One month ago I decided to create a new word. The word was ‘AntiGooglelianos’. The word itself was the first step for a new search experiment that we now call The AntiGooglelianos Experiment. As you can see above this is the very first time this result have been crawled, indexed, and stored by a search engine. It is the very first time in the internet world. As we see, Google has a big-time problem. Google performance is not human. It is ready for you, enjoy.
The word itself has had no meaning until today 19/03/2007.
Usually Antigooglelianos might mean in Spanish something like “those against the Google system”. But today, we define Antigooglelianos as “the hermeneutic and logic experiment that in the near future will save Google from itself”. 🙂 That means that an important search philosophy question should be thought about:
I have created the word… so can I assign to Antigooglelianos whatever meaning I wish?
The answer is no, not at all.
Meaning is not a private thing. It is shared. I welcome you to use and understand my word. Once you do, it will be alive.
Once The Machine links to it, antigooglelianos will have a “google meaning” pretty much different from the human meaning, not created and understood yet.
This is basic:
The Machine Link,
the Human Being Think.
Search Philosophy lesson 1
The important thing is that before being a result, AntiGooglelianos document has been previously crawled, indexed, and stored in a database accessible by search engine. And before everything, it was in my mind.
The first Search philosophy question is the following:
Does this mean that the meaning has been scored by a search engine?
What comes to my mind is another question:
How can a search engine score a 1-1 result that came from my Mind?
May be our initial question is wrong and it is not scored?
To be scored it must have a logic order and also a meta-order.
We should accept then that a 1-1 result should be taken as a 1-1 Model and accepting the main logic its antecedents are my mind, with independence of the anchor and the query, the result itself is nonsense…
If we consider the Term Count Model Theory and assuming that the query Antigooglelianos is a result is a 1-1 result, we might SEO expect that the Machine (the database vs search engine) is going to also find, from today onwards, terms that could be used to expand or reformulate the query.
From a human point of view, this is nonsense because AntiGooglelianos does not exist and has no meaning.
From a Machine point of view, this is correct because AntiGooglelianos is a query term and once it implies 1 result then it will also imply new meta-term queries.
This is the idea:
I have created a word.
I have created a result.
Can I control then potencial results of my Word?
The thing about Antigooglelianos… The thing is that right now Antigooglelianos is a Vector. I THINK GOOGLE gave birth to it with a single Antigooglelianos link anchor.
A simple technique:
1 anchor link –> h1.
The we copy and paste a surfed text.
And we apply a secret I THINK GOOGLE tecnhnique.
Right now we have to face the real problem:
Antigooglelianos is a real vector!
How can be this possible?
It is not relevant,
it got no meaning,
it doesn´t exist…
Re-reading LSI basics, If “documents containing these terms should be more relevant to the initial query”, then we have started a legimite SEO circle with new terms.
So in a 1-1 Model, 1 link implies 1 result, 1 anchor implies 1 result
If 1 anchor/link implies 1 result then 1 result implies 1 anchor/link
This looks like a tautology but obviously it is not.
This is the Angooglelianos Problem.
A big-time problem if you understand:
what I mean
the concept of model
the Google inexorability
AntiGooglelianos was mine.
I created it.
But from the very first time it was crawled, indexed, and stored, Antigooglelianos is no longer mine.
It is quite amazing to find out that a human being can actually create a new word and can spread this new information around the world. But today we can openly say that the meaning of the Machine and the meaning of the human being are positively not the same. As all major Human Beings search incorporate unsual anchor text and link forms in their documents, search engines use sophisticated algorithms that manipulate content as well as linkage into their ranking schemes
So what is the purpouse of a term count model? AntiGooglelianos itself is not a topic neither a keyword. It is virgin and cannot be taken as popular or a financially lucrative query. We are not teaching. We as Human Beings are giving to the Machine the opportunity of Learning because we are like Contexts.
Machine might learn to think but if and only if we – as context – tell the truth.
Search Philosophy lesson 2
In implementations consistent with the principles of the invention of Google 2007:
the history data of the document ANTIGOOGLELIANOS may include data relating to: document inception dates; document content updates/ changes; query analysis; link-based criteria; anchor text (e.g., the text in which a hyperlink is embedded, typically underlined or otherwise highlighted in a document); traffic; user behavior; domain- related information; ranking history; user maintained/generated data (e.g., bookmarks); unique words, bigrams, and phrases in anchor text; linkage of independent peers; and/or document topics.
So in a 1-1 Result Model,
which is the main ranking component?
Is the result itself?
In response to a search query the main ranking component generate an initial set of relevant documents, so is the 1-1 Result Model the initial set of relevant documents, the First Set?
Can be a single document be a set of documents?
Search Philosophy lesson 3
When the documents are associated with a search query (e.g., identified as relevant to the search query), search engine may sort the documents based on the ranking score and return the sorted set of documents to the client that submitted the search query?
- So, what is the real meaning of “relevant” for one result that is the only one, a 1-1?
- So, what is the real meaning of “relevant” for one result that has no meaning?
- So, what is the real meaning of “relevant” for one result that is a set of itself?
As we see in a 1-1- Result Model the ranking of the initial set of relevants documents is the same than the relevance score for each document in the initial ser of documents.
The foregoing definition then implies that receiving the search query Antigooglelianos from a user, the list of relevant documents are based on the match between the term of the query, the anchor of the result and in the identity between the relevance score and the relevance of the document.
