Abstract

Problems involving code-mixed language are often plagued by a lack of resources and an absence of materials to perform sophisticated transfer learning with. In this paper we describe our submission to the Sentimix Hindi-English task involving sentiment classification of code-mixed texts, and with an F1 score of 67.1%, we demonstrate that simple convolution and attention may well produce reasonable results.

Abstract

Resolution of POF based pressure sensor was improved by suppressing voltage fluctuation in LED-fiber-Photodiode setup using time windowing technique. This reduced the standard deviation by 98%, thereby showing significant improvement in gait analysis images

Abstract

With the development in the field of quantum physics, several methods for building a quantum computer have emerged. These differ in qubit technologies, interaction topologies, and noise characteristics. In this article, insights are given into the circuit‐ centric architecture design of Noisy Intermediate‐Scale Quantum (NISQ) devices. The dependence of the circuit size, circuit depth on the interaction and connection between different qubits present in quantum hardware are discussed. A noise‐aware procedure is presented which helps in determining the optimal interactions between different qubits of a quantum chip to execute a given circuit in the most efficient way possible. In this article, the 5‐qubit hardware in a noiseless setting is illustrated with an example. Also, a benchmark‐driven analysis is performed to show the importance of noise adaptivity in determining the hardware reliability. It is concluded that a generalized and flexible procedure such as this approach can aid in determining the design of hardware accurately for which the circuit runs efficiently, that is, with the least number of clock cycles, the lowest gate operations, and noise‐based errors

Abstract

A majority of the real world applications involve mapping the task at hand to an optimization problem, whose solutions are either fully known or can be approximated by relaxing some of its constraints. In general, the increment in problem size or augmentation of additional constraints can increase the hardness of the problem. This means that the computational resources required in solving them scale-up exponentially, making them computationally intractable. Hence, researchers worldwide have been working on developing efficient tools and techniques for solving these problems efficiently with the available computing power. With the recent d

Abstract

Measurement of the electric field radiations by ultra-wideband antenna is a challenging task. In general, all the antennas are measured for the radiation pattern and return loss to estimate their overall performance. However, when it comes to the antennas that are fed by very high voltages (HV), in the order of few hundreds of kilo volts and the input is not continuous, the measurements become complex. There arises an ambiguity in deciding about the peak value of the high electric fields generated, especially when the HV electric discharge mechanism is involved in the radiation of the antenna. The electric discharges are, in general, not repetitive. This paper proposes a measurement technique that is well suited for measuring the high electric fields, especially in the order of hundreds of kV/m. The proposed measurement procedure is verified and validated by experiments. The repeatability and reliability of the proposed measurement procedure are checked and presented here.

Abstract

We introduce PeeledHuman - a novel shape representation of the human body that is robust to self-occlusions. PeeledHuman encodes the human body as a set of Peeled Depth and RGB maps in 2D, obtained by performing raytracing on the 3D body model and extending each ray beyond its first intersection. This formulation allows us to handle self-occlusions efficiently compared to other representations. Given a monocular RGB image, we learn these Peeled maps in an end-to-end generative adversarial fashion using our novel framework - PeelGAN. We train PeelGAN using a 3D Chamfer loss and other 2D losses to generate multiple depth values per-pixel and a corresponding RGB field per-vertex in a dual-branch setup. In our simple non-parametric solution, the generated Peeled Depth maps are back-projected to 3D space to obtain a complete textured 3D shape. The corresponding RGB maps provide vertex-level texture details. We compare our method with current parametric and non-parametric methods in 3D reconstruction and find that we achieve state-of-the-art-results. We demonstrate the effectiveness of our representation on publicly available BUFF and MonoPerfCap datasets as well as loose clothing data collected by our calibrated multi-Kinect setup.

Abstract

Shayari is a form of poetry mainly popular in the Indian subcontinent, in which the poet expresses his emotions and feelings in a very poetic manner. It is one of the best ways to express our thoughts and opinions. Therefore, it is of prime importance to have an annotated corpus of Hindi shayaris for the task of sentiment analysis. In this paper, we introduce SUKHAN, a dataset consisting of Hindi shayaris along with sentiment polarity labels. To the best of our knowledge, this is the first corpus of Hindi shayaris annotated with sentiment polarity information. This corpus contains a total of 733 Hindi shayaris of various genres. Also, this dataset is of utmost value as all the annotation is done manually by five annotators and this makes it a very rich dataset for training purposes. This annotated corpus is also used to build baseline sentiment classification models using machine learning techniques.

