Kedar Sambhoos

Enhancements to high level data fusion using graph matching and state space search

The intent of this paper is to show enhancements in Levels 2 and 3 fusion capabilities through a new class of models and algorithms in graph matching. The problem today is not often lack of data, but instead, lack of information and data overload. Graph matching algorithms help us solve this problem by identifying meaningful patterns in voluminous amounts of data to provide information. In this paper we investigate a classical graph matching technique for subgraph isomorphism. A complete implementation of a heuristic approach (since the problem under consideration is NP-Hard) using an inexact isomorphism technique has been used. The heuristic approach is called Truncated Search Tree algorithm (TruST), where the state space of the problem is constrained using breadth and depth control parameters. The breadth and depth control parameters are then studied using design of experiment based inferential statistics. Finally, a software implementation of the procedure has been completed.

A fuzzy graph matching approach in intelligence analysis and maintenance of continuous situational awareness

In intelligence analysis a situation of interest is commonly obscured by the more voluminous amount of unimportant data. This data can be broadly divided into two categories, hard or physical sensor data and soft or human observed data. Soft intelligence data is collected by humans through human interaction, or human intelligence (HUMINT). The value and difficulty in manual processing of these observations due to the volume of available data and cognitive limitations of intelligence analysts necessitate an information fusion approach toward their understanding. The data representation utilized in this work is an attributed graphical format. The uncertainties, size and complexity of the connections within this graph make accurate assessments difficult for the intelligence analyst. While this graphical form is easier to consider for an intelligence analyst than disconnected multi-source human and sensor reports, manual traversal for the purpose of obtaining situation awareness and accurately answering priority information requests (PIRs) is still infeasible. To overcome this difficulty an automated stochastic graph matching approach is developed. This approach consists of three main processes: uncertainty alignment, graph matching result initialization and graph matching result maintenance. Uncertainty alignment associates with raw incoming observations a bias adjusted uncertainty representation representing the true value containing spread of the observation. The graph matching initialization step provides template graph to data graph matches for a newly initialized situation of interest (template graph). Finally, the graph matching result maintenance algorithm continuously updates graph matching results as incoming observations augment the cumulative data graph. Throughout these processes the uncertainties present in the original observations and the template to data graph matches are preserved, ultimately providing an indication of the uncertainties present in the current situation assessment. In addition to providing the technical details of this approach, this paper also provides an extensive numerical testing section which indicates a significant performance improvement of the proposed algorithm over a leading commercial solver.

Discovering patterns in social networks with graph matching algorithms

Social media data are amenable to representation by directed graphs. A node represents an entity in the social network such as a person, organization, location, or event. A link between two nodes represents a relationship such as communication, participation, or financial support. When stored in a database, these graphs can be searched and analyzed for occurrences of various subgraph patterns of nodes and links. This paper describes an interactive visual interface for constructing subgraph patterns called the Graph Matching Toolkit (GMT). GMT searches for subgraph patterns using the Truncated Search Tree (TruST) graph matching algorithm. GMT enables an analyst to draw a subgraph pattern and assign labels to nodes and links using a mouse and drop-down menus. GMT then executes the TruST algorithm to find subgraph pattern occurrences within the directed graph. Preliminary results using GMT to analyze a simulated collection of text communications containing a terrorist plot are reported.

Performance evaluation methods for data-fusion capable tactical platforms

Relatively little research has been done so far to define a technically fair yet affordable method for performance evaluation (PE) for data fusion (DF) processes. PE methodology is required to evaluate performance of alternative DF process designs that are being developed to handle the increasing complexity of modern day applications. This paper addresses the distributed fusion problem and gives quantitative insights into the inter dependencies and the consistency measures between distributed fusion measures of performance. Building on our prior works, our recommended PE methodology based upon the dual node network (DNN) DF & resource management (DF&RM) architecture is employed herein in a case study involving track picture consistency across multiple platforms. This research focuses in particular on distributed level 1 DF PE applications for the Air Force Flight Test Center (AFFTC), in support of new test and evaluation procedures that will be required for advanced, fusion-capable tactical aircraft.

Data association and soft data streams

This paper discusses the challenges of and possible methods for data association in the domain of counterinsurgency where “soft/linguistic” data is an important input data type. An overview of the processing operations from input to construction of fused estimates is described. The design issues that are discussed and require further exploration to yield a workable and efficient association process include developing an input batching logic, finding efficient ways to search between graphs, and the selection of appropriate semantic similarity metrics to associate nodes and arcs. Additionally, the solution to a multi-dimensional assignment problem and graph merging techniques will need to be defined. The application of data association in this type of environment has potential to yield an improved, comprehensive data graph which will aid in reducing search time and provide more accurate results for analysts making real time decisions in the real world.

Designing a Performance Evaluation Methodology for Data Fusion-capable Tactical Aircraft V

[Abstract] In previous papers, we documented our evolving research that expanded on and formalized an approach to the design of a Performance Evaluation (PE) methodology for Data Fusion (DF)-based tactical aircraft systems. We have shown that the design of a PE process for any multi-sensor or multi-aircraft fusion-based system involves the design of a separate data fusion process involving association and estimation functions for PE purposes per se. Our publications to date have developed the theoretical and architectural groundings for this new PE process, and several case studies have been carried out to show sample implementations of the principles of this new methodology. In addition, some limitedobjective parametric experiments have also been carried out that show the application of the new evaluation methodology for typical tactical aircraft problems. In the current paper, we summarize and cumulate the findings of these past works, and show our most recent AFOSR/AFFTC-sponsored research efforts related to extending the design and application of this methodology to air-to-air engagement problems involving what are called higherlevels of data fusion capability (situation and threat estimation) and the employment of electronic warfare systems. The paper discusses the detailed strategies for data association, metrics estimation, and also the analytical techniques that exploit the formality of the methods of Statistical Design of Experiments (DOE) and Analysis of Variance (ANOVA) for these fusion applications.

Searching social networks for subgraph patterns

Software tools for Social Network Analysis (SNA) are being developed which support various types of analysis of social networks extracted from social media websites (e.g., Twitter). Once extracted and stored in a database such social networks are amenable to analysis by SNA software. This data analysis often involves searching for occurrences of various subgraph patterns (i.e., graphical representations of entities and relationships). The authors have developed the Graph Matching Toolkit (GMT) which provides an intuitive Graphical User Interface (GUI) for a heuristic graph matching algorithm called the Truncated Search Tree (TruST) algorithm. GMT is a visual interface for graph matching algorithms processing large social networks. GMT enables an analyst to draw a subgraph pattern by using a mouse to select categories and labels for nodes and links from drop-down menus. GMT then executes the TruST algorithm to find the top five occurrences of the subgraph pattern within the social network stored in the database. GMT was tested using a simulated counter-insurgency dataset consisting of cellular phone communications within a populated area of operations in Iraq. The results indicated GMT (when executing the TruST graph matching algorithm) is a time-efficient approach to searching large social networks. GMT’s visual interface to a graph matching algorithm enables intelligence analysts to quickly analyze and summarize the large amounts of data necessary to produce actionable intelligence.