BII Academy

Put theory into practice

June — Wed, 29 Sep 2010 07:39:41 +0000

In July we talked about Prof. Dr. habil. Ulrich Reimer paper “Learning a Lightweight Ontology for Semantic Retrieval in Paitent-Center Information Systems“. Now we will illustrate how this can be applied in day to day applications.

We have developed a trend barometer that uses the ai-one core technology to observes and analyze for example the Internet (news platforms, online news, RSS feeds, blogs etc.) The trend barometer finds in context and topics discussed the current keywords, that is the semantic trends, and builds a dynamic ontology on a daily basis. Similarly, ai-one can be applied as trend barometer or analysis tool on documents or databases. This opens up the possibility to compare documents, even databases, as regards content. The number of possible applications are almost infinite.

Von der Theorie in die Praxis

In Juli haben wir über Prof. Dr. habil. Ulrich Reimer wissenschaftliche Veröffentlichung “Learning a Lightweight Ontology for Semantic Retrieval in Paitent-Center Information Systems” berichtet. Jetzt werden wir aufzeigen, wie diese Erkenntnisse in der Praxis angewendet werden.

Wir haben den ai-one Trendbarometer entwickelt. Der Trendbarometer beobachtet und analysiert z.B. das Internet (News Plattformen, Online-Zeitungen, FEEDs, Blogs etc.) und ermittelt aus dem besprochenen Kontext und Themen, die tagesaktuellen Schlüsselwörter, also die semantischen Trends, und baut daraus eine dynamische Ontologie. Anstelle vom Internet, kann ai-one als Trend- oder Analyse-Barometer auf Dokumente oder Datenbanken angewandt werden. Daraus erschliesst sich die Möglichkeit, Datenbanken oder Dokumente inhaltlich untereinander zu vergleichen. Es gibt eine fast unermessliche Vielzahl von möglichen Anwendungen.

Why do iPhone or Mac users get so excited?

June — Wed, 25 Aug 2010 14:11:27 +0000

One of the things iPhone or Mac users in general get so exited about, is the search function in those devices. The users start to type, and immediately the search function displays what it finds… The more the user keys in, the more accurate the answer. The ai-one technology offers exactly the same user comfort and tops it with its semantic functions.

Furthermore it is often the case that the users don’t know how to phrase their query or how to start the search, in exactly this situation the ai-one semantic technology plays to its strength in proposing the right question! This is one of the many applications of ai-one. Brainup Systems GmbH launches now the first version of the application: *Brainup data intelligence*. Users can experience how useful an application is, that can make sense and understand the content of the text. Using the application is as easy as the search function in Apple’s devices.

The secret of understanding content and context is to recognize the inherent (intrinsic) semantics, which is based on the data fed into the application. ai-one can autonomously recognize the intrinsic semantics, the meaning of words used in a certain content or context. ai-one automatically makes the semantic and associative connections and builds a context ontology of the text.

Weshalb sind Mac und iPhone User so begeistert?

Eine der Funktionen im iPhone oder im Mac, die von allen Nutzern besonders geschätzt wird, ist die Suche in den Apple Geräten. Der Nutzer beginnt einfach, zu tippen und die Suchapplikation baut sofort erste Antworten auf. Je mehr und genauer der Nutzer die Suchanfrage formuliert, um so exakter die Antworten. Die ai-one Technologie offeriert exakt den gleichen Bedienkonfort und darüberhinaus mit semantischen Funktionen.

Doch für viele Nutzer stellt sich das Problem, wie die Frage überhaupt formuliert sein soll, um etwas bestimmtes zu finden. Genau da hackt ai-one ein und hilft dem Nutzer dank den automatischen semantischen Funktionen die richtige Frage zu finden.

Das Geheimnis ist, Inhalte, Worte und Bedeutungen automatisch zu erkennen und die inheränte (intrinsische) Semantik (Bedeutung) zu verstehen. ai-one kann eben diese Zusammenhänge herstellen und Schlüsse ziehen. ai-one zeigt dem Nutzer die vorhandenen Zusammenhänge und bildet diese Ontologien selbständig auf.

Brainup Part 2

June — Tue, 13 Jul 2010 14:52:44 +0000

Brainup Part 1

June — Tue, 13 Jul 2010 14:49:44 +0000

Michael Krause and Walter Diggelmann are talking about Brainup and data intelligence. The company Brainup Systems GmbH has developed with brainup a novel and innovative solution for private and and professional users that implements the semantic technology of ai-one.

