VELOXITI’S KNOWLEDGE GRAPH TOOLS
Click through the tabs below to view Veloxiti’s knowledge graph tool’s screenshot examples and explanations. If you have any questions about our tools and how we implement them to benefit our clients please contact us!
“Magnify relationships to make the haystack disappear.”
Representing your world as a structured model is a design activity for business teams focused on productivity, competitive advantage, innovation or sharing expertise. Developing models is a key cost driver for stronger artificial intelligence. The tools your development team chooses will determine a return on investment but thanks to decades of successful projects, we have innovations and talent to share.
vExtract™ helps your team build operational models from text sources faster and easier. Businesses utilize our knowledge-extraction tool to identify and validate domain specific expertise organized in a graph that can be viewed in edge-based radial or hierarchical layout. Text-based sources contain insights about such things as relationships between people, connections between journal articles, training concepts in manuals, or even shared ideas within chat feeds. Creating assets from text sources to enhance enterprise knowledge management efforts is a valuable activity, it is also a primary consideration within our cognitive solutions development kit.
We help enterprise teams extract and organize text-based information so that algorithms can utilize the same information. Stronger artificial intelligence is synonymous with knowledge capture and knowledge graph design.
vExtract™ is the first graph tool in our cognitive solutions development kit that helps capture and connect ontological concepts. Your decades of expertise sourced from text combined with our decades of engineering talent will magnify relationships. Although with minimal training it is possible your team can extract high level concepts and discover semantic relationships on their own. The resulting graph is the foundation for a variety of artificial intelligence techniques.
The dynamic nature of our graph platform allows for additional development considerations. We are a first step for integration and continuous improvement where structured expertise is a strategic asset – taking your team beyond data.
“Describe your solution by engineering expertise into structured graphs”
Inspired by scientifically recognized cognitive (neuroscience) techniques, our tools focus on the development of domain specific graphs optimized for an algorithm. Each graph part, node or connector, holds deep descriptions and variety of logic operators which a cognitive algorithm processes. After decades of evaluation and testing, our knowledge graph engineering tools have emerged as a critical part of our cognitive solutions development kit stack. The process to engineer knowledge graphs combines philosophy, cognitive psychology, and computer science.
vGraph™ is the core visual knowledge graph editor in our cognitive solutions development kit stack. We believe that stronger artificial intelligence requires a human touch moving extracted insights into the core graphing tool or manually building models to create executable software. Engineers then design structured representations of the problem and solution in two formal graph templates. vGraph™ model types are acyclic, directed graphs based on established Belief, Desire, and Intent (BDI) architectures. The final result is executable software ready for testing.
Our graph types are linked which provides a broader and more stable representation of the world, thereby increasing capability and reliability in critical domains like managing complex machines or identifying human behaviors.
vGraph™ cognitive solutions development tool integrates with many neuroscience-based software techniques like Associative Memory, Case-Based Reasoning (CBR), or Natural Language Processing (NLP). All of these approaches are driven by underlying knowledge representation strategies and each approach has strengths and weaknesses depending upon the problem to be solved. Our knowledge representation technique is based on earlier scientific work in Artificial Intelligence which focused on the development of efficient graph formats and types as well as algorithms that traverse graphs.
vGraph™ templates represent the State-of-the-World in two formal graph types, one called Observe-Orient Graph the other Decide-Act Graph. Observe-Orient Graph holds representations of beliefs, and the Decide-Act Graph holds representations of Goals (i.e., desires) and Plans (i.e., intentions). Though the graphs differ with respect to their detailed implementation, both the OO and DA Graphs are acyclic, directed graphs, and work as a unified system.
We not only have extremely powerful cognitive engines, but we also have a dynamic graph strategy. These two factors combined with three decades of cognitive solutions innovation give us the confidence to make amazing outcomes possible.
vGraph™ connects a variety of techniques to meet complexity, budget, or requirements and can integrate with other artificial intelligence or learning algorithms. We also teach others how to do what we do – with a little help from scientifically proven cognitive tools.
“What gets measured, gets improved.”
