Adapt HyLighter for a wide variety of uses and desired outcomes.
In the age of the internet, many of us carry out projects or assignments that require combining pieces of information from multiple online sources and the collective intelligence of groups to accomplish something of value. Such activities are fundamental to the original purpose of the internet and the essence of our information economy. However, until HyLighter came along, support for multi-document sensemaking (e.g., write a paper, proposal, or report, answer a complex question, solve a difficult problem, make an important decision, develop an innovative idea) was largely missing from the IT landscape.
Even with the digital workplace transformation well underway and a workforce that is increasingly open to emerging technologies such as Cloud computing, AI, and various communication technologies (e.g., chat, voice, video, among others), this gap has remained as a major pain point. Word-processing applications like Word and GoogleDocs are for authoring, editing, and collaborative review of individual documents. Online annotation systems like Scrible, hypothes.is, and Lacuna Stories, among hundreds of open source and commercial annotation systems, are largely for adding meta-content (i.e., content about content) to individual documents and Web pages and sharing with others.
HyLighter is a flexible platform for combining high-value pieces of information from multiple online sources to produce something of value. HyLighter provides this capability by assimilating machine analytics and other machine components into collaborative sensemaking networks. This approach redefines the human/machine division of labor to (a) make machine-scale analytics more compatible with human-scale sensemaking practices and (b) transform data/information into usable knowledge.
Executive decision makers often rely on teams of analysts or researchers to gather and summarize information related to making a critical decision. This can put the executive in the disquieting position of having to rely more on intuition than a grasp of the relevant evidence. With HyLighter in the picture, analysts create a Knowledge Trail and write a report based on its content. A Knowledge Trail is a set of high-value HyLights (i.e.,color-coded fragments of text linked to related comments and other content about the fragments extracted from multiple documents) and arranged in a meaningful order with high-speed links back to their exact locations within their sources.
Analysts provide the executive decision maker(s) with a report and related Knowledge Trail. The executive can (a) rapidly navigate the Knowledge Trail to efficiently acquire familiarity with key content in a large document collection and (b) instantly link back from pieces of information in the report of the analysts to relevant content in the sources. Through this process, the executive is better prepared to question the analysts and make a decision based on knowledge of available evidence. Among other desirable effects, HyLighter maintains an historical record of conversations tied to important sections of documents for auditing past decisions and informing future actions.
The HyLighter tacit knowledge management (TKM) system helps organizations capture, share, and retain knowledge gained from on-the-job experience by superior performers. Use cases for TKM include navigating banking, financial, or insurance regulations; following legal, medical, safety, maintenance, or security procedures; or carrying out manufacturing processes. In a typical scenario, a high-performing individual or team creates a Knowledge Trail for a specific task by following the four steps listed below. A Knowledge Trail is a set of high-value HyLights (i.e., color-coded fragments of text linked to related comments and other information about each fragment) extracted from multiple documents and arranged in a meaningful order with high-speed links back to their exact locations within their sources:
The growth of the Internet has made large numbers of sources and huge volumes of information available for gathering competitive intelligence. HyLighter is your competitive intelligence (CI) assistant for anticipating changes in market conditions. CI is focused on acquiring knowledge about relevant developments in your industry (and related industries) so that your company can make timely adjustments to changing market circumstances. CI is not only concerned with analyzing the competition but, also, staying informed about new and emerging developments in your industry, consumer behavior, and broader trends. CI may include gathering and analyzing information from diverse sources (e.g., market analysis reports, government records, and news outlets) and may focus on a wide array of topics (e.g., examining the implications of potentially disruptive technologies, pending or potential changes in laws and regulations, changes in customer tastes and preferences). HyLighter assists CI analysts in combining pieces of information from multiple online sources to acquire strategic awareness in a rapidly changing world.
A major challenge of the digital information age is how to tap into large volumes of online information and the collective intelligence of diverse groups to generate new knowledge, solve difficult problems, and drive innovation. HyLighter enables new forms of interactions between humans, machines, and online content that have the potential to (a) improve outcomes of sensemaking activities that involve large collections of online documents and diverse groups, such as inter- and transdisciplinary literature reviews and (b) make machines more capable of assisting humans in such sensemaking efforts. HyLighter helps non-technical users (and technical users, too) to find, organize, and synthesize high-value information from multiple sources. As a consequence of interacting with the system, users become more capable of thinking within a domain or problem space and machines become more capable of assisting users in their sensemaking efforts.
You are probably familiar with the Pareto principle or 80/20 rule. This principle provides a rough guide about typical distributions based on the observation that things like effort and output are not distributed evenly. For example, if you run a document review with a team of 10 people, you might find that 20% of the workers produce 80% of the results.
