Browsing by Author "Anand Kumar, M."

Now showing 1 - 20 of 99

A comparative analysis of machine comprehension using deep learning models in code-mixed hindi language
(Springer Verlag service@springer.de, 2019) Viswanathan, S.; Anand Kumar, M.; Padannayil, K.P.
The domain of artificial intelligence revolutionizes the way in which humans interact with machines. Machine comprehension is one of the latest fields under natural language processing that holds the capability for huge improvement in artificial intelligence. Machine comprehension technique gives systems the ability to understand a passage given by user and answer questions asked from it, which is an evolved version of traditional question answering technique. Machine comprehension is a main technique that falls under the category of natural language understanding, which exposes the amount of understanding required for a model to find the area of interest from a passage. The scope for the implementation of this technique is very high in India due to the availability of different regional languages. This work focused on the incorporation of machine comprehension technique in code-mixed Hindi language. A detailed comparison study on the performance of dataset in several deep learning approaches including End to End Memory Network, Dynamic Memory Network, Recurrent Neural Network, Long Short-Term Memory Network and Gated Recurrent Unit are evaluated. The best suited model for the dataset used is identified from the comparison study. A new architecture is proposed in this work by combining two of the best performing networks. To improve the model with respect to various ways of answering questions from a passage the natural language processing technique of distributed word representation was performed on the best model identified. The model was improved by applying pre-trained fastText embeddings for word representations. This is the first implementation of machine comprehension models in code-mixed Hindi language using deep neural networks. The work analyses the performance of all five models implemented, which will be helpful for future researches on Machine Comprehension technique in code-mixed Indian languages. © Springer Nature Switzerland AG 2019.
A reasoning based explainable multimodal fake news detection for low resource language using large language models and transformers
(Springer Nature, 2025) LekshmiAmmal, H.R.; Anand Kumar, M.
Nowadays, individuals rely predominantly on online social media platforms, news feeds, websites, and news aggregator applications to acquire recent news stories. This trend has resulted in an increase in the number of available social media platforms, online news feeds, and news aggregator applications. These news platforms have been accused of spreading fake news to gain more attention and recognition. Earlier, this misinformation or fake news used to be propagated only in the text form. However, with the advent of technology, now it is spread in multimodal forms, such as images with text, videos, and audio with textual content. Currently, the automatic fake news detection models are focused on high resource languages and superficial output. Social media users need clarity and reasoning when it comes to identifying fake news, rather than just a superficial classification of news as fake. Providing context, reasoning, and explanations can help users understand why certain news is misleading or false. Hence, a multimodal system has to be developed to identify and justify fake news. In this proposed work, we have developed a multimodal fake news system for the Low Resource Language Tamil with reasoning-based explainability. The dataset for this proposed work is retrieved from fact-check websites and official news websites. We have experimented with different combinations of models for visual and text modalities. Further, we integrated LLM-based image descriptions into our model with the text and visual features, resulting in an F1 score of 0.8736. We used the Siamese model to determine the similarity of the news and its image descriptions. Additionally, we conducted error analysis and used explainable artificial intelligence to explore the reasoning behind our model’s predictions. We also present the textual reasoning for the model’s predictions and match them with images. © The Author(s) 2025.
A Study of Machine Translation Models for Kannada-Tulu
(Springer Science and Business Media Deutschland GmbH, 2023) Hegde, A.; Shashirekha, H.L.; Anand Kumar, M.; Chakravarthi, B.R.
Over the past ten years, neural machine translation (NMT) has seen tremendous growth and is now entering a phase of maturity. Despite being the most popular solution for machine translation (MT), it performs sub-optimally on under-resourced language pairs due to lack of parallel corpora as compared to high-resourced language pairs. The implementation of NMT techniques for under-resourced language pairs is receiving the attention of researchers and has resulted in a significant amount of research for many under-resourced language pairs. In view of the growth of MT, this paper describes a set of practical approaches for investigating MT between Kannada and Tulu. These two languages belong to the family of Dravidian languages and are under-resourced due to lack of tools and resources particularly the parallel corpus for MT. Since there are no parallel corpora for the Kannada-Tulu language pair for MT, this work aims to construct a parallel corpus for this language pair. As manual construction of parallel corpus is laborious, data augmentation is introduced to enhance the size of the parallel corpus along with suitable preprocessing techniques. Different NMT schemes such as recurrent neural network (RNN) baseline, bidirectional recurrent neural network (BiRNN), transformer-based NMT with and without subword tokenization, and statistical machine translation (SMT) models are implemented for MT of Kannada-Tulu and Tulu-Kannada language pairs. Empirical results reveal that the impact of data augmentation increases the bilingual evaluation understudy (BLEU) score of the proposed models. Transformer-based models with subword tokenization outperformed the other models with BLEU scores 41.82 and 40.91 for Kannada-Tulu and Tulu-Kannada MT, respectively. Â© 2023, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.
