9 research outputs found
Impact of global warming and other climatic condition for generation of wind energy and assessing the wind potential for future trends
No full text
Improving Intent Classication By Automatic Data Augmentation Using Word Sense Disambiguation
No full textabstract: Virtual digital assistants are automated software systems which assist humans by understanding natural languages such as English, either in voice or textual form. In recent times, a lot of digital applications have shifted towards providing a user experience using natural language interface. The change is brought up by the degree of ease with which the virtual digital assistants such as Google Assistant and Amazon Alexa can be integrated into your application. These assistants make use of a Natural Language Understanding (NLU) system which acts as an interface to translate unstructured natural language data into a structured form. Such an NLU system uses an intent finding algorithm which gives a high-level idea or meaning of a user query, termed as intent classification. The intent classification step identifies the action(s) that a user wants the assistant to perform. The intent classification step is followed by an entity recognition step in which the entities in the utterance are identified on which the intended action is performed. This step can be viewed as a sequence labeling task which maps an input word sequence into a corresponding sequence of slot labels. This step is also termed as slot filling.
In this thesis, we improve the intent classification and slot filling in the virtual voice agents by automatic data augmentation. Spoken Language Understanding systems face the issue of data sparsity. The reason behind this is that it is hard for a human-created training sample to represent all the patterns in the language. Due to the lack of relevant data, deep learning methods are unable to generalize the Spoken Language Understanding model. This thesis expounds a way to overcome the issue of data sparsity in deep learning approaches on Spoken Language Understanding tasks. Here we have described the limitations in the current intent classifiers and how the proposed algorithm uses existing knowledge bases to overcome those limitations. The method helps in creating a more robust intent classifier and slot filling system.Dissertation/ThesisMasters Thesis Computer Science 201
Hana: An Open-Domain Chatbot Application for Language Learning
No full textabstract: Learning a new language can be very challenging. One significant aspect of learning a language is learning how to have fluent verbal and written conversations with other people in that language. However, it can be difficult to find other people available with whom to practice conversations. Additionally, total beginners may feel uncomfortable and self-conscious when speaking the language with others. In this paper, I present Hana, a chatbot application powered by deep learning for practicing open-domain verbal and written conversations in a variety of different languages. Hana uses a pre-trained medium-sized instance of Microsoft's DialoGPT in order to generate English responses to user input translated into English. Google Cloud Platform's Translation API is used to handle translation to and from the language selected by the user. The chatbot is presented in the form of a browser-based web application, allowing users to interact with the chatbot in both a verbal or text-based manner. Overall, the chatbot is capable of having interesting open-domain conversations with the user in languages supported by the Google Cloud Translation API, but response generation can be delayed by several seconds, and the conversations and their translations do not necessarily take into account linguistic and cultural nuances associated with a given language
Representation, Exploration, and Recommendation of Music Playlists
No full textabstract: Playlists have become a significant part of the music listening experience today because of the digital cloud-based services such as Spotify, Pandora, Apple Music. Owing to the meteoric rise in usage of playlists, recommending playlists is crucial to music services today. Although there has been a lot of work done in playlist prediction, the area of playlist representation hasn't received that level of attention. Over the last few years, sequence-to-sequence models, especially in the field of natural language processing have shown the effectiveness of learned embeddings in capturing the semantic characteristics of sequences. Similar concepts can be applied to music to learn fixed length representations for playlists and the learned representations can then be used for downstream tasks such as playlist comparison and recommendation.
In this thesis, the problem of learning a fixed-length representation is formulated in an unsupervised manner, using Neural Machine Translation (NMT), where playlists are interpreted as sentences and songs as words. This approach is compared with other encoding architectures and evaluated using the suite of tasks commonly used for evaluating sentence embeddings, along with a few additional tasks pertaining to music. The aim of the evaluation is to study the traits captured by the playlist embeddings such that these can be leveraged for music recommendation purposes. This work lays down the foundation for analyzing music playlists and learning the patterns that exist in the playlists in an end-to-end manner. This thesis finally concludes with a discussion on the future direction for this research and its potential impact in the domain of Music Information Retrieval.Dissertation/ThesisMasters Thesis Computer Science 201
Effects of fisetin on hyperhomocysteinemia- induced experimental endothelial dysfunction and vascular dementia
No full textThis study was designed to investigate the effect of fisetin (FST) on hyperhomocysteinemia (HHcy) - induced experimental endothelial dysfunction (ED) and vascular dementia (VaD) in rats. Eight groups of Wistar rats were randomly divided as control, vehicle control, L-methionine, FST (5, 10 and 25 mg/kg, p.o.), FST-per se (25 mg/kg, p.o.) and donepezil (0.1 mg/kg, p.o.) respectively. L-Methionine administration (1.7 g/kg, p.o.) for 32 days induced HHcy. ED and VaD induced by HHcy were determined by vascular reactivity measurements, behavioral analysis using Morris water maze and Y-maze along with a biochemical and histological evaluation of thoracic aorta and brain tissues. Administration of L-methionine developed behavioral deficits, triggered brain lipid peroxidation (LPO) and compromised brain acetylcholinesterase activity (AChE) and reduced the levels of brain superoxide dismutase (SOD) ,brain catalase (CAT),brain reduced glutathione (GSH) and serum nitrite, increased serum homocysteine and cholesterol levels. These effects were accompanied by decreased vascular NO bioavailability, marked intimal thickening of the aorta and multiple necrotic foci in brain cortex. HHcy induced alterations in the activities of SOD, CAT, GSH, AChE, LPO, behavioral deficits, ED and histological aberrations were significantly attenuated by treatment with fisetin in a dose-dependent manner. Collectively our results indicate that fisetin exerts endothelial and neuroprotective effects against HHcy-induced ED and VaD.The accepted manuscript in pdf format is listed with the files at the bottom of this page. The presentation of the authors' names and (or) special characters in the title of the manuscript may differ slightly between what is listed on this page and what is listed in the pdf file of the accepted manuscript; that in the pdf file of the accepted manuscript is what was submitted by the author
Recovery of renewable aromatic and aliphatic hydrocarbon resources from microwave pyrolysis/co-pyrolysis of agro-residues and plastics wastes.