At HyLighter, we use our technology on a regular basis for a variety of purposes. Recently, the CEO posted an important document in HyLighter and invited all six members of the business team to add their comments. The CEO added 20 comments and one other member of the business team added 25 comments. The other four members, for a variety of reasons, did not participate in the discussion.
Using HyLighter, the CEO was able to get input from everyone on the team in an efficient and productive manner. He opened the document in HyLighter and selected a dozen high-value HyLights to include in a Knowledge Trail. At the next business team meeting, all six members participated in a lively conversation focused on the dozen points raised by the Knowledge Trail. Through this method, the CEO was able to get input from everyone on the team.
Former Secretary of Defense, Donald Rumsfeld, speaking to the press in 2002, memorably used the phrase, “unknown unknowns,” to refer to a problem that keeps analysts and decision makers up at night:
There are known knowns; there are things we know that we know. There are known unknowns; that is to say there are things that we now know we don't know. But there are also unknown unknowns – there are things we do not know we don't know.
In other words, a known unknown is a question that an analyst knows to ask and pursue further. An unknown unknown is a question that an analyst does not even know to ask. In the age of Big Data, detecting the unknown unknowns represents both a challenge and opportunity facing knowledge workers across all sectors.
HyLighter establishes a feedback loop between machines and analysts that supports a systematic approach to the detection of unknown unknowns hidden in Big Data. As an operational definition, unknown unknowns are entities that are not in a subset of files exported from a larger collection but are related at some criterion-level of strength to entities that are in the subset.
In brief, a text analytics platform identifies all targeted entities in one or more large document repositories (e.g., Big Data) and measures various attributes (e.g., sentiment) and associations (e.g., the strength of relationships between each entity in the corpus and every other entity). When analysts import a promising subset of files (i.e., a collection) to HyLighter, the system converts machine analytics into Machine HyLights -- as if the machine was a specially-abled teammate with the ability to emphasize all targeted entities across the collection in gray highlighting and add related meta-content (e.g., category, attributes, and associations) to the margin. Guided by the Machine HyLights, analysts add a layer of human-generated HyLights (i.e., color-coded fragments of text linked to related comments and other information about each fragment) across the collection.
Once the HyLight layer is sufficiently developed, analysts create Knowledge Trails (i.e., a set of high-value HyLights extracted from multiple documents and arranged in a meaningful order with high-speed links back to their exact locations within their sources) and feed their Knowledge Trails back to the text analytics platform. The machine finds all entities in the large repositories that are related to entities in the Knowledge Trail at some pre-determined level of strength but are not in the Knowledge Trail. The data that is returned to the analysts includes a list of entities in the large repositories that have criterion-level relationships with entities in the collection (i.e., previously unknown unknowns as operationally defined above). The analysts import a new subset of files identified by the text analytics application to HyLighter that include new entities detected by the machine for evaluation of relevance to the current task or mission.
HyLighter can help to prepare students for the 21st century information economy and an uncertain future. As rapid technological change eliminates jobs in every sector and new types of work are emerging that require new skill sets, the curricula in schools remain largely out of sync with global economic reality. As an historical analogy, literacy in Western Europe until the fifteenth century was largely limited to clergy, scholars, and the wealthy. With the arrival of Gutenberg’s innovation and the availability of low-cost reading material, literacy spread widely, especially when newspapers appeared in the 1600s. By analogy, the Internet, smart machines, and related technologies are the movable type and printing press of today, and data scientists are the literate elite. As Lanier has argued in Who Owns the Future (2013), just as the spread of literacy helped to fuel the growth of a merchant class and usher in the Renaissance period, the spread of Social Machines (i.e., a promising model for unifying machines and social processes for a wide range of purposes) will help fuel a resurgence of the diminished middle-class and, potentially, contribute to a Renaissance-like transformation of society. Arguably, HyLighter is at the forefront of Social Machines that are configured to accelerate learning and support machine-assisted sensemaking for tapping into large volumes of online information and the collective intelligence of diverse groups.
Anti-plagiarism software is among the most widely used category of software in education. Although plagiarism has become less common as a result of effective monitoring systems, a form of paraphrasing called patch writing is common among college students and others. Rather than copying a text word for word, the writer rearranges phrases and replaces words, often using a thesaurus.
When HyLighter is in use, the instructor requires the student to embed HyLight IDs within the paper and upload the paper to HyLighter. The system automatically creates a mashup of the embedded HyLights. Through this mechanism, the instructor has the option to navigate from a location in the paper to HyLights within the sources that provide background or evidence for the student’s statements and authentic work.