Abstract

In this paper, a dialogue system for Hospital domain in Telugu, which is a resource-poor Dravidian language, has been built. It handles various hospital and doctor related queries. The main aim of this paper is to present an approach for modelling a dialogue system in a resource-poor language by combining linguistic and domain knowledge. Focusing on the question answering aspect of the dialogue system, we identified Question Classification and Query Processing as the two most important parts of the dialogue system. Our method combines deep learning techniques for question classification and computational rule-based analysis for query processing. Human evaluation of the system has been performed as there is no automated evaluation tool for dialogue systems in Telugu. Our system achieves a high overall rating along with a significantly accurate context-capturing method as shown in the results.

Abstract

The Earth’s climate is not static; it changes according to the natural and anthropogenic climate variability. Anthropogenic forcing due to increase of greenhouse gases in the atmosphere has driven changes in climate variables globally. Changes in climatological variables have severe impact on global hydrological cycle affecting the severity and occurrence of natural hazards such as floods and droughts. Estimation of projections under climate signals with statistical and dynamic downscaling models and integration with water resource management models for the impact assessment have gained much attention. The fine-resolution climate change predictions of dynamic regional climate model (RCM) outputs, which include regional parameterization, have been widely applied in the hydrological impact assessment studies. Advancement of the Coordinated Regional Downscaling Experiment (CORDEX) program has enabled the use of RCMs in regional impact assessment which has progressed in recent years. CORDEX model outputs were considered to be valuable in terms of establishing large ensembles of climate projections based on regional climate downscaling all over the world. However, the simulations of RCM outputs have to be evaluated to check the reliability in reproducing the observed climate variability over a region. The present study demonstrates the use of bias-corrected CORDEX model simulation in analyzing the regional-scale climatology at river basin scale, Krishna river basin (KRB), India. The precipitation and temperature simulations from CORDEX models with RCP 4.5 were evaluated for the historical data for the period of 1965 to 2014 with India Meteorological Department (IMD) gridded rainfall and temperature data sets cropped over the basin. The projected increase

Abstract

Conservation efforts to protect biodiversity rely on an accurate identification process. In the case of plant identification, traditional methods used are manual, time-consuming and require a degree of expertise to operate. As a result, there is an increasing interest today for an automated plant identification system. Such a system can help in aiding plant-related education, promoting ecotourism, creating a digital heritage for plant species among many others. We propose a solution using modern convolutional neural network architectures which achieves stateof-the-art performance for plant classification in the wild. An exhaustive set of experiments are performed to classify 112 species of plants from the challenging Indic-Leaf dataset. The best performing model gives Top-1 precision of 90.08 and Top-5 precision of 96.90.

Abstract

Modeling atmospheric turbulence is a challenging problem since the light rays arbitrarily bend before entering the camera. Such models are critical to extend computer vision solutions developed in the laboratory to real-world use cases. Simulating atmospheric turbulence by using statistical models or by computer graphics is often computationally expensive. To overcome this problem, we train a generative adversarial network which outputs an atmospheric turbulent image by utilizing less computational resources than traditional methods. We propose a novel loss function to efficiently learn the atmospheric turbulence at the finer level. Experiments show that by using the proposed loss function, our network outperforms the existing state-of-the-art image to image translation network in turbulent image generation. We also perform extensive ablation studies on the loss function to demonstrate the improvement in the perceptual quality of turbulent images.

Abstract

This paper describes the Neural Machine Translation systems used by IIIT Hyderabad (CVIT-MT) for the translation tasks part of WAT-2019. We participated in tasks pertaining to Indian languages and submitted results for English-Hindi, Hindi-English, English-Tamil and Tamil-English language pairs. We employ Transformer architecture experimenting with multilingual models and methods for lowresource languages.

Abstract

This paper proposes the use of Shewhart test to reduce the number of data-transmissions in IoT networks. It is shown to outperform the widely-used least mean square (LMS) based data reduction method in terms of the number of data-transmissions, implementation complexity and mean square error (MSE) in prediction of time-series data at the sink node based on the partial transmissions of the measured time-series data from the sensor node. The paper also proposes the use of piggybacking and interpolation to further reduce the MSE of the estimated time-series data at the sink node without increasing the number of packet transmissions. The time-series data used for the comparison of data reduction algorithms is a set of measured temperature values in indoor and outdoor scenarios for four days using custom-designed wireless sensor nodes. To express the effectiveness of the piggybacked transmissions on battery lifetime, the total current consumption of the sensor node is measured for different number of piggybacks and corresponding battery lifetime is estimated. It is shown that the proposed piggyback approach significantly reduces the MSE at the cost of slight decrease in battery-lifetime. Published in: 2019 IEE

Abstract

In this paper, we study scalable parallel algorithms for estimating the farness-centrality value of the nodes in a given undirected and connected graph. Our algorithms consider approaches that are more suitable for sparse graphs. To this end, we propose four optimization techniques based on removing redundant nodes, removing identical nodes, removing chain nodes, and making use of decomposition based on the biconnected components of the input graph. We test our techniques on a collection of real-world graphs for the time taken and the average error percentage. We further analyze the applicability of our techniques on various classes of real-world graphs. We suggest why certain techniques work better on certain classes of graphs.

Abstract

Smart grid (SG) technology is regarded as the next generation of the power grid, which makes use of the two-way flows of electricity, as well as information, Pathways to a Smarter Power System. https://doi.org/10.1016/B978-0-08-102592-5.00013-2 © 2019 Elsevier Ltd. All rights reserved. 371 in order to build a broadly distributed automated energy delivery network. The Smart grid is a critical infrastructure that plays a critical role in the daily lives of people. Both the Smart Grid and its data must be protected from cyberattacks at all times [1, 2]. In SG, four components are present, namely sensing, control, communication, and actuation systems. The most important component of SG is the smart meter (SM), which consists of sensing and communication modules. In addition, there are service systems from the service providers (SPs), which consist of several modules for control, communication, and actuation. SMs are broadly used in homes for monitoring energy consumption in real time. Moreover, power pricing information to consumers is also provided by the SMs [3–6]. A detailed survey on SG including its applications can be found in [7–14]. In the following subsections, we now discuss the smart grid framework developed by the National Institute of Standards and Technology (NIST) and its taxonomy, based on domains and targeted research areas.

Abstract

It has long been understood that there is an interplay between affect and cognition((Kort & Reilly, 2003), but this interaction, based on the recent chess studies((Guntz, Crowley, Vaufreydaz, Balzarini, & Dessus, 2018), is much more intertwined than what the established theories postulate. To understand the underlying mechanisms in greater detail we propose an integrated model using Chrest and Nengo. We analyze the results based on simulations with data from previous empirical studies. Keywords: Affect, Cognition , Chess, Nengo , Chrest

Abstract

A method is proposed for time delay estimation (TDE) from mixed source (speaker) signals collected at two spatially separated microphones. The key idea in this proposal is that the crosscorrelation between corresponding segments of the mixed source signals is computed using the outputs of single frequency filtering (SFF) obtained at several frequencies, rather than using the collected waveforms directly. The advantage of the SFF output is that it will have high signal-to-noise ratio regions in both time and frequency domains. Also it gives multiple evidences, one from each of the SFF outputs. These multiple evidences are combined to obtain robustness in the TDE. The estimated time delays can be used to determine the number of speakers present in the mixed signals. The TDE is shown to be robust against different types and levels of degradations. The results are shown for actual mixed signals collected at two spatially separated microphones in a live laboratory environment, where the mixed signals contain speech from several spatially distributed speakers.