Indian Institute of Technology Madras

About: Indian Institute of Technology Madras is a facility organization based out in Chennai, Tamil Nadu, India. It is known for research contribution in the topics: Catalysis & Heat transfer. The organization has 20118 authors who have published 36499 publications receiving 590447 citations.

Functionalized graphene reinforced thermoplastic nanocomposites as strain sensors in structural health monitoring

Abstract: Functionalized graphene–polyvinylidene fluoride (f-G–PVDF) nanocomposite films were synthesized using a simple solvent casting technique. The investigation demonstrates the real time strain response of f-G–PVDF nanocomposites on the macro-scale under tensile loads and the use of this nanocomposite as strain sensor. This is ascribed to the variation of electrical properties of graphene–polymer upon mechanical deformation at the nanoscale.

Ablation-CAM: Visual Explanations for Deep Convolutional Network via Gradient-free Localization

Abstract: In response to recent criticism of gradient-based visualization techniques, we propose a new methodology to generate visual explanations for deep Convolutional Neural Networks (CNN) - based models. Our approach – Ablation-based Class Activation Mapping (Ablation CAM) uses ablation analysis to determine the importance (weights) of individual feature map units w.r.t. class. Further, this is used to produce a coarse localization map highlighting the important regions in the image for predicting the concept. Our objective and subjective evaluations show that this gradient-free approach works better than state-of-the-art Grad-CAM technique. Moreover, further experiments are carried out to show that Ablation-CAM is class discriminative as well as can be used to evaluate trust in a model.

Performance‐based optimal design and rehabilitation of water distribution networks using life cycle costing

Abstract: [1] A new multiobjective formulation is proposed for the optimal design and rehabilitation of a water distribution network, with minimization of life cycle cost and maximization of performance as objectives. The life cycle cost is considered to comprise the initial cost of pipes, the cost of replacing old pipes with new ones, the cost of cleaning and lining existing pipes, the expected repair cost for pipe breaks, and the salvage value of the pipes that are replaced. The performance measure proposed in this study is a modification to the resilience index to suit application to water distribution networks with multiple sources. A new heuristic method is proposed to obtain the solution for the design and rehabilitation problem. This heuristic method involves selection of various design and rehabilitation alternatives in an iterative manner on the basis of the improvement in the network performance as compared to the change in the life cycle cost on implementation of the alternatives. The solutions obtained from the heuristic method are used as part of the initial population set of the multiobjective, nondominated sorting genetic algorithm (NSGA-II) in order to improve the search process. Using a sample water distribution network, the modified resilience index proposed is shown to be a good indicator of the uncertainty handling ability of the network.