The genetic rules

June — Thu, 10 Jun 2010 07:06:51 +0000

The power of our approach lies in the genetic rules, the Hoffleisch’ Algorithm. This set of instructions is an universal and natural algorithm. This set of instructions enables the computer system to autonomously separate content from form, to attribute meaning to a word, to associate it to other things and to classify. The association can be attributed after encountering certain pieces of data or after a certain threshold. This behavior emulates the self-organizing principles of nature.

Each person stores information and views the real world in his or her own representation of the world. Programs with this set of genetic rules can ’see’ through this individual’s view and automatically detect the inherent information pattern and concepts in the entered data. The more data imported and the more interaction with the user, the more knowledge these programs gain. And now, the program also understands associative search (that is that the searching person does not need to search for an exact term). The search can be initiated with terms and nouns which we ourselves would associate with the search term. But searching with an exact term is equally possible and the program suggests terms, which could be associated with it. For example, the program might suggest to “summer” “hotel”, “book a flight”, or “rent a car” (if this terms were in any of the imported data). Programs based on this algorithm detect inherently existing patterns, of which we did not know of before hand and therefore could not even inquire about.

Tap into the wealth of your databases

Is this not what data-mining is supposed to do? To find answers to questions we may not have even thought of? To find patterns and associations? Brainup is ongoing project to unlock the knowledge lurking in our legacy systems, databases, archives, even on our workstations.

Intelligence in Computing Part 3

June — Fri, 04 Jun 2010 13:15:28 +0000

Intelligence in Computing Part 2

June — Fri, 04 Jun 2010 13:14:43 +0000

Intelligence in Computing Part 1

June — Fri, 04 Jun 2010 13:12:33 +0000

What can we achieve with intelligence in computing? Watch the video where Walt interviews Dr. Jim Hardt and Manfred Hoffleisch. In this interview Jim, Manfred and Walt tackle issues like having the computer behave in an open ended way – the way a human brain behaves and the advantages of the computer that can think like a biological brain.

Holistic approach or how ai-one™ helps solve the information overload

June — Fri, 04 Jun 2010 07:59:22 +0000

To date, the solutions for business intelligence are based on limited boolean logic and are not associative at all. Furthermore, the “garbage in garbage out” principle applies: A computer can not distinguish if the entered data is valuable or relevant. A search in a database only delivers useful information, when the person already knows before hand, what exactly he or she is searching.

Humans on the other hand store a piece of data in structures and by associating it with other things. When we recall an information, we have several alternative associative routes to retrieve the memory, Take the terms “warm”, “blue sky”, “ice cream”, “sunshine”, “barbecue”, “beach” and palmtree, there is a tendency to immediately form an associations with “summer”. Which a search in traditional business intelligence solutions and with traditional methods won’t deliver.

However, our holistic approach based upon the holosemantic neuronal network incorporates both the boolean logic and the creative and associative thought processes of humans. The holosemantic neuronal network and data space combined with the linguistic and genetic rules enables the computer to learn, to think and to reach decisions just in the same way as humans – without pre-programmed logic. This is biologically inspired intelligence.

In our third and final part we go into the details of the basis of biologically inspired intelligence.

Where is the meaning?

June — Thu, 03 Jun 2010 21:11:38 +0000

How do we cope with one of the biggest challenges in IT, the information overload? Theoretically, all data should be accessible, once it is entered into a computer. Indeed, the growing variety of business intelligence solution, operating systems, formats and databases adds to the challenge. However, inaccessible data is lost knowledge.

How do we access the knowledge?

Paradoxically, knowledge in a company is linked to a certain person or persons. This means that only with the help of this person can the wealth of the archives be unlocked. The creativity and knowledge of this person is based upon association. Knowledge is not just a collection of information, but it stands in context,with something else. A piece of data gets a meaning for us when we associate it with other things. When we encounter a piece of data, our first action is usually to attempt to find a way to attribute meaning to it. And we do this by associating it with other things.

The semantic trap!

It’s so easy for humans, to understand semantic or meaningful information. We either understand it immediately or we ask because we do not understand. At the latest, if we have to raise and educate children, we realize, that there is so much useless information out there. And we also realize that if we feed useless information, we corrupt the quality of exiting knowledge – the GIGO principle (Garbage IN = Garbage OUT). “We must be carefully, what we teach to a semantic system, it’s like educate a human” … says Manfred Hoffleisch, head of R&D at ai-one inc.

Next: how ai-one™ helps solve this problem.