The Future of Search - Martin White

Intranet Focus Ltd
The future of search
Findwise Findability Day
30 May 2013
Martin White
Managing Director, Intranet Focus Ltd.
martin.white@intranetfocus.com
@intranetfocus

Agenda
 A very short history of search and big data
 Information as a business asset
 Metadata tagging
 Finding people and expertise
 Mobile enterprise search
 Big data and enterprise search
 Content analytics
 Semantic search
 Sentiment analysis
 Digital workplaces
www.intranetfocus.com

In the beginning…!
Charles Bourne 1963
Gerard Salton 1965
John Mashey 1997

No sign of improvement!

Metadata tagging
 There is an increasing awareness of the value in tagging
content
 However this value is not enough to justify human tagging
even if consistency issues can be addressed
 There is a significant amount of research being undertaken
into passive (computer-generated) metadata
 ISO/IEC JTC1 SC32 WG2 is the Working Group that
develops international standards for metadata and related
technologies
 ISO 25964 is the international standard for thesauri and
interoperability with other vocabularies
 This is why organisations need information scientists!

Searching for expertise
 Searching for people and expertise is the single most
important task in any organisation
 The challenges are summed up by David Snowden
 Knowledge can only be volunteered, it cannot be conscripted.
 We only know what we know when we need to know it.
 In the context of real need few people will withhold their
knowledge.
 The way we know things is not the way we report we know things.
 We always know more than we can say, and we always say more
than we can write down.
 http://www.gurteen.com/gurteen/gurteen.nsf/id/km-seven-principles
 Relying on people creating expertise profiles is not the
answer

http://www.basistech.com/name-indexer/whitepaper.html

http://www.hpl.hp.com/techreports/2012/HPL-2012-102.pdf
http://www.research.ibm.com/haifa/dept/imt/ct_st_public.shtm l

Mobile enterprise search

http://designingthesearchexperience.com/

Mobile search
 Will require a radical re-appraisal of the search process
 For the first time users will want to
 Start a search on a desk top and transfer it to a mobile device
 Start a search on a mobile device and transfer it to a desktop
 Search as a team
 Print out results on a remote basis
 Use natural search interfaces such as voice and swipe
 Use GPS as a search parameter
 Virtually no research has yet been carried out on mobile
search outside of some stand-alone consumer services
 Collaborative search is an extension of mobile search
 Will stimulate voice technologies

http://research.microsoft.com/en-us/um/people/ryenw/papers/WangWSDM2013.pdf
http://dmrussell.net/CHI2010/docs/p951.pdf

Big Data and Enterprise Search
 The ratio between structured data and unstructured data
will vary between companies and sectors
 But both will be present and each has a dependency on
the other
http://info.marklogic.com/post-relational-reality-dbta-survey-2012.html

The best of both worlds
 The benefits come from being able to integrate
Big Data with Enterprise Search
http://www.aiim.org/pdfdocuments/IW_Big-Data_2012.pdf

Visual data analysis
http://queue.acm.org/detail.cfm?id=2146416

Search-based applications
 To make these work requires a lot of business analysis and
the combination of structured and unstructured
information
 A number of challenges have still not been fully addressed
 Given the high level of dissatisfaction with search will users trust
an SBA?
 How open will the SBA be to personal optimisation?
 How sensitive will the SBA be to changes in data and information
quality?
 Will it be possible to connect SBAs on different platforms?
 Solving these and other challenges will require a cross-
functional and cross-technology teams which are normally
absent in most organisations

Content analytics

http://www.pwc.com/us/en/press-releases/2012/pwc-releases-new-technology-forecast.jhtml

Sentiment analysis
 Sentiment analysis enables organisations to monitor a wide
range of information sources for positive and negative
signals about author opinions
 Analysis can be carried out at
 Document level
 Sentence level
 Aspect based
 Comparative level
 Convergence of search and text analytics
http://cacm.acm.org/magazines/2013/4/162501-techniques-and-applications-for-sentiment-analysis/fulltext

Semantic technologies
 Semantic technologies include
 Auto-categorization
 Concept analysis
 Concept and entity extraction
 Content data normalizing
 Federating and de-duplicating
 Natural language processing (NLP)
 Sentiment analysis
 Text mining and text analytics
 There is not a single ‘semantic’ technology and each
vendor and research group has a different view and
different priorities

Monitoring the enterprise
 During the course of the working day we need to keep a
track of
 Emails
 Blogs
 Microblogs
 Wikis
 Discussion forums
 Team sites
 Document additions and updates
 RSS feeds
 External information services
 We need more than just the delta changes, but also to
monitor
 Trends
 Emerging issues

Activity stream analysis
http://www.research.ibm.com/haifa/dept/imt/ct_st_public.shtm

Digital workplaces
 Although there is still no consensus on a definition for a
‘digital workplace’ there is considerable interest in trying to
understand the value of a digital workplace
 Search, and in particular search-based applications and
search-as-a-service, could have a core role to play as an
integration platform across multiple applications and
repositories, including those of suppliers and customers

Digital workplace framework
Jane McConnell http://netjmc.com

Digital workplaces
Search is likely to be the core platform technology

Questions and discussion
Martin.white@intranetfocus.com @intranetfocus

The Future of Search - Martin White

Recommandé

Recommandé

Contenu connexe

Similaire à The Future of Search - Martin White

Similaire à The Future of Search - Martin White (20)

Plus de Findwise

Plus de Findwise (20)

Dernier

Dernier (20)

The Future of Search - Martin White