Looks over priced, should have picked up one of the new startups that have been taking Tableau market share.
Salesforce is buying data visualisation specialist Tableau for $15.7bn as it looks to up its analytics game. As a CRM platform heavyweight, Salesforce collects caboodles of data – and the deal could make information presented by products like Customer 360 prettier to look at and maybe, just maybe, easier to understand. This …
Well? Let's go crazy?
So n-gram parsing vs. AI-parsing.
In the fields of computational linguistics and probability, an n-gram is a contiguous sequence of n items from a given sample of text or speech.
AI-parsing distills both continuous and discontinuous sequences of several words belonging to several parts of speech, a combination of a noun with other parts of speech at least one of which is a verb or an adjective.
For example there is a sentence
alice and bob exercise merrily, she trains a lot.
n-gram parsing will probably not be able to recognize that "alice and bob" are names, because there are no capital letters present. Next, n-gram will get "a contiguous sequence" for the first clause (if it sees the coma) - "alice and bob exercise merrily." For the second clause - "she trains a lot". And no weights! No statistics! (Ask Larry E if you don't trust me, will you?)
AI-parsing, using dictionary definition, detects that "alice" and "bob" are actually the names "Alice" and "Bob." After that, AI-parsing constructs two patterns for the first clause
- Alice exercise merrily
- Bob exercise merrily
For the second clause AI-parsing gets also two patterns:
- she trains a lot.
- Alice trains a lot.
Then AI-parsing counts weight for all patterns:
- Alice exercise merrily - 0.25
- Bob exercise merrily - 0.25
- she trains a lot - 0.5
- Alice trains a lot - 0.5
And finally comes the index by dictionary definitions and synonyms clusters. There are three here.
Do you see the difference? AI-parsing got four indexed by dictionary, understood by computer and weighted patterns, n-gram only two and no statistics.
How do causal relationships exist in AI database? For example, a relationship between this sentence
- alice and bob exercise merrily, she trains a lot.
and this paragraph:
- Alice is tired. She ran a lot, swam and trained merrily.
In both the sentence and the paragraph there is the pattern "Alice trains(ed) a lot", which establishes a relationship.
But the timestamp of the sentence is older than for the paragraph, which allows to establish the cause-and-effect: the sentence is a parent for the paragraph. So searching "Does Alice exercise merrily?" you may get the answer "Alice is tired".
United States Patent 8,516,013
14. The computer system of claim 9 in which said facility configured to extract predicative phrases is further configured to assign to the subtexts information regarding the date of their origin.
AI database is a blockchain technology.
Microsoft Corp. today debuted a new version of its Power BI, which competes in the same category as Tableau Software Inc. Analysts can now extend the tool’s default vocabulary with technical terms unique to their industry, company-specific terms and synonyms to help it answer queries more accurately.
I told you a few hours ago, didn't I?
If the computer does not see individual words, their parts of speech and dictionary definitions - it can not construct the pattern "Alice trains a lot - 0.5" and the information is lost. Do you want to search your database knowing that you are destined to lose? Because of that Microsoft allowed to add technical terms and synonyms.
A dictionary is a universal index, that is the same definition can be used at any time and in any place. Therefore, in AI database you can instantly find everything you want, nothing is lost and Microsoft drifts toward this database.
What companies like OpenAI offer? They create their own dictionaries, which contain definitions only on phrases (patterns), not on singular words. Also, OpenAI uses not reduced to a single system and form, the same definitions but different texts. I guess Microsoft doesn't want OpenAI idea and decided to use words.
Biting the hand that feeds IT © 1998–2022