Department of Computer Science & Engineering

&ldquo;IIT Jodhpur is committed to impart quality education to facilitate the creation of qualified, competent and responsible human resources to meet the emerging technological challenges of the world. It is working towards creating an equitable learning platform that is positive, accessible and effective in the face of the modern societal and scientific advancements.

Flexibility in curriculum, in all its academic programs, encompassing elements of teaching, research, creativity, innovation and entrepreneurship is the most unique feature of IIT Jodhpur. The Department of Science and Technology (DST), under the National Mission on Cyber Physical Systems, has identified IIT Jodhpur as a Technology &amp; Innovation Hub in the area of Augmented Reality and Virtual Reality. The institute has also been nominated as the coordinator for the Jodhpur City and Knowledge Cluster by the Office of Principal Scientific Adviser, Government of India.

The sprawling 852-acre IIT Jodhpur campus is a majestic tribute to the deep-rooted diverse cultural heritage of the region. The campus is an architectural masterpiece, which is slated to become an international exemplar of sustainability with net-zero energy, and advanced water and waste management strategies.&rdquo;

IIT Jodhpur

School of Artificial Intelligence and Data Science (AIDE)

Functional Magnetic Resonance Imaging (fMRI) has opened ways to look inside active human brain. However, fMRI signal is an indirect indicator of underlying neuronal activity and has lowtemporal resolution due to acquisition process. This paper proposes autoregressive hidden Markov model with missing data (AR-HMM-md) framework which aims at addressing aforementioned issues while allowing accurate capturing of fMRI time series characteristics. The proposed work models unobserved neuronal activity over time as sequence of discrete hidden states, and shows how exact inference can be obtained with missing fMRI data under the "Missing not at Random" (MNAR) mechanism. This mechanism requires explicit modelling of the missing data along with the observed data. The performance is evaluated by observing convergence characteristic of loglikelihoods and classification capability of the proposed model over existing models for two fMRI datasets. The classification is performed between real fMRI time series from a task-based experiment and randomly-generated time series. Another classification experiment is performed between children and elder subjects using fMRI time series from resting-state data. The proposed model captured the fMRI characteristics efficiently and thus converged to better posterior probability resulting into higher classification accuracy over existing models for both the datasets. © 2016 ACM.

World University root node

ACM International Conference Proceeding Series

Autoregressive hidden Markov model with missing data for modelling functional MR imaging data

In the real world, we often encounter varying membership grades due to varying information source values. The fuzzy rough set model is refurbished to develop probabilistic variable precision fuzzy rough set (P-VP-FRS) to deal with this imprecision. The main inspiration behind the proposed P-VP-FRS is our inability to precisely represent the imprecision, which necessitates generalization in the approximations. The adjustable parameters in P-VP-FRS control the tradeoff between the generalization and accuracy. A few measures for quality of approximation and generalization are proposed. The usefulness of P-VP-FRS is shown through a case study. © 1993-2012 IEEE.

IEEE Transactions on Fuzzy Systems

Probabilistic variable precision fuzzy rough sets

Introducing recent advances in the machine learning techniques to state-of-the-art discrete choice models, we develop an approach to infer the unique and complex decision making process of a decision-maker (DM), which is characterized by the DM's priorities and attitudinal character, along with the attributes interaction, to name a few. On the basis of exemplary preference information in the form of pairwise comparisons of alternatives, our method seeks to induce a DM's preference model in terms of the parameters of recent discrete choice models. To this end, we reduce our learning function to a constrained non-linear optimization problem. Our learning approach is a simple one that takes into consideration the interaction among the attributes along with the priorities and the unique attitudinal character of a DM. The experimental results on standard benchmark datasets suggest that our approach is not only intuitively appealing and easily interpretable but also competitive to state-of-the-art methods. © 1989-2012 IEEE.

IEEE Transactions on Knowledge and Data Engineering

On learning of choice models with interactive attributes

In this paper, we study maximal monotone inclusions from the perspective of gap functions. We propose a very natural gap function for an arbitrary maximal monotone inclusion and will demonstrate how naturally this gap function arises from the Fitzpatrick function, which is a convex function, used to represent maximal monotone operators. This allows us to use the powerful strong Fitzpatrick inequality to analyse solutions of the inclusion. We also study the special cases of a variational inequality and of a generalized variational inequality problem. The associated notion of a scalar gap is also considered in some detail. Corresponding local and global error bounds are also developed for the maximal monotone inclusion. © 2015, Springer Science+Business Media New York.