A transformer-based architecture for fake news classification
(Springer, 2021) Mehta, D.; Dwivedi, A.; Patra, A.; Anand Kumar, M.
In today’s post-truth world, the proliferation of propaganda and falsified news poses a deadly risk of misinforming the public on a variety of issues, either through traditional media or on social media. Information people acquire through these articles and posts tends to shape their world view and provides reasoning for choices they take in their day to day lives. Thus, fake news can definitely be a malicious force, having massive real-world consequences. In this paper, we focus on classifying fake news using models based on a natural language processing framework, Bidirectional Encoder Representations from Transformers, also known as BERT. We fine-tune BERT for specific domain datasets and also make use of human justification and metadata for added performance in our models. We determine that the deep-contextualizing nature of BERT is effective for this task and obtain significant improvement over binary classification, and minimal yet important improvement in six-label classification in comparison with previously explored models. © 2021, The Author(s), under exclusive licence to Springer-Verlag GmbH Austria, part of Springer Nature.
Advancements in Financial Document Structure Extraction: Insights from Five Years of FinTOC (2019-2023)
(Institute of Electrical and Electronics Engineers Inc., 2023) Kang, J.; Patel, M.M.; Agrawal, A.; Simhadri, S.; Srinivasa, R.; Bellato, S.; Anand Kumar, M.; Tsang, N.D.; El-Haj, M.
In this comprehensive paper, we present a detailed overview of the Financial Table Of Content extraction shared task series, FinTOC, conducted over a span of five years from 2019 to 2023. This paper serves as a retrospective analysis of the key developments in the field of financial document structure extraction. The FinTOC series, hosted within the framework of the Financial Narrative Processing (FNP) workshop, has been instrumental in shaping the landscape of Natural Language Processing (NLP) in the financial domain. Our analysis delves into the diverse methodologies proposed by participants across all editions, shedding light on the innovative strategies employed to tackle the intricate challenge of extracting structured information from financial documents. We explore the evolution of techniques, from traditional rule-based approaches to cutting-edge deep learning models, showcasing the dynamic nature of NLP advancements. Furthermore, our study investigates the introduction of multilingual datasets by the organizers, highlighting the importance of cross-lingual analysis in financial document processing. We also examine the contributions made by participants in augmenting the training data with external sources, showcasing the collaborative spirit of the NLP community in enhancing the quality and size of the shared training dataset. Â© 2023 IEEE.
An Approach for Waste Classification Using Data Augmentation and Transfer Learning Models
(Springer Science and Business Media Deutschland GmbH, 2023) Kumsetty, N.V.; Bhat Nekkare, A.B.; Kamath Sâ€¤, S.; Anand Kumar, M.
Waste segregation has become a daunting problem in the twenty-first century, as careless waste disposal manifests significant ecological and health concerns. Existing approaches to waste disposal primarily rely on incineration or land filling, neither of which are sustainable. Hence, responsible recycling and then adequate disposal is the optimal solution promoting both environment-friendly practices and reuse. In this paper, a computer vision-based approach for automated waste classification across multiple classes of waste products is proposed. We focus on improving the quality of existing datasets using data augmentation and image processing techniques. We also experiment with transfer learning based models such as ResNet and VGG for fast and accurate classification. The models were trained, validated, and tested on the benchmark TrashNet and TACO datasets. During experimental evaluation, the proposed model achieved 93.13% accuracy on TrashNet and outperformed state-of-the-art models by a margin of 16% on TACO. Â© 2023, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.
An Effective Diabetic Retinopathy Detection Using Hybrid Convolutional Neural Network Models
(Springer Science and Business Media Deutschland GmbH, 2023) Kumar, N.; Ahmed, R.; Venkatesh, B.H.; Anand Kumar, M.
Loss of vision in the present era of the developing world is mainly caused by diabetic retinopathy. More than 103 million people are believed to be affected. It is estimated that around 40 million beings have diabetes in the United States, and according to the World Health Organization (WHO), 347 million people are living with the disease globally. Diabetic retinopathy (DR) is a long-term diabetes-related eye condition. Roughly, 45–50% of the American citizens suffering from diabetes undergo some unique stages that can be categorized. When DR is diagnosed on a timely basis, the possibility of it extending to the course of vision impairment can be delayed and stopped, though this is not entirely true and a very daunting task because it seldom reveals any symptom before it escalates to a stage of no return to effectively treat it. The paper uses convolutional neural network models to achieve an effective classification for diabetic detection of retinal fundus images. © 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.
