Intelligent Information Management

The Benefits of AI in Information Management and vice-versa

Information management in general can benefit greatly from implementations of artificial intelligence:

1. improvements in content recognition – identifying, extracting and deriving more informational value from content which in turn can even help improve the methods of AI used in content recognition;

2. possibility of automatic or machine assisted content manipulation – after decades of content management based solely on human intervention several processing patterns have been identified that could support automatic operations on the content if the elements of these patterns could have been identified correctly in the process of intelligent content recognition;

3. improvements in content search – intelligent content recognition constitutes a foundation for new methods of intelligent search that can yield more accurate results from existing (and new) content by identifying new (additional) associations between individual units of content;

4. improvements in content distribution (and presentation) – intelligent recognition of content will enable information providers to supply their users with content that better suits their needs and will help the providers to better adapt to the users' expectations. Identifying the needs of the end-user is one of the most important tasks any information provider (manager) faces when supplying information to the end-user. The implementation of artificial intelligence in the field of user profiling could just as well revolutionize content management altogether!

At the same time artificial intelligence can benefit greatly from information management:

1. Large number of best practice solutions – over the years information providers have created libraries of best practice solutions both from the area of general content management as well as from several areas of specific information management cases. Most of these practices can be translated into machine logic thus setting a base for rapid development of content management related AI solutions;

2. Large amounts of acquired content – the practices in content management mentioned above originate in an even larger collection of actual examples of content, of real creation methods (writing techniques), of real-life consumption needs, and of actual data manipulation possibilities. They provide both a strong basis for research into possibilities of AI implementations and a very large development and testing environment where new AI methods can be developed efficiently and quite rapidly;

3. Large environments with real-life situations – as mentioned information management systems deal with creation and/or acquisition of information, but at the same time they need to deal with consumption of information as well. Over the years serious investments of time and resources have been made in the area of analyzing the actual needs and expectations of end-users. The results of these analyses are used practically every day in contemporary information management systems to accommodate information consumption by best matching the needs and expectations of end-users. Most of these methods support translation into machine logic which can both improve user experience directly and aid in AI development by supplying new principles of learning;

4. Long history of information processing – judging by the long history of information management it is also highly probable that most of the cases of either information supply or demand have already been identified and responded to appropriately. The typical cases, at least, have so far been encountered and catered to on a day-to-day basis for many years, but many marginal and exceptional cases have also been identified and appropriate remedies have been developed for them. All these experience can help in development of AI solutions – both in information management and in the general field of AI.

No implementation of artificial intelligence in information management should even be attempted without a platform that could support the basic principles of content related development:

1. Intelligent content management even on the pure elementary (core) level – an appropriate solution foundation must include efficient methods of: data acquisition, information recognition, classification and association. These methods must be a natural part of the platform and they must at the same time be replaceable, upgradeable and scalable;

2. The ability to implement new methods of information management as soon as they evolve – an appropriate solution foundation must support instant implementation of any new methods of data manipulation – be it methods that require human intervention or automated methods. As soon as new methods reach the conceptual level the platform must support instant concept proofing, and as soon as new concepts are proven the platform must support instant implementation of results. 21^st century development must be swift and efficient;

3. Process and workflow management support – since most information management is done in business environments where many human and machine resources are actively involved in achieving results, the appropriate solution foundation must support such live and dynamic environments. Full support must be given to each individual person (or machine) to give them uninhibited access to information while at the same time all workflow-related scenarios must have full support as well. A team of individuals (humans and machines alike) achieves maximum efficiency through the ability to do their individual tasks without any unnecessary inhibitions and through synergetic effects of complementary (mutually contributive) tasks on the team level;

4. User profiling functionalities – an appropriate solution foundation must be user-aware – user management must be fully supported – i.e. complete user history must be available from which individual user preferences can be recognized, the foundation must support explicit user requests for information in order to provide the user with expected answers as soon as possible, also implicit activity tracking must be supported in order to provide expected results to users without burdening them further by asking them to specify their needs more accurately;

5. Business intelligence support – although business intelligence has little or nothing in common with artificial intelligence its results can aid in development of information related AI solutions. Business intelligence in information management systems revolves around two key issues: the analysis of existing information, their value and use on one hand, and the forecast of anticipated information, their anticipated value and use on the other. The essential role of business intelligence in the development of AI can be in defining research and development priorities to best complement actual information management needs – this way the results of any AI development can be put to real production tests synchronously with the development of information.

One solution foundation has been in development for the last year, and is now ready for real implementations – either as a foundation of information management production environments or as a foundation of new development environments. It conforms to all the principles mentioned above.

So far the Anthology Information Management Foundation (IMF) and the Anthology Intelligent Content Environment (ICE) have been built with just such complex implementations in mind, so any development work can theoretically begin as soon as the basic project guidelines are set. It would, however, be more appropriate and efficient that project plans be fully determined before any actual work is done. Usable results can be achieved steadily throughout the development phases, but some fundamental priorities will have to be set in order to enable achievements of the highest quality.

More information on Anthology (in Slovenian language) is available directly from the authors.