Today’s businesses utilize numerous data points to better understand both internal and external team performance data. New insights are required to enhance digital productivity but also to monitor human responses for critical deficiencies.
vHumanMetrics software was originally designed for battlefield leadership so that they could monitor the productivity of analysts at a workstation. Leaders began to discover real-time human-computer interaction deficiencies – and now this successful technology is ready for enterprise applications.
vHumanMetrics tracks productivity from each user class or across the world, it identifies and catalogs novel metrics, automatically tracking behaviors, all while integrating multiple data sources to build actionable real-time performance intelligence. This human-centered tool is unique in the degree to which it fuses subject matter expertise and powerful technology.
Generating user level productivity intelligence is extremely important for enterprise management teams. Measuring and improving digital information management has the potential to greatly influence the success real-time operations such as medical billing, supply chain, network administration, display controls or any interaction in a digital domain. Effective enterprise managers can now measure reactions to unexpected and unanticipated changes across their business while remaining focused on resolving problems. Empower your employees or users to identify problems and integrate shared tasks to accomplish team goals.
vHumanMetrics is a multi-faceted tool that can be customized to understand the effectiveness of human interaction within any digital domain. It is not a tool to rate employees or displace human workforces, it is simply a software measurement tool used to conduct team performance diagnostics. This new tool provides data about training inefficiencies, interface design issues, cognitive limitations under stress, or building an overall human workforce health score.
Demonstrating the value of a knowledge graph and its engine is no trivial task especially when the domain is large. To facilitate system evaluation, we developed a real-time knowledge graph analysis tools. Customer’s use our graph metrics to evaluate both individual and team performance while interacting with large complex graphs. Essentially this provides a window into the working components of our technology.
Our engine traverses the graph model and records the systems beliefs about the world, human-computer goals to be achieved, and required plans to be executed. The essential idea behind the metrics tool is to capture real-time instances of graph pathways to document what is happening in the world and what its human users intend to do about what they know. By measuring what is happening in the graphs, we can make inferences about employee and team performance.
The underlying activation of the knowledge graph structures can provide a dynamic real-time model of ‘human intention’ that has the potential to support the analytical community with novel classes of metrics. Some of these novel metrics might be:
- Cognitive workload (number of concurrently active goals across time)
- Current active plans and goals in the DA Graph
- Time and predicted time to complete ongoing tasks
- Predict collaborations between users based on shared plans & goals
- Force synchronization (timeliness of distributed sub-goal satisfaction)
Curious about how knowledge graph based tools can be put to use? Check out some examples below of successful projects we’ve done. Still have questions on how this technology could be used for your purposes? Contact us and tell us about your project needs.
Prescient® / VehicleDNA®
Note: VehicleDNA® is the commercial version of Prescient® – which has been used by the U.S. Department of Defense.
Prescient® / VehicleDNA® makes it possible to sample a vehicle in less than 10 minutes.
Oil evaluation is a two-step process. The first step is sample testing to identify wear metals and contaminants such as iron, copper, or soot and to determine sample viscosity. The second step is sample evaluation, a task that focuses on interpreting the meaning of test data. Evaluation provides the expert insight needed for determining a component’s condition.
For many years, oil testing equipment has been too expensive for use in maintenance depots or smaller service areas. However, reasonably priced test equipment exists today that is simple to use and robust enough for severe environments.
Of course, test data have no value until they are interpreted. Prescient® / VehicleDNA® addresses the all-important need for accurate sample evaluation. It uses both statistical algorithms and an expert rule-base that captures the knowledge of a world class oil evaluator. In addition, it not only diagnoses and recommends corrective actions but also explains the reasons for its assessment.
On-site analysis eliminates frustrating, time-consuming, and expensive delays. VehicleDNA® makes it possible to evaluation a truck’s condition while it is in the depot, not when it is on the road. On-site analysis dramatically reduces lost revenue due to maintenance recalls.
Furthermore, since Prescient®/VehicleDNA® software lives “in the cloud”, a vehicle’s condition history is always available at any depot and any time.