That means that in order to refine the relevance score of Antigooglelianos, we can create a score value only using a “are referenced by other documents in the set of documents technique”.
Consistent with aspects of the invention, the ranking score is a value that attempts to quantify the quality of the documents.
In implementations consistent with the principles of the invention, the score is based, at least in part, on the history data from history component.
This score is associated with a document based, at least in part, on user maintained or generated data.
“According to an implementation consistent with the principles of the invention, information regarding document topics may be used to generate (or alter) a score associated with a document. For example, search engine may perform topic extraction (e.g., through categorization, URL analysis, content analysis, clustering, summarization, a set of unique low frequency words, or some other type of topic extraction). Search engine may then monitor the topic(s) of a document over time and use this information for scoring purposes.”
I do not think so.
Search Philosophy lesson 4
What is a Keyword? What is a Topic?
It is quite simple.
Keyword and Topic deep down does not exist either: we create them.
Once we create Keywords and Topics we can postively assure they are a topic or a keyword.
The Machine will agree.
We master – as Human beings – the decision.
And the most important thing: without being “sponsored”.
The Machine will require specific detailed skills or knowledge tuned to a specific task (knowledge engineering bottleneck).
We will give the Machine a new task, new knowledge.
In the near future, the Machine will “discover” new knowledge – my new word – from its large databases (data mining) once crawled, indexed, and stored.
The Machine will apply numerical optimization, Probability, Statistics, and so on. And this is the funniest thing of this meaningless new word ANTIGOOGLELIANOS. Its Absurd Power.
I give the Machine a new word to classify in its database: but the percentage of words correctly classified or use of a Database of human- labeled new words does not mean the Machine understands what AntiGooglelianos means.
Classified does not mean Understand.
Human Being Mash
Human Being label.
The thing is that we can create and manipulate Keywords and Topic because we master the Language.
This implies Topics and Keywords are under our control. And at the same time out of our control.
All this underscores the fact that users perception of relevancy and machine perception of relevancy (relevance scores) are 2 different things.
This is a fascinating problem we have to face.
Given ANTIGOOGLELIANOS sample input and output pairs for a useful target function.
let checker boards labeled with the correct move, allow be extracted from record of expert play.
Be a potentially arbitrary sequences of game moves and their final game results.
How the Machine will assign credit blame to individual moves given only one indirect feedback?
Will the Machine asisign it because this is a 1-1 result?
The Machine will learn. The Machine will can query an “unlabeled” word.
So Machine can construct an arbitrary example and query an oracle for its label.
In fact, Machine can design and run experiments directly in the environment without any human guidance but only
Because I believe the Machine is a Checker, I am playing with the Machine.
This is my choice:
I Choose a Move.
I have a board.
My word is “legal”
So I create new moves
ChooseMove (board, legal-moves) ? best-move
“Or could learn an evaluation function, V(board) ? R, that gives each board position a score for how favorable it is.
V can be used to pick a move by applying each legal move, scoring the resulting board position, and choosing the move that results in the highest scoring board position..”
As a Human being, I can train the Machine creating a single word that estimate training values for intermediate (non-terminal) serps positions by the estimated value of their successor in an actual game trace.”
I mean: anchor/link –> H1, where successor(b) is the next board position where it is the program’s move in actual play.
“Values towards the end of the game are initially more accurate and continued training slowly “backs up” accurate values to earlier board positions.”
This is basic in SEO.
What I am doing with my new ANTIGOOGLELIANOS word is conducting controlled cross-validation search experiment to compare various methods on a variety of Topics/keywords-benchmark datasets.
I have the ability to fit training data. The Machine will debug.
All pages might contain more than one element. In this sense, determing a set of results based on the search terms, we can decided which Keywords and Topics are the best ones. Even more: we should create them.
I THINK GOOGLE has the ability to fit training data and let the Machine be a Machine.
On the other hand, let check my domain name I THINK GOOGLE. It is a overloaded Matching-sentence. In a way, is a Topic – number 1 in 124 millions -. Far away from this search situation, my word Antigooglelianos is a new word in a 1-1 model, a 1-1 result …
So, should be taken as a keyword?
Again, in implementations consistent with the principles of the invention of Google 2007:
If “All pages are refered …”
What should be done, Human being…?
Which are the next steps should be taken?
This is the final Search Philosophy Lesson:
Before of including a relevant link to this text link Traslation of Antigooglelianos using Google … I mean, a link to the automatic google translation service, obviously I have to upload my document in advance.
So I cannot add a Link to the translation before the document has being uploaded.
This is impossible.
That´s logically means that Links come after documents.
The conclusion must proceed from primary premises that are indemonstrable premises, for one cannot know things of which one can give no demonstration, since to know demonstrable things in any real sense is just to have a demonstration of them. The premises must be Causal, Better known and Anterior; Causal, because we only know a thing when we have learned its cause, Anterior because anteriority is implied by causation, previously known not only in our second sense, viz. that their meaning is understood, but that one knows that they exist. by Aristotle
The Results are Documents … only after being linked?
In a 1-1 Model, a result is not only an Output. In the same way, a query is not only an Input.
Antigooglelianos is not a query: is an input.
Antigooglelianos is not a result: is an output.
The SEO task is create an input that is a result; and an output that is a query.
Thank you for reading!
Do you want to read more about Search Philosophy?