Features | i-gat

GASET’s Main Functions

GASET’s search engine has four main unique functions not available in other common or specialized search engines. It’s AI motivated special capabilities make it possible to outrank any similar technology available commercially. We hope that the following summary will give your organization a clear and informative understanding of our project’s features. We think that a critical look by search engine experts and consultants will assist you in anticipating our project’s real value.

Power Search

The main idea behind Power Search is to assist the web surfer by anticipating multi level web searching tasks. This will be done by using the web as a base for our self modified and customizable re-learner agent. Its auto query adjustments and multistage searching and its re-defined/re-adjusted web repository, will lead to the adaptation of conceptual base construction by re-implementing its newly acquired conceptual/contextual knowledge foundation. Such technology is in place to assist in determining the web pages identities, semantic and perception relations with other web pages/ sites in a way that take into account interchangeable compatibility in regard to serving web surfer’s needs to have complex searching process, compatible with the density levels of his query.

Our innovation will enable our system to understand a query, such as “car manufacturing”, as a task which needs to be achieved. It will then find resources, solutions, technologies … etc, related to the specific query and organize it in one block of information along with other related links despite its initial location on the web. Its profound dependency on self-directed multitasks handling functions, advance interactive interface, its relay on rigid and dynamic formulas that autonomously double check its own mechanism and results for accuracy, guarantee that the web surfer will deal with a friendly system that will save him/her time and provide the most comprehensive, multifaceted results.

Topic Search

Will be used by web surfers to locate web pages with topics which match his/her query. We use a pioneering method to decipher web page framework (either in parts or as a whole) and concept. It will also check multiple facets of the page in regard to its relevance to already stored, analyzed and confirmed concept containers that our system uses for evaluation, ranking and re-modifying our web repositories. If, for example, the topic search was “car manufacturing”, the match will be semantically/ syntactically correlated with sentences and words such as, machinery, raw material, tools, sales, testing ... etc in a harmonized linguistic form. Unlike the practice followed by the entire main search engines, the results and pages containing the matches will be more related to the topic than the calculating of the number of times the query words “car” and “Manufacturing” repeat on the web pages regardless of its semantic relation.

We used complicated logarithmic technology to enable our system to constantly self modify and configure its own rules and ranking factors in an attempt to keep the unique linguistic forms, dialogue types and cultural lingo differences in prospect. Such procedures will enable the project to be updated autonomously.

QA (Question Answering)

Considered by our staff to be the project’s “Gem”. This section of our project is capable of analyzing (based on a strong dynamic and specialized linguistic knowledge base) the user’s complex and multi-facets question and then locating, generating and customizing the proper answers. Such technology employs hand crafted and web extracted linguistic rules and patterns which are capable of virtually realizing the question concept and its contextual relations to other parts of its linguistic block. It will then find or customize a sentence or a paragraph which meets the answer criteria in a logical manner (optional corresponding thesaurus, expected forms of answers… etc).

This task will be done by anticipating the searcher’s question category based on foreseeing related questions – answers results compiled from our “QA” dynamic and specialized knowledge database. This is done by engaging the user in some sort of a dialogue which is extremely helpful in narrowing the received results (such technology will be employed upon the full integration of the needed web resources into our system – a huge hardware dependency task which we don't possess as of yet).

Multimedia Search

The future of web surfing will be determined by the user’s need for strong analytical application capable of tagging and grouping massive amounts of multimedia resources that use the internet as a storage facility. Our ranking technique will be the other decisive factor. It will be based on the concept from its associated information (not only titles and other html tags). Information in the form of critiques and opinions (without the downloadable file) available on other locations, and conceptually related to the query subject, will be used for enhancing and classifying the similar downloadable files.

Our multi-faceted NLP parser will set the stage for the needed item classifications, which will be used to search specialized multimedia web resources such as photos, books, papers, videos, audios, software … etc. coupled with an advance in customized matching technology using conceptual, contextual and efficient methods. In contrary to other methods, we regard the web resources as ours to use without a need to re-categorize and re-restore it in specialized web sites (such as: Google™ video, YouTube ... etc). We have succeeded in creating special techniques which we use to tag and filter such multimedia sources in its original location.

