Linked Open Data Technologies are under the engine hood of more and more applications and services. Search engines development was a field were lots of pioneer work was done, in recent years, and were now powerful products came and will come to market. I have had the chance to talk to Amar-Djalil Meazour about the vibrant sector of semantic search.
Thomas Thurner: Exalead is one of the few European market leaders in search engine technology. You are also one of the Use Case partners in the LOD2 project. What are Exalead’s interests and motives in participating in the project?
Amar-Djalil Meazour: Exalead is the only remaining European actor which indexes the Web and provides a public interface to this interface (on exalead.com/search). In June 2010, Exalead was acquired by Dassault Systèmes, opening vast new market perspectives for our search-based solutions. We are redesigning our strategy to adapt to the requirements and challenges of our new customers, building more business applications to efficiently address customers’ technological constraints, domain requirements, etc. We are currently building vertical search-based applications on top of our generic technological platform CloudView to address industries, domain by domain. Human Resources is one of the strategic domains that Exalead is targeting, with already successful customer use cases with fish4, Randstat, Sogeti, and others.
One prime differentiator with our competitors is the use semantics in every stage of the application workflow. Capturing meaning of unstructured and semi-structured information is an unavoidable challenge for turning web-based information into actionable data. For this, Exalead invests intensively in R&D in Natural Language Processing techniques, in Information Extraction, and text and data mining. And continuing our Web based traditions, we invest in continuing to increase the efficiency and power of our underlying information Retrieval technology.
We feel our involvement in LOD2 is crucial for the future of search based applications. As linked and open data continues to grow, it is imperative to not only index this data, but to integrate open linked data with internal enterprise data and with data extracted from unstructured web and professional sources. Since LOD2 addresses real-life applications and use cases based on the linked open datasets, we know that lessons learned will bring great added value to domain applications, which is perfectly in line with Exalead’s new strategy and makes our involvement in this project straightforward. By leading the enterprise use case, we expect that LOD2 will provide us with new advanced features and technologies for addressing our HR customers with our HR portfolio.
Thomas Thurner: Can you tell us a little bit more about the LOD2 use case Enterprise Linked Data that is lead by Exalead please?
Amar-Djalil Meazour: In enterprise IT environments, company data is stored in repositories and databases spread over different locations and accessible through different applications. This data is often duplicated when processed, leading to both unnecessary redundancies and gradual inconsistencies as data is changed in one place and not the other. In addition, most IT applications have little semantic processing, so that accessing information can be hit-or-miss for users unacquainted with the exact form in which the data is stored.
The web, with its wealth of data, should provide a source for enriching internal data, for example by supplying missing information about entities stored in enterprise IT software. But the promise has so far remained unrealized, as semantic web techniques have only begun to spread.
Internally the same confusion reigns, as corporate IT systems, using different software applications to manage the same data, impose different incompatible formats, making the dataflow processing tough and cumbersome. Software heterogeneity feeds the massive development efforts needed to make diverse vendor data formats interoperate. In a semantic web world, where data is freed from format, yet explicitly identified, and reusable, much of the current unnecessary IT communication problems will disappear leading to more efficient and more powerful IT applications.
To prototype these future solutions, the LOD2 enterprise linked data use case imports the LOD paradigm and practices in the enterprise environment to enhance performance of IT systems. We will develop an integrated application that uses the intercommunicability of linked data to cross enterprise boundaries between enterprise IT systems (ERP, HR ERP, CRM, etc.). In this use case, we harmonize data formats using standard vocabularies, applying semantic technologies to identify, characterize and annotate enterprise data. This internal data can then be completed directly using external sources of linked information, such as trusted LOD cloud sources, via an open mash up of data, exploiting LOD2 linking mechanisms. To demonstrate the benefit of these linkage, we are developing a support application for the employee recruiting process, addressing one of our strategic markets of Human Resources IT software solutions. The application will be developed as an integrated service using CloudView platform and a dedicated workflow of LOD2 stack components.
Thomas Thurner: In December 2010 Exalead introduced its service Cloudview 360, basically a semantic mash up engine for enterprise data.
How does Cloudview utilize semantics and what role does the Semantic Web play in this product?
Amar-Djalil Meazour: Exalead is reshaping its portfolio offer to provide domain-specific “value”, necessitating rich semantic processing. Our CloudView platform includes a Semantic Factory, incorporating a wide variety of semantic processing modules which annotate and transform data in different ways, such as recognizing elements of a taxonomy, adding in classification results, and allowing fuzzy matching, or cross lingual access to data. An essential innovation of the Semantic Factory is its modularity, allowing processing modules to be arranged in any desired sequence, with outputs from one module fluidly serving as inputs to another. Exalead CloudView’s Semantic Factory visual configurability is paralleled by a rich search mashup platform that allows search results to be chained and displayed in interconnected windows. The Semantic Web formalisms provide a carrier for all the semantic resources, and exchange of data within the factory and within mashups.
Thomas Thurner: From an Exalead perspective: What are the next big trends in the search engine market, especially when it comes to enterprise search?
Amar-Djalil Meazour: The biggest trend is the replacement of search platforms by search-based applications in which search is a component over a larger application. This use of search as a component is possible since search input and output is now richly semantically structured, which allows for semantically pertinent operations on the output of search. It is in search-based applications that LOD2 results, linked data and the semantic web, will find their full power. Rather than the classic case of a use asking a question and then having to process a human readable search engine results page (10 links and snippets), search based applications can turn a user request into a rich output of interlinked pie charts, maps, graphs, and structured information.
Another trend is continued use of distributed outsourced infrastructures for hosting and managing IT applications of companies, in cloud architectures. Enterprise search solutions will adapt to this trend by providing less “on premises” solutions, deployed on site, and more search based applications delivered as Saas services. Only the effective consumed service will be charged reducing the global costs of a search based application. Beyond this trend, we will also see, as tools for extracting linked data from unstructured web sources progress, more Data as a Services solutions. On line reputation monitoring is one example of DaaS.
About Exalead
Founded in 2000 by search engine pioneers, Exalead is a global software provider in the enterprise and Web search markets. Exalead’s worldwide client base includes leading companies such as PSA Peugeot Citroën, ViaMichelin, GEFCO, American Greetings and Sanofi Aventis R&D, and more than 100 million unique users a month use Exalead’s technology for search.
Today, Exalead is reshaping the digital content landscape with a platform that uses advanced semantic technologies to bring structure, meaning and accessibility to previously unused or under-utilized data in the disparate, heterogeneous enterprise information cloud. The system collects data from virtually any source, in any format, and transforms it into structured, pervasive, contextualized building blocks of business information that can be directly searched and queried, or used as the foundation for a new breed of lean, innovative information access applications.
This revolutionary technology and others make Exalead one of the most important companies to watch in the fields of digital content search, discovery, management, security, and storage.