No full textThe present study focussed on recovering the valuable carbon resources from agro-residues (wheat straw, rice husk) and waste plastics (polypropylene, polystyrene) using microwave pyrolysis and co-pyrolysis. The main objective of this study is to investigate the effect of the susceptor blending mechanism on the co-pyrolysis product distribution. Graphite was mixed with feedstock in a new approach to achieving homogeneity, and microwave power of 600 W was used. The average heating rate (52-67 (°C/min)), microwave energy required (2267-2936 (J/g)), heat energy utilized (1410-1444 (J/g)), and conductive heat losses (85-110 (J/g)) were analyzed. The selectivity of cyclic alkanes and alkenes (65.5%) was found to be high in polypropylene pyrolysis oil. Polystyrene pyrolysis oil predominantly contained cyclooctatetraene (61%) compound. Bio-oil obtained from wheat straw predominantly contained aromatic hydrocarbons (85%), whereas rice husk oil also contains high selectivity aromatic hydrocarbons (37.8%) along with aliphatic hydrocarbons (54.9%). The co-pyrolysis oils has high selectivity of aromatics.The corresponding author acknowledges the PanditDeendayal Petroleum University management for providing the faculty seed grant
Balloon-borne aerosol–cloud interaction studies (BACIS): field campaigns to understand and quantify aerosol effects on clouds
No full textA better understanding of aerosol–cloud interaction processes is important to quantify the role of clouds and aerosols on the climate system. There have been significant efforts to explain the ways aerosols modulate cloud properties. However, from the observational point of view, it is indeed challenging to observe and/or verify some of these processes because no single instrument or platform has been proven to be sufficient. Discrimination between aerosol and cloud is vital for the quantification of aerosol–cloud interaction. With this motivation, a set of observational field campaigns named balloon-borne aerosol–cloud interaction studies (BACIS) is proposed and conducted using balloon-borne in situ measurements in addition to the ground-based (lidar; mesosphere, stratosphere and troposphere (MST) radar; lower atmospheric wind profiler; microwave radiometer; ceilometer) and space-borne (CALIPSO) remote sensing instruments from Gadanki (13.45◦ N, 79.2◦ E), India. So far, 15 campaigns have been conducted as a part of BACIS campaigns from 2017 to 2020. This paper presents the concept of the observational approach, lists the major objectives of the campaigns, describes the instruments deployed, and discusses results from selected campaigns. Balloon-borne measurements of aerosol and cloud backscatter ratio and cloud particle count are qualitatively assessed using the range-corrected data from simultaneous observations of ground-based and space-borne lidars. Aerosol and cloud vertical profiles obtained in multi-instrumental observations are found to reasonably agree. Apart from this, balloon-borne profiling is found to provide information on clouds missed by ground-based and/or space-borne lidar. A combination of the Compact Optical Backscatter AerosoL Detector (COBALD) and Cloud Particle Sensor (CPS) sonde is employed for the first time in this study to discriminate cloud and aerosol in an in situ profile. A threshold value of the COBALD colour index (CI) for ice clouds is found to be between 18 and 20, and CI values for coarse-mode aerosol particles range between 11 and 15. Using the data from balloon measurements, the relationship between cloud and aerosol is quantified for the liquid clouds. A statistically significant slope (aerosol–cloud interaction index) of 0.77 found between aerosol backscatter and cloud particle count reveals the role of aerosol in the cloud activation process. In a nutshell, the results presented here demonstrate the observational approach to quantifying aerosol–cloud interactions.Geoscience and Remote Sensin