Search Results Page

Improvements to the regular search engine results page were introduced in order to give the searcher an option to choose between receiving our version of the result page snippets or:

1. Related page Keywords list (most important related words).

2. Complete related Paragraphs (paragraphs matching the query concept).

3. Downloads (list downloadable parts, used with multimedia specialized search).

Our friendly “scroll down window”, when used, will keep the search results page proportionate while giving searchers the ability to view page concept without the need to open the actual web pages. In the near future our system user will be able to view related parts of matching web pages without the need to leave our GASET result page. Our UBO technique “User Behavior - results analyzing –Observation” will save more time by identifying web surfer’s results page analyzing patterns and his/her subjects of interests.

GASET AI inspired B2C Platform

As mentioned in GASET project’s profile document, such technology will be used to launch an intelligent B2C platform which will be a part of the GASET initiative (similar to Google Froogle system). Some of GBSET business oriented technologies will be adapted and modified to meet the consumer’s needs. B2B virtual dealing, E-Negotiating and business tasks handling, with its payments and security related issues will be modified to meet such challenges. This application has great potential with the GASET search tools (QA, Power Search …. etc) playing a decisive role in its success. Interactive technology will assist business owners in their marketing campaigns since it provides customers with hassle free services. Revenue generating method will be taken into consideration as a part of GASET complete financial profile.

Samples of other services

The tools used are superior to other similar tools employed by standard search engines due to the application of special web searching techniques that are based on conceptually, semantically and contextually analyzing factors. (Details available upon request):

GDM Concept Suggestion: will find directories based results conceptually matching user query

Similar Page: will locate pages similar to the ones the user found to be matching the original query

Search within Results: by enhancing user query to locate similar web pages to the one he/she favored

Local Search: will search within the user’s desired geographical locations

G.A.S.E.T System - Features

G.A.S.E.T’s Main Functions

G.A.S.E.T’s search engine has four main unique functions not available in other common or specialized search engines. Its AI-motivated special capabilities make it possible to outrank any similar technology available commercially. We hope that the following summary will give your organization a clear and informative understanding of our project’s features. We think that a critical look by search engine experts and consultants will assist you in anticipating our project’s real value.

 Power Search

The main idea behind Power Search is to assist the web surfer by anticipating multi-level web searching tasks. This will be done by using the web as a base for our self-modified and customizable re-learner agent. Its auto query adjustments and multistage searching and its re-defined/re-adjusted web repository will lead to the adaptation of conceptual base construction by re-implementing its newly acquired conceptual/contextual knowledge foundation. Such technology is in place to assist in determining the web pages identities, semantic and perception relations with other web pages/ sites in a way that takes into account interchangeable compatibility regarding serving web surfer’s needs to have a complex searching process, compatible with the density levels of his query.

Our innovation will enable our system to understand a query, such as “car manufacturing”, as a task that needs to be achieved. It will then find resources, solutions, technologies … Etc., related to the specific query and organize it in one block of information along with other related links despite its initial location on the web. Its profound dependency on self-directed multitasks handling functions, advanced interactive interface, its relay on rigid and dynamic formulas that autonomously double-check its mechanism and results for accuracy, guarantee that the web surfer will deal with a friendly system that will save him/her time and provide the most comprehensive, multifaceted results.

 Topic Search

Will be used by web surfers to locate web pages with topics that match his/her query. We use a pioneering method to decipher web page framework (either in parts or as a whole) and concept. It will also check multiple facets of the page regarding its relevance to already stored, analyzed, and confirmed concept containers that our system uses for evaluation, ranking, and re-modifying our web repositories. If, for example, the topic search was “car manufacturing”, the match will be semantically/ syntactically correlated with sentences and words such as machinery, raw material, tools, sales, testing ... Etc. in a harmonized linguistic form. Unlike the practice followed by the entire main search engines, the results, and pages containing the matches will be more related to the topic than the calculating of the number of times the query words “car” and “Manufacturing” repeat on the web pages regardless of their semantic relation.

