ABHISHEK NARAYANAN | PES University (original) (raw)

Uploads

Papers by ABHISHEK NARAYANAN

Research paper thumbnail of Towards Open Ended and Free Form Visual Question Answering: Modeling VQA as a Factoid Question Answering Problem

Research paper thumbnail of Document Embedding Generation for Cyber-Aggressive Comment Detection using Supervised Machine Learning Approach

Cyber-bullying may be defined as the employment of technological means for the purpose of harassi... more Cyber-bullying may be defined as the employment of technological means for the purpose of harassing, threatening, embarrassing, or targeting a particular person. It is also possible for Cyber-bullying to have occurred accidentally. One of the major challenges in identifying cyber-bullying or cyber-aggressive comments is to detect a sender’s tone in a particular text message, email or comments on social media, since what a person may consider to be a joke, may act as a hurting insult to another. Nevertheless, cyber-bullying may prove to be non-accidental in specific cases where a repetition in the pattern of text in emails, messages, and online posts is existent. In order to curb such a social threat, this Paper proposes the usage of a combination of document embeddings along with different supervised machine learning algorithms to get optimized results in flagging cyber-aggressive comments. Extensive experimentation indicates that the SVM model with rbf kernel combined with document...

Research paper thumbnail of A Novel Algorithm for Deraining Videos Using Range-Domain Filtering with Spatio-Temporal Correspondence

2019 Fifth International Conference on Image Information Processing (ICIIP), 2019

Unconstrained environment has always been a challenge to handle in computer vision and outdoor vi... more Unconstrained environment has always been a challenge to handle in computer vision and outdoor video analytics. Not only the visual perception but also the algorithmic efficiency at outdoor environment vastly impacted due to adverse weather conditions like rain, fog, snow, etc. These weather conditions largely degrade video quality by inducing complex noise into video frames especially due to mutual motion between capture assembly/ camera and object/ scene of interest. This greatly hampers video analytics and computer vision in various real time applications such as intelligent transportation systems, smart surveillance etc. Hence, restoring clean video from adverse effect of environment is one of the most important challenge in video analytics. With the aim of tackling this problem this paper proposes an efficient method of deraining videos by applying bilateral filtering on time-sliced video frames. The proposed idea exploited the strength of range-domain filtering possessed by bi...

Research paper thumbnail of Recurrent Neural Network Architectures with Trained Document Embeddings for Flagging Cyber-Aggressive Comments on Social Media

The emancipation of communication systems implies that bullying is no longer limited to schoolyar... more The emancipation of communication systems implies that bullying is no longer limited to schoolyards or street corners. This harassment has also infiltrated victims’ homes, via email, cell phones, and social media websites, in the form of cyber-bullying. For those who suffer such harassment, the effects can be devastating. The aim of our experimentation therefore, is to create a model using natural language processing to automate the flagging of cyber-aggressive comments with greater efficiency than the state-of-the-art models. In this task, this Paper proposes the application of Recurrent Neural Network Architectures to effectively analyze and flag such cyber aggressive comments on social media. The semantic document embeddings learnt by Doc2Vec [1] were fed into three recurrent neural network architectures for sequence classification. Our extensive experimentation indicates that an LSTM model using document embeddings outperforms the state-of-the-art models using standard feature e...

Research paper thumbnail of IronSense: Towards the Identification of Fake User-Profiles on Twitter Using Machine Learning

With the rampant escalation in the usage of online social media, there has been an uncurbed upsur... more With the rampant escalation in the usage of online social media, there has been an uncurbed upsurge in the number of fake user profiles which have infiltrated social networks, and has become a formidable threat to cyber-security. It is imperative to identify such fake profiles at the earliest since such malevolent accounts are often exploited to perpetrate fraud activities, retrieve personal or confidential information from victims, spread false propaganda online or to threaten and bully victims, ensuring that their original identities remain camouflaged. Though such profiles often look realistically convincing, there exist patterns in their behavioural tendencies. In order to curb the existence of such online frauds and help users distinguish between real or possibly fake profiles on Twitter, this paper proposes the extraction of key features such as number of friends, followers, statuses, that these behavioural trends of fake and legitimate users, learnt by various machine learnin...

