Hakia - Search for Better Search

Hakia recently initiated “The Search for Better Search” initiative via a focused poll taken from some of Web 2.0's best technical blogs. The results reveal an overwhelming and compelling need for a better search capability. Make no mistake about it; raising the bar for the people at hakia has nothing to do with hype or beating Google really, but about the art of transcendence. The vision there is about elevating the world's expectations and thinking so that search and the Web can transcend to the next level.

For those of you just arriving from another galaxy, hakia is a new search entity based on the vision that search is about meaning derived from context and relevance. To a logical person of a certain type nothing could be simpler really, but the methodology and technology are of course rather much more complicated. Trying to explain hakia's structure has been a difficult task for me; even after conversations with Dr. Riza Berkan, CEO of hakia I was enlightened to the point of illumination but still woefully inept to describe the hakia processes to my friends and readers.

Yesterday I was delighted to chat with Melek Pulatkonak, COO and President of hakia, and once again I was fascinated to talk with a person who is gracious, supremely intelligent and without the slightest inkling of subterfuge. Melek helped me to realize quite a few things I was having some trouble with and, she took an inordinate amount of time answering even the most simplistic questions. I make these points for one purpose only, and that is to illustrate that symbolically everything is important. Excellent people and visions are the gateway to fantastic innovation. When I tell our readers about a brilliant rocket scientist lady who explained patiently a very complex concept, this is a symbol of what our expectations should be at the least.

The results of the Search for Better Search poll reveal a similar symbolic truth. Nearly 65 percent of the technology crowd at the most influential blogs on Web 2.0 revealed their insight for the need of improved search capability and relevance.

The poll was conducted from April 1^st thru April 26^th and is linked along with the commentary to the following sites: Battellemedia; ResourceShelf; GigaOM; TechCrunch; VentureBeat; Read/WriteWeb; and Ars Technica. 

Results of the Poll “What do you think a search engine should do?” 

• 39% understand my question the way a human does. (165 votes)
• 25% bring highly relevant results with semantic precision. (107 votes)
• 9% give me more control of search options. (39 votes)
• 7% have user interface that is simple and powerful  (31 votes)
• 5% allow me to participate to rank search results. (21 votes)
• 5% learn my search behavior in a trusted environment. (20 votes)
• 4% chat with me to help with my search. (19 votes)
• 3% provide personalization tools and agents. (12 votes)
• 3% none of the above! (11 votes)

The Real Issue

Search and the Web as we know them are still in a phase of development. People naturally cling to familiar and useful utilities, places and technologies. We are simultaneously repelled by innovation and change and drawn to the possibilities this dynamic reveals. What we really want is the ability to be understood and helped by the tools we employ. Hakia in their effort at science, art and commercial enterprise (in that order I might add) is addressing one of the most important and complex problems that exist in the next evolution of the Web.

According to Dr. Berkan and Melek the single most important part of the semantic search is the “middle layer” of the technology that is the heart of the attempt to produce a true representation of how the human mind performs searches. In his own words Dr. Berkan provides a much more eloquent explanation of this aspect and why previous attempts have failed to provide a solution.

“A long list of trials has been attempted, all avoiding to change the middle layer technology to enable a full-scale, compromised, semantic natural language processing (NLP). NLP and Web search do not go hand-in-hand unless the middle-ware is invented. Once this obstacle is out of the way, then the possibilities become endless.”

This middle architecture's key element is the QDEX system of data storage gateways to replace the vertical index systems in use so far today. This “top secret” and revolutionary system allows semantically rich data to be processed very rapidly. No vertical index system could ever process semantic data quickly enough to provide a result within milliseconds. Neither Dr. Berkan nor Melek Pulatkonak would represent this aspect of their process as anything beyond advanced technology. These are not secretive people, but they are also not presumptive scientists. I mentioned what I thought of the entire system in both conversations and was answered both times with a kind and knowing laugh.

What you are essentially looking at in the QDEX aspect is at the least a part of what could be considered a brain. Melek will probably kill me for saying this, because she made no claim of such a giant leap from stored, mathematically manipulated data to sentient intelligence. This is why other semantic projects have failed in the past, there was no representation of an element that truly emulated the way a human brain actually functions.

So, how does this “brain” or portion thereof communicate the billions of electrochemical impulses and communications in order to understand in human language a concept or idea? The answer is Ontological Semantics. With OntoSem imposed on our brain (replete with cells) what we have done is infused a system of analysis closely resembling our own reasoning system. OntoSem is constantly researching meaning-based inferencing and reasoning according to the hakia labs description.

What this and the auxiliary capacity of the SemanticRank algorithms essentially approximate or emulate is the way human beings understand language, meaning and ultimately queries from searches. Combine all this with the Dialogue Communications System being developed and the leap from simple computer data base to AI is complete. Hakia is developing what I call “the lower limits” of artificial intelligence. I have been wrong before, but I bet $100 bucks my illustrious friends at hakia will not fault my assumption, especially given my limited scope of interaction with the system.

Actually, being right or wrong is of little import. What is crucially important is for my readers and people everywhere to elevate level of inquiry, thinking, expectations and anticipation of what could be one of the breakthroughs of the 21^st century. Isn't it more interesting to dream and anticipate that these wonderful scientists are endeavoring to create something like I describe for humanity?

I can only encourage you all to visit the hakia labs pages and Ontological Semantics to try yourselves to grasp these complex and fascinating concepts. The poll illuminates the knowing need for something a step up on the evolutionary scale and the people and technology of hakia are demonstrating the probability that we can get there.

My mash up of how to visualize hakia as the beginning of a new search, what does it remind you of? Ok squint if you must then.

If you enjoyed this post, make sure you subscribe to profy RSS feed!