Abstract

Development of Force-based seismic design is historical, and every earthquake has thrown new challenges for already existing assumptions and principles, which helped in the evolution of force-based design. Although many of the assumptions were answered but some remained unaltered and unanswered. Hence, this paved way for developing new theories and philosophies and one such theory is Direct Displacement-Based seismic Design (DDBD). Indian Standard codes, IS 1893 and IS 13920, underwent a major revision in the year 2016, considering the latest developments in Force based design principles. Current work is an attempt to understand the performance comparison between latest IS 1893 & IS 13920 with DDBD method. Initially, three buildings i.e., 3, 6 and 9 storey buildings were designed using DDBD at inter-storey drift of 4% with a performance objective to achieve “No collapse”. The spectrum developed using IS code equations does not represent ideal displacement spectrum shape and hence PGA, T and CA dependent equations are used for DDBD methodology of design. PGA of 0.36g is used in the current study which represents the highest seismic zone in India. Later these buildings were subjected to lateral displacement using displacement-based pushover analysis method. The inelastic strength along with drifts achieved by DDBD are compared with the IS code compliant buildings. From the comparison, it was observed that inelastic strength achieved by 6,9 storey buildings designed according to FBD are greater than the buildings designed according to DDBD at “No Collapse”. For 3 storey structure, the strength achieved by FBD is less than DDBD. If zone is reduced from Zone V to IV or PGA is increased from 0.36g, then inelastic strength achieved by DDBD with performance objective “No Collapse” would be more than FBD requiring revision of response reduction values for different categories of buildings or PGA values in respective seismic zones. The inelastic drifts achieved by both the methods are comparable only for 3 storey structure.

Abstract

An earthquake Mw7.8 occurred on 25 April 2015 in a seismically dominant area of Nepal. No major earthquakes (>Mw7.0) were recorded in the past two decades, though Nepal is one of the active seismic regions in the world. The combined effect of the main event and the aftershocks, which are temporally distributed over 40 days, has led to a huge devastation, killing thousands of people in the epi-central area. A complete seismic catalog and influencing parameters such as type of fault, epi-central/hypo-central distance, type of soil are required for a region to assess seismic hazard scenario. The seismic hazard can be represented in terms of Peak Ground Acceleration (PGA), which is the most widely used parameter in strong motion studies. The main objective of this paper is to understand the characteristics of ground motion of the 2015 Nepal earthquake. Since the available ground motion records are limited, an attempt has been made to generate ground motion records using a modified semi-empirical approach. Contour maps of PGA and intensity have also been generated. It is observed that around 22% of buildings experienced an intensity of VI on MMI scale at Bageswari mandal. A few buildings have experienced an intensity of VII in Thankot, Seuchatar, Satikhel, Mahadevathan, Indrayani, Chhaimale, and Balambu.

Abstract

In last two decades, operational modal analysis is widely used for modal identification of structures which uses ambient responses of structure for determining the modal parameters and Frequency Domain Decomposition is one of the most popular method, but seismic excitations and high damping values of structures do not satisfy the assumptions of this classical method. The dynamic modal parameters derived from earthquake responses are more reliable than those from small amplitude responses since dynamic properties of civil structures are dependent on amplitude. In recent years, a new algorithm known as refined Frequency Domain Decomposition (rFDD) is developed which uses structural response to strong ground motions that can be short in duration and is applicable for structures with light to heavy damping. The refined Frequency Domain Decomposition (rFDD) algorithm was initially developed using synthetic seismic response signals as input and later real strong motion response records were used to demonstrate the effectiveness of this algorithm for modal dynamic identification of structures under real earthquake excitations. This method is well developed to counter the problems that normally arise in modal identification and in cases of heavy damping and non-stationarity of the signal, making it lengthy and tedious to process. In this paper the main feature of this method which is Integrated Power Spectral Density matrix computation is taken and applied on mild earthquakes responses of structure. The modal parameters identified are then compared with target values obtained from classical FDD method using ambient responses and interpreted, presenting a case for usage of rFDD in simplified manner in modal dynamic identification of structures. Keywords: Operational Modal analysis, mild earthquake response, refined Frequency Domain Decomposition