Aspect Based Neural Recommender Using Adaptive Prediction
(Institute of Electrical and Electronics Engineers Inc., 2023) Bhojwani, A.; Jolly, V.; Goel, S.; Anand Kumar, M.
Recommendation systems are information processing systems that analyze user behavior and make suggestions relevant to the users' interests. These systems recommend movies, products etc., based on many different factors. These recommended items are the items which are most likely in the interest of the users. This work focuses on aspects of users' textual comments to make recommendations that are relevant to users. We propose a model that recommends the items based on extracted aspects, taking the reviews and information from the items. Based on the item's aspect level importance, a user may rate them high or low, and we have modelled our recommendation system considering the same. This calculation is based on the neural co-attention mechanism. The proposed model points to various shortcomings of the model that were introduced before. We have used the datasets from amazon, which are extracted using web scraping. The existing recommendation system predicts ratings alone, while the proposed model gives rankings and ratings. The model's efficiency is checked using precision, recall and F1 score. Â© 2023 IEEE.
Automated Traffic Light Signal Violation Detection System Using Convolutional Neural Network
(Springer, 2020) Bordia, B.; Nishanth, N.; Patel, S.; Anand Kumar, M.; Rudra, B.
Automated traffic light violation detection system relies on the detection of traffic light color from the video captured with the CCTV camera, detection of the white safety line before the traffic signal and vehicles. Detection of the vehicles crossing traffic signals is generally done with the help of sensors which get triggered when the traffic signal turns red or yellow. Sometimes, these sensors get triggered even when the person crosses the line or some animal crossover or because of some bad weather that gives false results. In this paper, we present a software which will work on image processing and convolutional neural network to detect the traffic signals, vehicles and the white safety line present in front of the traffic signals. We present an efficient way to detect the white safety line in this paper combined with the detection of traffic lights trained on the Bosch dataset and vehicle detection using the TensorFlow object detection SSD model. Â© 2020, Springer Nature Singapore Pte Ltd.
Bot and gender identification from twitter notebook for PAN at CLEF 2019
(CEUR-WS ceurws@sunsite.informatik.rwth-aachen.de, 2019) Radarapu, R.; Vishwakarma, Y.; Sai Gopal, A.S.; Anand Kumar, M.
The popularity of social media raises a concern about the quality of content over its platforms. The quality of data is important, especially for fair and considerable predictive analysis. If the quality of data is less, it may result in the prediction of wrong circumstances of an event. This causes misleading trending problems and more importantly, the sensitive stock price may fluctuate. The contents available on social media can be corrupted and overflowed by bots. There are a variety of bots available such as Spam Bots, Influence Bots, etc. Our target is to identify such bots on Twitter. Twitter data is mostly used by data analysts for applications related to scientific predictions or opinion analysis. This working note is capitalized on earlier approaches and Machine Learning (ML) approaches used to classify between a bot and human and find the gender further for interesting studies in crime detection etc. By sharing many attributes for user profiles, we have identified the pattern to find out that the given user is a bot or human based on the tweets posted. Â© 2019 for this paper by its authors. Use permitted under Creative Commons License Attribution 4.0 International (CC BY 4.0). CLEF 2019, 9-12 September 2019, Lugano, Switzerland.
Building Intelligent Embodied AI Agents for Asking Clarifying Questions
(Institute of Electrical and Electronics Engineers Inc., 2023) Lakhtaria, D.; Chhabra, R.; Taparia, R.; Anand Kumar, M.
Interactive embodied agents are intelligent agents present inside some virtual environment and interact with humans to do tasks we want them to simulate. The communication takes place in some verbal or non-verbal form. This kind of communication gives a better and richer experience and more possibility for users to be satisfied with the carried-out task. An intelligent agent analyzes the environment it is in and performs tasks with all its knowledge based on the interaction. The task is carried out in the game environment where the Agent resides. There are some challenges to an agent that it needs to address, as well as to understand the instruction given to it. The next challenge is asking for clarification if the instruction given is unclear. In this research paper, we have discussed the problems mentioned above using the dataset provided by the NeurIPS 2022 IGLU Challenge which is a labeled dataset. We have discussed the various approaches for the problem mentioned. Â© 2023 IEEE.