Research paper thumbnail of VQA as a factoid question answering problem: A novel approach for knowledge-aware and explainable visual question answering

Image and Vision Computing

Research paper thumbnail of Character Level Neural Architectures for Boosting Named Entity Recognition In Code Mixed Tweets

2020 International Conference on Emerging Trends in Information Technology and Engineering (ic-ETITE)

Research paper thumbnail of Towards Open Ended and Free Form Visual Question Answering: Modeling VQA as a Factoid Question Answering Problem

Research paper thumbnail of Document Embedding Generation for Cyber-Aggressive Comment Detection using Supervised Machine Learning Approach

Cyber-bullying may be defined as the employment of technological means for the purpose of harassi... more Cyber-bullying may be defined as the employment of technological means for the purpose of harassing, threatening, embarrassing, or targeting a particular person. It is also possible for Cyber-bullying to have occurred accidentally. One of the major challenges in identifying cyber-bullying or cyber-aggressive comments is to detect a sender’s tone in a particular text message, email or comments on social media, since what a person may consider to be a joke, may act as a hurting insult to another. Nevertheless, cyber-bullying may prove to be non-accidental in specific cases where a repetition in the pattern of text in emails, messages, and online posts is existent. In order to curb such a social threat, this Paper proposes the usage of a combination of document embeddings along with different supervised machine learning algorithms to get optimized results in flagging cyber-aggressive comments. Extensive experimentation indicates that the SVM model with rbf kernel combined with document...

Research paper thumbnail of A Novel Algorithm for Deraining Videos Using Range-Domain Filtering with Spatio-Temporal Correspondence

2019 Fifth International Conference on Image Information Processing (ICIIP), 2019

Unconstrained environment has always been a challenge to handle in computer vision and outdoor vi... more Unconstrained environment has always been a challenge to handle in computer vision and outdoor video analytics. Not only the visual perception but also the algorithmic efficiency at outdoor environment vastly impacted due to adverse weather conditions like rain, fog, snow, etc. These weather conditions largely degrade video quality by inducing complex noise into video frames especially due to mutual motion between capture assembly/ camera and object/ scene of interest. This greatly hampers video analytics and computer vision in various real time applications such as intelligent transportation systems, smart surveillance etc. Hence, restoring clean video from adverse effect of environment is one of the most important challenge in video analytics. With the aim of tackling this problem this paper proposes an efficient method of deraining videos by applying bilateral filtering on time-sliced video frames. The proposed idea exploited the strength of range-domain filtering possessed by bi...

Research paper thumbnail of Recurrent Neural Network Architectures with Trained Document Embeddings for Flagging Cyber-Aggressive Comments on Social Media

The emancipation of communication systems implies that bullying is no longer limited to schoolyar... more The emancipation of communication systems implies that bullying is no longer limited to schoolyards or street corners. This harassment has also infiltrated victims’ homes, via email, cell phones, and social media websites, in the form of cyber-bullying. For those who suffer such harassment, the effects can be devastating. The aim of our experimentation therefore, is to create a model using natural language processing to automate the flagging of cyber-aggressive comments with greater efficiency than the state-of-the-art models. In this task, this Paper proposes the application of Recurrent Neural Network Architectures to effectively analyze and flag such cyber aggressive comments on social media. The semantic document embeddings learnt by Doc2Vec [1] were fed into three recurrent neural network architectures for sequence classification. Our extensive experimentation indicates that an LSTM model using document embeddings outperforms the state-of-the-art models using standard feature e...

Research paper thumbnail of IronSense: Towards the Identification of Fake User-Profiles on Twitter Using Machine Learning

With the rampant escalation in the usage of online social media, there has been an uncurbed upsur... more With the rampant escalation in the usage of online social media, there has been an uncurbed upsurge in the number of fake user profiles which have infiltrated social networks, and has become a formidable threat to cyber-security. It is imperative to identify such fake profiles at the earliest since such malevolent accounts are often exploited to perpetrate fraud activities, retrieve personal or confidential information from victims, spread false propaganda online or to threaten and bully victims, ensuring that their original identities remain camouflaged. Though such profiles often look realistically convincing, there exist patterns in their behavioural tendencies. In order to curb the existence of such online frauds and help users distinguish between real or possibly fake profiles on Twitter, this paper proposes the extraction of key features such as number of friends, followers, statuses, that these behavioural trends of fake and legitimate users, learnt by various machine learnin...

Research paper thumbnail of VQA as a factoid question answering problem: A novel approach for knowledge-aware and explainable visual question answering

Image and Vision Computing

Research paper thumbnail of Character Level Neural Architectures for Boosting Named Entity Recognition In Code Mixed Tweets

2020 International Conference on Emerging Trends in Information Technology and Engineering (ic-ETITE)