shared with the malignant spawn of the Sirius Cybernetics Corporation.
Sounds like Oxford Semantic Technologies is using something along the lines of John Sowa's conceptual graphs/schema (Conceptual Structures was mid-80s so not particularly new tech ;)
Semantic graphs presumably annotate syntactic graphs with meaning (semantics). As far as I can see explicit "meaning" is notably missing from LLMs - seems to me they are mostly a syntactic representation.
Linguistics gets rather deep very quickly once you start thinking about the relationships between structure (syntax) and meaning (semantics.)
Even applying mathematical meaning to simple (procedural) programming language statements (denotational semantics) is pretty heavy going.
--
Living in a world in which every appliance and device panders to and anticipates the owner's every wish or desire cannot be healthy. Like letting a djinni out of its bottle and ultimately pure hell I should think.
A general problem with these adaptive systems is that we people are also adaptive systems. We're pretty good at adapting to something if it stays still for long enough. If you have an adaptive system (person) trying to adapt to an adaptive system (electronic device), and vice versa, literal chaos can ensue.
And knowing this, I wouldn't even start trusting my (hypothetical) smart fridge enough to leave all the ordering of milk to it... until it has a solid track record of ordering milk... which it's not going to get because I don't trust it to order milk.
"you said you liked <x> when you were 12 so here's <x> related items every christmas until you're 40"
Will AI realise that because i ate loads of apples last week maybe i just don't fancy them this week? or will it see an upward trend of apples and decide that's my defining feature and buy them forever?
I'd love my proclivities to be consistent and predictable but they're not, so the AI's going to be as confused as i am.
"What's your hobby this month then?"
"haven't decided yet let's try running"
"I'm a super computer and i would never have guessed that one in a million years"