Cardiovascular Disease Prediction Using Machine Learning
(Institute of Electrical and Electronics Engineers Inc., 2022) Prajwal, K.; Tharun, K.; Navaneeth, P.; Anand Kumar, M.
As the human population increases, so is the chance of getting diseases. There are many illnesses globally, and one of the biggest problems faced by the hospital systems today is the lack of technology to know when the patients are ill. One such illness is Cardiovascular Disease or CVD. It refers to any heart disease, vascular disease, or blood vessel disease. According to WHO, more people die of CVD's worldwide than any other cause. It affects the low and middle-income countries more. It is very hard for people living alone to contact the hospital when they are sick. Therefore, we have developed a model that can detect when a patient is ill and report back to the hospital. The system currently only identifies patients with heart disease and reports back to the hospital. We decided to go with heart disease identification because it is one of the most deadly diseases, and the risk of patients dying because of heart disease is high. Predicting whether a patient has heart disease or not is very clearly a classification problem. Therefore, we have used five models to classify. We take several factors such as blood sugar level, age, cholesterol level, and many more and give the outcome based on the input. Â© 2022 IEEE.
Categorizing Relations via Semi-supervised Learning Using a Hybrid Tolerance Rough Sets and Genetic Algorithm Approach
(Springer Science and Business Media Deutschland GmbH, 2022) Agrawal, S.; Ahmed, R.; Anand Kumar, M.; Ramanna, S.
In the last few decades, we have seen a tremendous increase in the amount of data available on the web. There have been significant advances in constructing knowledge bases consisting of relations from the text data. These relations are words in the text often represented as pairs (Noun, Context), for example (Disease, Symptom), which can be classified into some predefined category to give us some useful information. Categorization of relations using tolerance-rough set based semi-supervised learning algorithm (TPL) have been successfully demonstrated in several works. However, an unexplored problem is the automatic selection of hyper parameters of the TPL algorithm. This paper proposes a genetic algorithm-based approach (TPL-GA) for optimizing the hyper-parameters that are fundamental to the TPL algorithm. The proposed approach was tested on two standard datasets drawn from different domains representing two different languages: English and Hindi text. © 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.
CISER: Customized Institute Specific Search Engine for Retrieving Research Papers
(Institute of Electrical and Electronics Engineers Inc., 2023) Sankar, S.; Muslihuddeen, H.; Ostwal, S.; Sathvika, P.; Anand Kumar, M.
This paper proposes a methodology of a search engine system for searching research papers customized to our institute students. Most of the courses are associated with course projects where students face difficulties in finding the best research papers associated with the course. So, here we propose a customized mechanism to search the research papers published by the faculties of the institute. The input for the proposed search engine can either be the course name or the topic itself. We give users two options: search by course name and topic. If the course name is given as input, we get the corresponding keywords for the course, and then we implement semantic similarity on the Author Keywords. If the user searches by topic, we perform semantic similarity using the given topic and the Author Keywords of the research papers. We have also created a web interface using Django. Â© 2023 IEEE.
Citation Intent Classification Using Transformers
(Institute of Electrical and Electronics Engineers Inc., 2024) Rakshith Gowda, H.C.; Raj, K.S.; Anand Kumar, M.
As the world of scholarly research continues to grow, the intricate network of citations serves as the foundation of academic discussion, symbolizing the interweaving of concepts and the dissemination of information. The study of citations in scientific literature is important for discovering new knowledge, retrieving information, and analyzing discourse. However, manually categorizing citation functions is a slow and biased process. To address this, we conducted research on automated citation function classification in astrophysics literature by creating and evaluating deep learning models. We also introduce the FOCAL dataset, which stands for Functions of Citations in Astrophysics Literature, includes astrophysics articles with manually labelled citation functions. Our approach uses language features, citation contexts, and domain knowledge to classify citation functions. Results show that our method accurately identifies citation functions, indicating its potential for improving citation analysis. Â© 2024 IEEE.
Classification of Censored Tweets in Chinese Language using XLNet
(Association for Computational Linguistics (ACL), 2021) Ahmed, S.S.; Anand Kumar, M.
In the growth of todayâ€™s world and advanced technology, social media networks play a significant role in impacting human lives. Censorship is the overthrowing of speech, public transmission, or other details that play a vast role in social media. The content may be considered harmful, sensitive, or inconvenient. Authorities like institutes, governments, and other organizations conduct Censorship. This paper has implemented a model that helps classify censored and uncensored tweets as a binary classification. The paper describes submission to the Censorship shared task of the NLP4IF 2021 workshop. We used various transformer-based pre-trained models, and XLNet outputs a better accuracy among all. We fine-tuned the model for better performance and achieved a reasonable accuracy, and calculated other performance metrics. Â© 2021 Association for Computational Linguistics.
