As enterprise enters further into the digital age, data has become the strategic asset that knowledge workers, small or large, rely on to guide their decisions. However, managing such large volumes of data has exposed some unprecedented challenges for the enterprises. Enterprises have learned that the data that they hold, comes in a variety of formats, resides in different and distributed systems and is specific to the organization and its domain.
Setting these challenges as the backdrop, IBMâ€™s Watson division has built solutions that not only allow for data connectivity but also the analysis of unstructured data and its customization to an enterprise domain.
IBM Watson Explorer is Watsonâ€™s flagship product for text analytics and discovery. Watson Explorer allows enterprises to access multiple data sources from a single unified platform. The productâ€™s content analytics capabilities allow knowledge workers to take a step further and analyze the unstructured, textual, content. Furthermore, to tailor the analytics to an enterprise domain, Watson Explorer offers the ability to create rules that allow the curation of the analytics to the userâ€™s data. For example, an enterprise, letâ€™s say a bank, can create rules for terms that are specific to its business. On the other hand, a medical device company can create an entirely different set of rules that are more specific to its own industry.
A rule based technique offers high precision, and is easier to start with, but can
be difficult to scale. Understandably, creating rules to encompass the linguistics nuances for a very large dataset, such as that of an entire industry, can be quite a laborious and time intensive effort. Furthermore, this effort typically requires the assistance of natural language computer scientists. The recently launched offering by IBM, Watson Knowledge Studio, plays an integral role in taking on these exact challenges.
Watson Knowledge Studio employs machine learning and natural language capabilities to create language models that are specific to an industry or enterprise.
The machine learning capabilities allow users to teach the system with only a subset of the documents. A model trained in this fashion can then expand its knowledge to data sets that are much larger. Furthermore, to truly allow the industry experts to customize the analysis, all the model training is done through a web-interface that does not require the user to write a single line of code. After the subject matter experts build their machine learning models, they can download and import them into Watson Explorer. The machine learning models run in tandem with the rules written to add to Watson Explorerâ€™s search functionality and unstructured content analytics capabilities.
For more information on how Watson Knowledge Studio works, check out the video demo.
Best of both worlds
Watson Explorer can uniquely utilize both rule based and machine learning based custom language models. This allows the platform to reap the benefits of both. By using rule based capabilities users can quickly customize, with high precision, the analytics to their domain. Whereas, with Watson Knowledge Studioâ€™s machine learning capabilities users can scale to larger data sets faster, easily manage complex models and efficiently deal with ambiguity in the data (untrained cases).
Incorporating both these methodologies within one platform allows for faster, more accurate and customized analytics. With both products working together as a knowledge platform, users can expect to reduce time-to value, enable team collaboration and further increase the ROI of Watson Explorer by reducing training efforts.
For more information on the integration of Watson Explorer and Watson Knowledge Studio register for the webinar