Suresh Subramanian

Methodology for hyperspectral image classification using novel neural network

A novel feed forward neural network is used to classify hyperspectral data from the AVIRIS sensor. The network applies an alternating direction singular value decomposition technique to achieve rapid training times. Very few samples are required for training. 100 percent accurate classification is obtained using test data sets. The methodology combines this rapid training neural network together with data reduction and maximal feature separation techniques such as principal component analysis and simultaneous diagonalization of covariance matrices, for rapid and accurate classification of large hyperspectral images. The results are compared to those of standard statistical classifiers.

Subpixel object detection using hyperspectral imaging for search and rescue operations

Time critical search & rescue (s&r) operations often requires the detection of small objects in a vast area. While an airborne search can cover the area, no operational instrumental tools currently exist to actually replace the human operator. By producing the spectral signature of each pixel in a spatial image, multi- and hyper-spectral imaging (HSI) sensors provides a powerful capability for automated detection of subpixel size objects that are otherwise unresolved objects in conventional imagery. This property of HSI naturally lends itself to s&r operations. A lost hiker, skier, life raft adrift in the ocean, downed pilot or small aircraft wreckage targets, can be detected from relatively high altitude based on their unique spectral signatures. Moreover, the spectral information obtained allows the search craft to operate at substantially reduced spatial resolution thereby increasing scene coverage without a significant loss in detection sensitivity. The paper demonstrates the detection of objects as small as 1/10 of an image pixel from a sensor flying at over 6 km altitude. A subpixel object detection algorithm using HSI, based on local image statistics without reliance on spectral libraries is presented. The technique is amenable to fast signal processing and the requisite hardware can be built using inexpensive off the shelf technology. This makes HSI a highly attractive tool for real-time, autonomous instrument-based implementation. It can complement current visual-based s&r operations or emerging synthetic aperture radar sensors that are much more expensive.

Chemical detection using the airborne thermal infrared imaging spectrometer (TIRIS)

A methodology is described for an airborne, downlooking, longwave infrared imaging spectrometer based technique for the detection and tracking of plumes of toxic gases. Plumes can be observed in emission or absorption, depending on the thermal contrast between the vapor and the background terrain. While the sensor is currently undergoing laboratory calibration and characterization, a radiative exchange phenomenology model has been developed to predict sensor response and to facilitate the sensor design. An inverse problem model has also been developed to obtain plume parameters based on sensor measurements. These models, the sensors, and ongoing activities are described.

Efficient Anytime Anywhere Algorithms for Closeness Centrality in Large and Dynamic Graphs

Recent advances in social network analysis methodologies for large (millions of nodes and billions of edges) and dynamic (evolving at different rates) networks have focused on leveraging new high performance architectures, parallel/distributed tools and novel data structures. However, there has been less focus on designing scalable and efficient algorithms to handle the challenges of dynamism in large-scale networks. In our previous work, we presented an overarching anytime anywhere framework for designing parallel and distributed social network analysis algorithms that are scalable to large network sizes and can handle dynamism. A key contribution of our work is to leverage the anytime and anywhere properties of graph analysis problems to design algorithms that can efficiently handle network dynamism by reusing partial results, and by reducing re-computations. In this paper, we present an algorithm for closeness centrality analysis that can handle changes in the network in the form of edge deletions. Using both theoretical analysis and experimental evaluations, we examine the performance of our algorithm with different network sizes and dynamism rates.

Thermal Infrared Imaging Spectrometer (TIRIS) status report

The TIRIS is a pushbroom long wave infrared imaging spectrometer designed to operate in the 7.5 to 14.0 micrometer spectral region from an airborne platform, using uncooled optics. The focal plane array is a 64 by 20 extrinsic Si:As detector operating at 10 K, providing 64 spectral bands with 0.1 micrometer spectral resolution, and 20 spatial pixels with 3.6 milliradians spatial resolution. A custom linear variable filter mounted over the focal plane acts to suppress near field radiation from the uncooled external optics. This dual- use sensor is developed to demonstrate the detection of plumes of toxic gases and pollutants in a downlooking mode.

Effectively Handling New Relationship Formations in Closeness Centrality Analysis of Social Networks Using Anytime Anywhere Methodology

The flood of real time social data, generated by various social media applications and sensors, is enabling researchers to gain critical insights into important social modeling and analysis problems such as the evolution of social relationships and analysis of emergent social processes. However, current computational tools have to address the grand challenge of analyzing large and dynamic social networks within strict time constraints before the available social data can be effectively utilized. The computational issues are further exacerbated by the network size, which can range in the millions of nodes, and by the need for analytical tools to work with various computational architectures. Existing methodologies primarily deal with dynamic relationships in social networks by simply re-computing the results, and relying on massive parallel and distributed processing resources to maintain time constraints. In previous work, we introduced an overarching parallel/distributed algorithm design framework called the anytime anywhere framework, which leverages the inherent iterative property of graph algorithms to generate partial results, whose quality increase with the processing time, and which efficiently incorporates network changes. In this paper, we focus on closeness centrality algorithm design for dynamic social networks where new relationships are formed due to edge additions. Using both theoretical analysis and empirical results, we will demonstrate how this algorithm efficiently reuses the partial results and reduces the need for re-computations.

