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

Text to Speech (TTS) is a form of speech synthesis where the text is converted into a spoken human-like voice output. The state of the art methods for TTS employs a neural network based approach. This work aims to look at some of the issues and limitations present in the current works, specifically Tacotron-2, and attempts to further improve its performance by modifying its architecture. The modified model uses Transformer network as a Spectrogram Prediction Network (SPN) and WaveGlow as an Audio Generation Network (AGN). For the modified model, performance improvements are seen in terms of the speech output generated for corresponding texts, the inference time taken for audio generation, and a Mean Opinion Score (MOS) of 4.10 (out of 5) is obtained. © 2020 IEEE.

World University root node

2020 7th International Conference on Soft Computing and Machine Intelligence, ISCMI 2020

Natural Text-to-Speech Synthesis by Conditioning Spectrogram Predictions from Transformer Network on WaveGlow Vocoder

Automating the classification of camera-obtained microscopic images of White Blood Cells (WBCs) and related cell subtypes has assumed importance since it aids the laborious manual process of review and diagnosis. Several State-Of-The-Art (SOTA) methods developed using Deep Convolutional Neural Networks suffer from the problem of domain shift - severe performance degradation when they are tested on data (target) obtained in a setting different from that of the training (source). The change in the target data might be caused by factors such as differences in camera/microscope types, lenses, lighting-conditions etc. This problem can potentially be solved using Unsupervised Domain Adaptation (UDA) techniques albeit standard algorithms presuppose the existence of a sufficient amount of unlabelled target data which is not always the case with medical images. In this paper, we propose a method for UDA that is devoid of the need for target data. Given a test image from the target data, we obtain its `closest-clone' from the source data that is used as a proxy in the classifier. We prove the existence of such a clone given that infinite number of data points can be sampled from the source distribution. We propose a method in which a latent-variable generative model based on variational inference is used to simultaneously sample and find the `closest-clone' from the source distribution through an optimization procedure in the latent space. We demonstrate the efficacy of the proposed method over several SOTA UDA methods for WBC classification on datasets captured using different imaging modalities under multiple settings.

Fulltext

IEEE TRANSACTIONS ON MEDICAL IMAGING

Target-Independent Domain Adaptation for WBC Classification Using Generative Latent Search

Extending the temperature measurement range and sustaining the system performance are important requirements in Rayleigh lidars. While replacing hardware and increasing the power aperture product are ways of achieving these objectives, the data analysis approach is another means for achieving the same. A method to retrieve atmospheric temperatures using the penalized maximum likelihood method after denoising the backscattered signal using the dictionary learning technique is presented and compared with the conventional method. The proposed combination has the advantage of improving the measurement range and reducing the standard error (SE) in temperatures. The penalized maximum likelihood function is solved using the method of successive approximations, and the SE in temperature is calculated using Monte Carlo simulations. Observations from the Rayleigh lidar at the National Atmospheric Research Laboratory, India, are used for testing the approach. When compared with the conventional method, the SE in temperatures improved by 5K at 84 km, and the average height improvement was about 6 km. © 2020 Society of Photo-Optical Instrumentation Engineers (SPIE).

Journal of Applied Remote Sensing

Dictionary learning technique and penalized maximum likelihood for extending measurement range of a Rayleigh lidar

Research Anthology on Smart Grid and Microgrid Development

A Machine Learning Approach for Anomaly Detection to Secure Smart Grid Systems

SARS-CoV-2 is a highly virulent and deadly RNA virus causing the Covid-19 pandemic and several deaths across the world. The pandemic is so fast that any concrete theory of sudden widespread of this disease is still not known. In this work, we studied and analyzed a large number of publicly available SARS-CoV-2 genomes across the world using the multifractal approach. The mutation events in the isolates obey the Markov process and exhibit very high mutational rates, which occur in six specific genes and highest in orf1ab gene, leading to virulent nature. f ( $\alpha$ ) analysis indicated that the isolates are highly asymmetric (left-skewed), revealing the richness of complexity and dominance by large fluctuations in genome structure organization. The values of Hq and Dq are found to be significantly large, showing heterogeneous genome structure self-organization, strong positive correlation in organizing the isolates and quite sensitive to fluctuations in and around it. We then present multiple-isolates hosts-virus interaction models and derived Price equation for the model. The phase plane analysis of the model showed asymptotic stability type bifurcation. The competition among the mutant isolates drives the trade-off of the dominant mutant isolates, otherwise confined to the present hosts. ### Competing Interest Statement The authors have declared no competing interest.

bioRxiv

Complexity in SARS-CoV-2 genome data: Price theory of mutant isolates

IIIT-Delhi

On DenseResNet: residual connection finetuning

Trusted AI

We show that given an embedding of an O(log n) genus bipartite graph, one can construct an edge weight function in logarithmic space, with respect to which the minimum weight perfect matching in the graph is unique, if one exists. As a consequence, we obtain that deciding whether such a graph has a perfect matching or not is in SPL. In 1999, Reinhardt, Allender and Zhou proved that if one can construct a polynomially bounded weight function for a graph in logspace such that it isolates a minimum weight perfect matching in the graph, then the perfect matching problem can be solved in SPL. In this paper, we give a deterministic logspace construction of such a weight function for O(log n) genus bipartite graphs.

45th International Symposium on Mathematical Foundations of Computer Science …

Efficient Isolation of Perfect Matching in O (log n) Genus Bipartite Graphs

Structural analysis helps in parsing the mathematical expressions. Various approaches for structural analysis have been reported in literature, but they mainly deal with online and printed expressions. In this work, two-dimensional, stochastic context-free grammar is used for the structural analysis of offline handwritten mathematical expressions in a document image. The spatial relation between characters in an expression has been incorporated so that the structural variability in handwritten expressions can be tackled. © 2020, Springer Nature Singapore Pte Ltd.

Department of Electrical Engineering

Advances in Intelligent Systems and Computing

Structural Analysis of Offline Handwritten Mathematical Expressions

The cutting-edge technologies have improved the journey of vehicle users on the road using different vehicular-based structures (vehicular ad-hoc networks (VANETs), vehicular cloud computing (VCC), and Internet of Vehicles (IoV)). Vehicles generally transmit different road-side data (e.g., traffic, details of nearby vehicles, emergency/environment, location, etc.) from/to another end in a public environment. Hence, three key challenges on data are as follows; (1) secure and effective data storage in the vehicular cloud (VC) (2) efficient data extraction from the VC and (3) reliable data transmission between two entities. Some general methods are to search/store data from/in the cloud, but these techniques are only for data storage and extraction. Next, various communication schemes are proposed for VANETs, VCC, and IoV, but they need high computational resources, and these schemes are vulnerable to multiple security attacks. We find that there is no system, which provides all these data operations (storage, searching, and communication) for vehicle users. In this paper, we propose a dependable system (named as SearchCom) for vehicle applicants to overcome above-mentioned challenges and provide operational facility. The SearchCom is analyzed to verify its security strengths against various security attacks and its performance results in execution time, memory requirement, communication overhead, searching, and energy consumption. © 2020 ACM.