We used complicated logarithmic technology to enable our system to constantly self-modify and configure its own rules and ranking factors in an attempt to keep the unique linguistic forms, dialogue types, and cultural lingo differences in prospect. Such procedures will enable the project to be updated autonomously.

 QA (Question Answering)

Considered by our staff to be the project’s “Gem”. This section of our project is capable of analyzing (based on a strong dynamic and specialized linguistic knowledge base) the user’s complex and multi-facet questions and then locating, generating, and customizing the proper answers. Such technology employs hand-crafted and web-extracted linguistic rules and patterns which are capable of virtually realizing the question concept and its contextual relations to other parts of its linguistic block. It will then find or customize a sentence or a paragraph that meets the answer criteria in a logical manner (optional corresponding thesaurus, expected forms of answers… etc.).

 Multimedia Search

The future of web surfing will be determined by the user’s need for strong analytical applications capable of tagging and grouping massive amounts of multimedia resources that use the internet as a storage facility. Our ranking technique will be the other decisive factor. It will be based on the concept from its associated information (not only titles and other HTML tags). Information in the form of critiques and opinions (without the downloadable file) available in other locations, and conceptually related to the query subject, will be used for enhancing and classifying similar downloadable files.

Our multi-faceted NLP parser will set the stage for the needed item classifications, which will be used to search specialized multimedia web resources such as photos, books, papers, videos, audios, software … Etc.. coupled with an advance in customized matching technology using conceptual, contextual and efficient methods. Contrary to other methods, we regard the web resources as ours to use without a need to re-categorize and re-restore them in specialized websites (such as Google™ video, YouTube ... etc.). We have succeeded in creating special techniques which we use to tag and filter such multimedia sources in their original location.

 Search Results Page

Improvements to the regular search engine results page were introduced in order to give the searcher an option to choose between receiving our version of the result page snippets or:

1. Related page Keywords list (most important related words).

2. Complete related Paragraphs (paragraphs matching the query concept).

3. Downloads (list downloadable parts, used with multimedia specialized search).

Our friendly “scroll down window”, when used, will keep the search results page proportionate while giving searchers the ability to view page concepts without the need to open the actual web pages. Shortly, our system users will be able to view related parts of matching web pages without the need to leave our G.A.S.E.T result page. Our UBO technique “User Behavior - results analyzing –Observation” will save more time by identifying web surfers’ results page analyzing patterns and his/her subjects of interest.

 G.A.S.E.T AI-inspired B2C Platform

As mentioned in the G.A.S.E.T project’s profile document, such technology will be used to launch an intelligent B2C platform which will be a part of the G.A.S.E.T initiative (similar to the Google Froogle system). Some of G.B.S.E.T's business-oriented technologies will be adapted and modified to meet the consumer’s needs. B2B virtual dealing, E-Negotiating, and business task handling, with its payments and security-related issues, will be modified to meet such challenges.

This application has great potential with the G.A.S.E.T search tools (QA, Power Search …. Etc.) playing a decisive role in its success. Interactive technology will assist business owners in their marketing campaigns since it provides customers with hassle-free services. The revenue-generating method will be taken into consideration as a part of G.A.S.E.T's complete financial profile.

 Samples of other services

The tools used are superior to other similar tools employed by standard search engines due to the application of special web searching techniques that are based on conceptually, semantically, and contextually analyzing factors. (Details available upon request):

• GDM Concept Suggestion: will find directory-based results conceptually matching the user query

• Similar Page: will locate pages similar to the ones the user found to be matching the original query

• Search within Results: by enhancing user query to locate similar web pages to the one he/she favoured

• Local Search: will search within the user’s desired geographical locations.

Summary of G.A.S.E.T. System Features