Incorporating Social Theories in Computational Behavioral Models

Computational social science methodologies are increasingly being viewed as critical for modeling complex individual and organizational behaviors in dynamic, real world scenarios. However, many challenges for identifying, representing and incorporating appropriate socio-cultural behaviors remain. Social theories provide rules, which have strong theoretic underpinnings and have been empirically validated, for representing and analyzing individual and group interactions. The key insight in this paper is that social theories can be embedded into computational models as functional mappings based on underlying factors, structures and interactions in social systems. We describe a generic framework, called a Culturally Infused Social Network (CISN), which makes such mappings realizable with its abilities to incorporate multi-domain socio-cultural factors, model at multiple scales, and represent dynamic information. We explore the incorporation of different social theories for added rigor to modeling and analysis by analyzing the fall of the Islamic Courts Union (ICU) regime in Somalia during the latter half of 2006. Specifically, we incorporate the concepts of homophily and frustration to examine the strength of the ICU’s alliances during its rise and fall. Additionally, we employ Affect Control Theory (ACT) to improve the resolution and detail of the model, and thus enhance the explanatory power of the CISN framework.

Modeling emergent border-crossing behaviors during pandemics

Modeling real-world scenarios is a challenge for traditional social science researchers, as it is often hard to capture the intricacies and dynamisms of real-world situations without making simplistic assumptions. This imposes severe limitations on the capabilities of such models and frameworks. Complex population dynamics during natural disasters such as pandemics is an area where computational social science can provide useful insights and explanations. In this paper, we employ a novel intent-driven modeling paradigm for such real-world scenarios by causally mapping beliefs, goals, and actions of individuals and groups to overall behavior using a probabilistic representation called Bayesian Knowledge Bases (BKBs). To validate our framework we examine emergent behavior occurring near a national border during pandemics, specifically the 2009 H1N1 pandemic in Mexico. The novelty of the work in this paper lies in representing the dynamism at multiple scales by including both coarse-grained (events at the national level) and finegrained (events at two separate border locations) information. This is especially useful for analysts in disaster management and first responder organizations who need to be able to understand both macro-level behavior and changes in the immediate vicinity, to help with planning, prevention, and mitigation. We demonstrate the capabilities of our framework in uncovering previously hidden connections and explanations by comparing independent models of the border locations with their fused model to identify emergent behaviors not found in either independent location models nor in a simple linear combination of those models.

Efficient Anytime Anywhere Algorithms for Vertex Additions in Large and Dynamic Graphs

Over the past decade, there has been a dramatic increase in the availability of large and dynamic social network datasets. Conducting social network analysis (SNA) on these networks is critical for understanding underlying social phenomena. However, continuously evolving graph structures require massive recomputations and conducting SNA is infeasible if the computations have to be restarted for every change. Many recent works have proposed large-scale graph processing systems and frameworks, but less attention has been given to scalable SNA algorithm designs that can efficiently adapt to dynamic graph changes. Moreover, continuously adapting to dynamic graph changes such as node/vertex/actor additions/deletions in a parallel/distributed computational environment can skew the initial graph partitions, leading to load imbalance issues and performance degradation. Previous approaches that focus on computing SNA measures on dynamic graphs either ignore this critical load-balancing aspect or focus only on measures that are straightforward and inherently adjustable to changes in the graph topology. In this work, we have designed an anytime anywhere closeness centrality algorithm that can efficiently incorporate vertex additions while avoiding massive recomputations, by leveraging a generic framework for designing parallel/distributed algorithms called anytime anywhere. Furthermore, we have also performed an analysis of the effectiveness of various processor assignment strategies to mitigate the load imbalances caused by dynamic graph changes.

Modeling insider threat types in cyber organizations

Insider threats can cause immense damage to organizations of different types, including government, corporate, and non-profit organizations. Being an insider, however, does not necessarily equate to being a threat. Effectively identifying valid threats, and assessing the type of threat an insider presents, remain difficult challenges. In this work, we propose a novel breakdown of eight insider threat types, identified by using three insider traits: predictability, susceptibility, and awareness. In addition to presenting this framework for insider threat types, we implement a computational model to demonstrate the viability of our framework with synthetic scenarios devised after reviewing real world insider threat case studies. The results yield useful insights into how further investigation might proceed to reveal how best to gauge predictability, susceptibility, and awareness, and precisely how they relate to the eight insider types.