Journal of Optimization Theory and Applications

Maximal Monotone Inclusions and Fitzpatrick Functions

Computational modeling of the spontaneous dynamics over the whole brain provides critical insight into the spatiotemporal organization of brain dynamics at multiple resolutions and their alteration to changes in brain structure (e.g. in diseased states, aging, across individuals). Recent experimental evidence further suggests that the adverse effect of lesions is visible on spontaneous dynamics characterized by changes in resting state functional connectivity and its graph theoretical properties (e.g. modularity). These changes originate from altered neural dynamics in individual brain areas that are otherwise poised towards a homeostatic equilibrium to maintain a stable excitatory and inhibitory activity. In this work, we employ a homeostatic inhibitory mechanism, balancing excitation and inhibition in the local brain areas of the entire cortex under neurological impairments like lesions to understand global functional recovery (across brain networks and individuals). Previous computational and empirical studies have demonstrated that the resting state functional connectivity varies primarily due to the location and specific topological characteristics of the lesion. We show that local homeostatic balance provides a functional recovery by re-establishing excitation-inhibition balance in all areas that are affected by lesion. We systematically compare the extent of recovery in the primary hub areas (e.g. default mode network (DMN), medial temporal lobe, medial prefrontal cortex) as well as other sensory areas like primary motor area, supplementary motor area, fronto-parietal and temporo-parietal networks. Our findings suggest that stability and richness similar to the normal brain dynamics at rest are achievable by re-establishment of balance. © 2016 Elsevier Inc.

Fulltext

NeuroImage

Does the regulation of local excitation-inhibition balance aid in recovery of functional connectivity? A computational account

Feed manufacturing faces enormous challenges and with the demand for good quality feed increasing gradually, it becomes essential to improve the processes in a feed mill. This article provides a brief overview of the different processes in feed manufacturing and identifies the critical process parameters. Five critical parameters are identified where the production rate is the output parameter. Mash feed size, steam temperature, conditioning time and feed rate are the input parameters. Artificial neural network is the methodology which is used to optimize the process parameters. Root mean squared error and coefficient of determination and computation time are used as performance measures and it is observed that Polak-Ribiere conjugate gradient backpropagation training function with log sigmoid - pure linear transfer function combination provided good results among the different available alternatives. The process parameters are then optimized using the appropriate ideal settings of neural network parameters. This model is extremely useful for the prediction of production rate for 1 specific recipe in a feed mill. © 2015 Elsevier B.V..

Computers and Electronics in Agriculture

Optimization of process parameters in feed manufacturing using artificial neural network

Several new aggregation operators are proposed in the context of multicriteria decision making (MCDM) in the linguistic domain. The proposed operators first infer the discrimination index, based on the extent of variability in the various linguistic evaluations against a criterion. This value is then utilized in the actual aggregation step to discriminate among the alternatives. Besides, the proposed operators also take into account the a priori weights associated with the criteria. The proposed concepts are illustrated through an example in group MCDM. © 2016 Wiley Periodicals, Inc.

International Journal of Intelligent Systems

Linguistic Discriminative Aggregation in Multicriteria Decision Making

This handbook is an endeavour to cover many current, relevant, and essential topics related to decision sciences in a scientific manner. Using this handbook, graduate students, researchers, as well as practitioners from engineering, statistics, sociology, economics, etc. will find a new and refreshing paradigm shift as to how these topics can be put to use beneficially. Starting from the basics to advanced concepts, authors hope to make the readers well aware of the different theoretical and practical ideas, which are the focus of study in decision sciences nowadays. It includes an excellent bibliography/reference/journal list, information about a variety of datasets, illustrated pseudo-codes, and discussion of future trends in research. Covering topics ranging from optimization, networks and games, multi-objective optimization, inventory theory, statistical methods, artificial neural networks, times series analysis, simulation modeling, decision support system, data envelopment analysis, queueing theory, etc., this reference book is an attempt to make this area more meaningful for varied readers. Noteworthy features of this handbook are in-depth coverage of different topics, solved practical examples, unique datasets for a variety of examples in the areas of decision sciences, in-depth analysis of problems through colored charts, 3D diagrams, and discussions about software. © 2017 by Taylor & Francis Group, LLC. All rights reserved.

Decision Sciences: Theory and Practice

Convex functions are central to the study of optimization. The main property that makes it pivotal to the study of optimization is that every local minimizer of a convex function over a convex set is a global minimizer. The very large number of problems that arise in engineering, business, economics, and finance can be modeled as a convex optimization problem. In this chapter we first focus on the various examples and important properties of convex functions, which includes the study of subdifferentials and their computation. We also focus on the various important models of convex optimization problems and then on their optimality and duality properties. We end the chapter by a study of monotone operators, variational inequalities, and of the proximal point method for convex optimization problems. © 2017 by Taylor & Francis Group, LLC.

Convex functions in optimization

We develop new methods for the representation of uncertainty in the granularized information source values by making use of the entropy framework in the possibilistic domain. An information-theoretic entropy function is used to map the information source values to information (entropy) values. We term a collection of such information values as an information set. The information values are then used in an adaptive form of this entropy function to formulate Shannon transforms. A few uncertainty measures are derived from these transforms for the quantification of uncertainty. Information set is also extended to other domains, such as probabilistic, intuitionistic, and probabilistic-intuitionistic domains. A biometric application is included to demonstrate the usefulness of the study. © 1993-2012 IEEE.