Friday, September 12, 2014

Google Aims to Archive All Human Knowledge

It says something to know that this can be contemplated. It is also excellent news as vast amounts of archived material can be made available. That this is truly huge is often not well understood. Yet imagine the mining industry alone which gathers hard data daily and has been doing so for centuries. Most of that was simply lost. Yet It all had value. That means even the raw data which may be thousands of individual assays.

The importance of doing this cannot be underestimated. Most interpretation is about a statistically significant number of eyeballs.

Thus a system in place that squirrels away data will stike gold.

Google Aims to Archive All Human Knowledge

AUG 21, 2014 01:59 PM ET // BY GLENN MCDONALD

New research details emerging this week suggest that Google is after the truth, the whole truth, and nothing but the truth.

According to an intriguing report in New Scientist, Google is building a next-generation information database called Knowledge Vault that’s designed to index and store what we can reasonably term facts. And not just some facts — the Vault is intended to continually catalog and store all facts about our world and our history.

It’s certainly an ambitious plan, and likely beyond the capability of even the largest crowdsourcing initiative. But that’s all right, because Knowledge Vault is fully automated and uses algorithms to turn raw data gathered online into bite-sized, usable chunks of knowledge.

What’s more, because of the nature of the knowledge base, the information would be readable by both machines and humans. You could query Knowledge Vault directly, as with a Google search. Or you could rely on your future smartphone — or digital assistant or robotic helper — to do the searching for you.

The Vault project is building upon Google’s existing crowdsourced database, Knowledge Graph, and so far has cataloged about 1.6 million facts. Google researchers will present a paper on Knowledge Vault next week at the Conference on Knowledge Discovery at Data Mining, in New York.

It’s all part of a larger initiative, in the information technology arena, to improve the manner in which we interact with machines and databases. Similar knowledge bases are being built by companies like Facebook, Amazon and Microsoft and IBM.

One of the first practical applications for these ultra-database systems is to create a new generation of virtual personal assistants. In other words, Siri is about to get a lot smarter and faster.

Down the line, the Knowledge Vault could serve as the foundation for advanced augmented reality networks. The database would provide instant data, via heads-up display, on virtually anything you look at. The Knowledge Vault could also be used, eventually, to model all of human history and society as a vast collection of pure data. That knowledge, in turn, could be extrapolated to make predictions about the future.

Turning all of history into data and prophesying the future? Those Google guys sure don’t monkey around.

No comments: