Balaraman Ravindran

Latent dirichlet allocation based multi-document summarization

Extraction based Multi-Document Summarization Algorithms consist of choosing sentences from the documents using some weighting mechanism and combining them into a summary. In this article we use Latent Dirichlet Allocation to capture the events being covered by the documents and form the summary with sentences representing these different events. Our approach is distinguished from existing approaches in that we use mixture models to capture the topics and pick up the sentences without paying attention to the details of grammar and structure of the documents. Finally we present the evaluation of the algorithms on the DUC 2002 Corpus multi-document summarization tasks using the ROUGE evaluator to evaluate the summaries. Compared to DUC 2002 winners, our algorithms gave significantly better ROUGE-1 recall measures.

Efficient computation of the shapley value for game-theoretic network centrality

The Shapley value--probably the most important normative payoff division scheme in coalitional games--has recently been advocated as a useful measure of centrality in networks. However, although this approach has a variety of real-world applications (including social and organisational networks, biological networks and communication networks), its computational properties have not been widely studied. To date, the only practicable approach to compute Shapley value-based centrality has been via Monte Carlo simulations which are computationally expensive and not guaranteed to give an exact answer. Against this background, this paper presents the first study of the computational aspects of the Shapley value for network centralities. Specifically, we develop exact analytical formulae for Shapley value-based centrality in both weighted and unweighted networks and develop efficient (polynomial time) and exact algorithms based on them. We empirically evaluate these algorithms on two real-life examples (an infrastructure network representing the topology of the Western States Power Grid and a collaboration network from the field of astrophysics) and demonstrate that they deliver significant speedups over the Monte Carlo approach. For instance, in the case of unweighted networks our algorithms are able to return the exact solution about 1600 times faster than the Monte Carlo approximation, even if we allow for a generous 10% error margin for the latter method.

Model Minimization in Hierarchical Reinforcement Learning

When applied to real world problems Markov Decision Processes (MDPs) often exhibit considerable implicit redundancy, especially when there are symmetries in the problem. In this article we present an MDP minimization framework based on homomorphisms. The framework exploits redundancy and symmetry to derive smaller equivalent models of the problem. We then apply our minimization ideas to the options framework to derive relativized options--options defined without an absolute frame of reference. We demonstrate their utility empirically even in cases where the minimization criteria are not met exactly.

Accurate mobile robot localization in indoor environments using bluetooth

In this paper, we describe an accurate method for localization of a mobile robot using bluetooth. We introduce novel approaches for obtaining distance estimates and trilateration that overcome the hitherto known limitations of using bluetooth for localization. Our approach is reliable and has the potential of being scaled to multi-agent scenarios. The proposed approach was tested on a mobile robot, and we present the experimental results. The error obtained was 0.427 ± 0.229 m, which proves the accuracy of our method.

Latent Dirichlet Allocation and Singular Value Decomposition Based Multi-document Summarization

Multi-Document Summarization deals with computing a summary for a set of related articles such that they give the user a general view about the events. One of the objectives is that the sentences should cover the different events in the documents with the information covered in as few sentences as possible. Latent Dirichlet Allocation can breakdown these documents into different topics or events. However to reduce the common information content the sentences of the summary need to be orthogonal to each other since orthogonal vectors have the lowest possible similarity and correlation between them. Singular Value Decompositions used to get the orthogonal representations of vectors and representing sentences as vectors, we can get the sentences that are orthogonal to each other in the LDA mixture model weighted term domain. Thus using LDA we find the different topics in the documents and using SVD we find the sentences that best represent these topics. Finally we present the evaluation of the algorithms on the DUC2002 Corpus multi-document summarization tasks using the ROUGE evaluator to evaluate the summaries. Compared to DUC 2002 winners, our algorithms gave significantly better ROUGE-1 recall measures.

Modeling basal ganglia for understanding parkinsonian reaching movements

We present a computational model that highlights the role of basal ganglia (BG) in generating simple reaching movements. The model is cast within the reinforcement learning (RL) framework with correspondence between RL components and neuroanatomy as follows: dopamine signal of substantia nigra pars compacta as the temporal difference error, striatum as the substrate for the critic, and the motor cortex as the actor. A key feature of this neurobiological interpretation is our hypothesis that the indirect pathway is the explorer. Chaotic activity, originating from the indirect pathway part of the model, drives the wandering, exploratory movements of the arm. Thus, the direct pathway subserves exploitation, while the indirect pathway subserves exploration. The motor cortex becomes more and more independent of the corrective influence of BG as training progresses. Reaching trajectories show diminishing variability with training. Reaching movements associated with Parkinsons disease (PD) are simulated by reducing dopamine and degrading the complexity of indirect pathway dynamics by switching it from chaotic to periodic behavior. Under the simulated PD conditions, the arm exhibits PD motor symptoms like tremor, bradykinesia and undershooting. The model echoes the notion that PD is a dynamical disease.

Improving legal information retrieval using an ontological framework

A variety of legal documents are increasingly being made available in electronic format. Automatic Information Search and Retrieval algorithms play a key role in enabling efficient access to such digitized documents. Although keyword-based search is the traditional method used for text retrieval, they perform poorly when literal term matching is done for query processing, due to synonymy and ambivalence of words. To overcome these drawbacks, an ontological framework to enhance the user’s query for retrieval of truly relevant legal judgments has been proposed in this paper. Ontologies ensure efficient retrieval by enabling inferences based on domain knowledge, which is gathered during the construction of the knowledge base. Empirical results demonstrate that ontology-based searches generate significantly better results than traditional search methods.

Adaptive network intrusion detection system using a hybrid approach

Any activity aimed at disrupting a service or making a resource unavailable or gaining unauthorized access can be termed as an intrusion. Examples include buffer overflow attacks, flooding attacks, system break-ins, etc. Intrusion detection systems (IDSs) play a key role in detecting such malicious activities and enable administrators in securing network systems. Two key criteria should be met by an IDS for it to be effective: (i) ability to detect unknown attack types, (ii) having very less miss classification rate. In this paper we describe an adaptive network intrusion detection system, that uses a two stage architecture. In the first stage a probabilistic classifier is used to detect potential anomalies in the traffic. In the second stage a HMM based traffic model is used to narrow down the potential attack IP addresses. Various design choices that were made to make this system practical and difficulties faced in integrating with existing models are also described. We show that this system achieves good performance empirically.

Efficient computation of the shapley value for centrality in networks

The Shapley Value is arguably the most important normative solution concept in coalitional games. One of its applications is in the domain of networks, where the Shapley Value is used to measure the relative importance of individual nodes. This measure, which is called node centrality, is of paramount significance in many real-world application domains including social and organisational networks, biological networks, communication networks and the internet. Whereas computational aspects of the Shapley Value have been analyzed in the context of conventional coalitional games, this paper presents the first such study of the Shapley Value for network centrality. Our results demonstrate that this particular application of the Shapley Value presents unique opportunities for efficiency gains, which we exploit to develop exact analytical formulas for Shapley Value based centrality computation in both weighted and unweighted networks. These formulas not only yield efficient (polynomial time) and error-free algorithms for computing node centralities, but their surprisingly simple closed form expressions also offer intuition into why certain nodes are relatively more important to a network.

Hierarchical activity recognition for dementia care using Markov Logic Network

Statistics reveal that globally, the aging population in different stages of dementia are struggling to cope with daily activities and are progressively becoming dependent on care takers thereby making dementia care a challenging social problem. Healthcare systems in smart environments that aim to address this growing social need require the support of technology to recognize and respond in an ubiquitous manner. To incorporate an efficient activity recognition and abnormality detection system in smart environments, the routine activities of the occupant are modeled and any deviation from the activity model is recognized as abnormality. Recognition systems are generally designed using machine learning strategies and in this paper a novel hybrid, data and knowledge-driven approach is introduced. Markov Logic Network (MLN) used in our design is a suitable approach for activity recognition as it integrates common sense knowledge with a probabilistic model that augments the recognition ability of the system. The proposed activity recognition system for dementia care uses a hierarchical approach to detect abnormality in occupant behavior using MLN. The abnormality in the context of dementia care is identified in terms of factors associated with the activity such as objects, location, time and duration. The task of recognition is done in a hierarchical manner based on the priority of the factor that is associated with each layer. The motivation for designing a hierarchical approach was to enable each layer to commence its computation after inferring from the lower layers that the ongoing activity was normal with regard to the associated factors. This hierarchical feature enables quick decisions, as factors that require immediate attention are processed first at the lowest layer. Experimental results indicate that the hierarchical approach has higher accuracy in recognition and efficient response time when compared to the existing approaches.