Classifying Emotional States Through EEG-Derived Spectrograms
(Institute of Electrical and Electronics Engineers Inc., 2024) Mahit Nandan, A.D.; Dhiraj Choudhary, D.; Godbole, I.; Anand Kumar, M.
Identifying emotional state of a person plays an important role in a multitude of applications such as affective computing, human-computer interaction and most importantly healthcare. Understanding and correctly identifying human emotions can improve mental health assessments, making it possible to obtain personalized treatment plans and increase the user experience across a variety of digital applications. Our work explores the feasibility of emotion classification using EEG signals recorded from multiple users while experiencing various emotions. Working with time series data of the brain activity in various participants, who are experiencing particular emotions over an interval of 15 seconds. To train and evaluate the classification model several kinds of machine learning and deep learning models such as CNN, RNN and LSTM, are employed. We compare general sequential models with image processing models in the task of classifying signal data. Observing the model's ability to generalize across the population. Â© 2024 IEEE.
Clustering Enhanced Encoderâ€“Decoder Approach to Dimensionality Reduction and Encryption
(Springer Science and Business Media Deutschland GmbH info@springer-sbm.com, 2021) Mukesh, B.R.; Madhumitha, N.; Aditya, N.P.; Vivek, S.; Anand Kumar, M.
Dimensionality reduction refers to reducing the number of attributes that are being considered, by producing a set of principal variables. It can be divided into feature selection and feature extraction. Dimensionality reduction serves as one of the preliminary challenges in storage management and is useful for effective transmission over the Internet. In this paper, we propose a deep learning approach using encoderâ€“decoder networks for effective (almost-lossless) compression and encryption. The neural network essentially encrypts data into an encoded format which can only be decrypted using the corresponding decoders. Clustering is essential to reduce the variation in the dataset to ensure overfit. Using clustering resulted in a net gain of 1% over the standard encoder architecture over three MNIST datasets. The compression ratio achieved was 24.6:1. The usage of image datasets is for visualization only and the proposed pipeline could be applied for textual and visual data as well. Â© 2021, The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.
Cognitive Chromatic Image Synthesis Using UNET and GAN
(Institute of Electrical and Electronics Engineers Inc., 2024) Choubey, D.; Patil, V.; Anand Kumar, M.
In the last decade, there has been a lot of interest for image colorization over a wide range of applications, especially in the restoration of old or damaged images. Because there are a lot of options when it comes to assigning color information, this problem is ill-posed by nature and is quite difficult to solve. Researchers have handled this issue in a variety of imaginative ways. More recent developments in automated colorization are focused on images that are repetitive in nature or images that require extensive editing. For instance, in such settings, semantic maps can be used as additional input to offer better control over the generalization of the colorization task with the help of conditional Deep Convolutional Generative Adversarial Networks (DCGANs).Our solution combines the techniques to allow computers to produce vivid visuals in this way. Monochrome or black and white images most of the times differ from the colored images in terms of visual detail and image content and colorizing them by hand is a tedious and often an artistic task. Â© 2024 IEEE.
Context Sensitive Tamil Language Spellchecker Using RoBERTa
(Springer Science and Business Media Deutschland GmbH, 2023) Rajalakshmi, R.; Sharma, V.; Anand Kumar, M.
A spellchecker is a tool that helps to identify spelling errors in a piece of text and lists out the possible suggestions for that word. There are many spell-checkers available for languages such as English but a limited number of spell-checking tools are found for low-resource languages like Tamil. In this paper, we present an approach to develop a Tamil spell checker using the RoBERTa (xlm-roberta-base) model. We have also proposed an algorithm to generate the test dataset by introducing errors in a piece of text. The spellchecker finds out the mistake in a given text using a corpus of unique Tamil words collected from different sources such as Wikipedia and Tamil conversations, and lists out the suggestions that could be the potential contextual replacement of the misspelled word using the proposed model. On introducing a few errors in a piece of text collected from a Wikipedia article and testing it on our model, an accuracy of 91.14% was achieved for error detection. Contextually correct words were then suggested for these erroneous words detected. Our spellchecker performed better than some of the existing Tamil spellcheckers in terms of both higher accuracy and lower false positives. Â© 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.