NVNallarasan Venkatachalam

Nallarasan Venkatachalam

*This is an algorithmically generated author record based on publicly available information.

Nallarasan Venkatachalam has a total of 9 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Curiosity-Guided Graph Neural Skill Discovery for Sparse-Reward Goal-Oriented Dialogue Systems

    by Abhishek Verma, Nallarasan V, Balaraman Ravindran

    Goal-oriented chatbots often operate in sparse-reward settings, where meaningful feedback (task success or user satisfaction) arrives only after many dialogue turns. This makes exploration and policy learning challenging. We propose a novel hierarchical reinforcement learning (HRL) framework that performs automated skill discovery using curiosity-driven exploration and graph-based temporal clustering. Dialogue trajectories are represented as a dynamic graph with utterance embeddings as nodes and transition probabilities as edges. A Graph Neural Network (GNN) combined with curiosity-weighted clustering identifies reusable dialogue skills (options) such as greeting, intent clarification, information gathering, and confirmation. These options are integrated into a hierarchical policy where a high-level policy selects skills and a low-level policy generates utterances. Evaluated on MultiWOZ 2.2 and a custom customer service dataset, the proposed framework achieves up to 22% higher task success rate and 18% shorter dialogues compared to strong HRL and flat RL baselines. The discovered skills demonstrate strong cross-domain transferability, making the framework practical for scalable, adaptive goal-oriented dialogue systems.

  • Open Access

    Harnessing Generative AI in Education: From Theory to Real-World Impact

    by Abhishek Verma, Nallarasan V

    The rise of generative artificial intelligence (GenAI) is transforming the education industry. GenAI models, particularly large language models (LLMs), have emerged as powerful tools capable of driving innovation, improving efficiency, and delivering superior services for educational purposes. This paper provides an overview of GenAI for educational purposes, from theory to practice. We have developed a chatbot for summarizing dialogues. In our research work, we have used strategies like zero-shot, one-shot, and few-shot inferencing and also fine-tuned the FLAN-T5 model to serve our purpose of summarization for educational tasks using PEFT (Parameter-Efficient Fine-Tuning) techniques like LoRA (Low-Rank Adaptation) and Prompt Tuning. We have also utilized the technique of Reinforcement Learning with PPO (Proximal Policy Optimization) and PEFT to generate less toxic summaries. The model's performance is quantitatively evaluated using the ROUGE metric and toxicity evaluation metrics. The chatbot can summarize dialogues and is of immense interest to users in the real world. In our research work, our findings demonstrate significant improvements in summarization quality and toxicity reduction, contributing to the development of safer and more effective AI systems.

  • Open Access

    Sentence Classification Using Transfer Learning with BERT

    by Abhishek Verma, Nallarasan V

    The rise of machine learning is enhancing various products in the software industry in a variety of ways by applying a diverse set of algorithms. One such application is to verify whether a sentence is grammatically correct or not. We find the use of such applications in Writing Assistance tools, Language Learning platforms, Automated Scoring platforms, Content Moderation platforms, and Email platforms. In our research, we present Sentence Classification using Transfer Learning with BERT (Bidirectional Encoder Representations from Transformers). We leverage transfer learning with the Hugging Face Transformers library to fine-tune the BERT model, incorporating hyperparameter sweeps via Weights and Biases to optimize learning rate, batch size, and the number of epochs. The process involves downloading and preprocessing the dataset, tokenizing sentences with the BERT tokenizer, and preparing input data with special tokens, padding, and attention masks. The dataset is split into training and validation sets, and a custom training loop is implemented to fine-tune the model, logging performance metrics to Weights and Biases. Additionally, a test function is provided to predict the grammaticality of example sentences using the fine-tuned model.

  • Modern Real-World Applications Using Data Analytics and Machine Learning

    by Vijayakumar Ponnusamy, Nallarasan V., Rajasegar R. S., Arivazhagan N., Gouthaman P.


  • Open Access

    Data Privacy and Ethics in Data Analytics

    by Rajasegar R. S., Gouthaman P., Vijayakumar Ponnusamy, Arivazhagan N., Nallarasan V.


  • Jamming Attack Mitigation in CR-IoT Using Game Theory

    by Nallarasan Venkatachalam, Kottilingam Kottursamy

    Jamming attack is Cognitive radio Internet of thing network disables the spectrum sharing and reduces the spectrum utilization .detection and mitigation of such attacks is the main component in realizing cognitive radio-based spectrum sharing in cognitive radio-based internet of thing network this work proposes a game theory-based jamming attack mitigation strategy. The problem of jamming attack mitigation is modeled as a zero-sum game and solved by finding Nash equilibrium. The cognitive node which tries to share the spectrum pays the zero-sum game with the jamming attacker and selects the best strategy of selecting the best frequency without getting into the jamming attack. The result of the proposed mechanism proves that the gaming mechanism can tackle the jamming attack.


  • Open Access

    Apache Mahout based Book Recommendation System

    by Abhishek Verma, V Nallarasan

    E-Commerce websites plays an important role in an individual’s life as it serves as the medium for online shopping with a huge audience. With the commencement of the pandemic due to novel coronavirus, the involvement of E-Commerce websites for shopping has drastically increased or more precisely it remains as the only medium to shop. With the increasing demand for online shopping on E-Commerce websites, the role of the Recommendation System has also become vital as it accomplishes the goal to make Personalized Recommendation for users. In this paper, we set out Apache Mahout-based Book Recommendation System to help recommend books to users. With this paper, we have described our project that recommends books to users on the basis of the user’s prior experience of purchase. The platform utilizing this recommendation system is developed using Spring Framework as a part of our project. The dataset used in the process is taken from Kaggle. Dataset has ratings for various books given by users. As a part of the User-based Collaborative Filtering recommendation technique, Euclidean Distance Similarity is used as a similarity measure along with Nearest N User Neighborhood and Generic User-Based Recommender to give quality results as compared to the existing system. To get the best quality recommendation we have obtained an evaluation score of 0.5 for Euclidean Distance Similarity

  • Open Access

    Cognitive Radio Jamming Attack Detection Using an Autoencoder for CRIoT Network 

    by Nallarasan V, Kottilingam Kottursamy

    IoT network-connected devices will be kept on increasing and will cross million, but it is impossible to allocate spectrum for those million and million of the devices. This spectrum scarcity can be handled by incorporating cognitive radio-based dynamic spectrum sharing, which is referred to as Cognitive Radio Internet of Things (CRIoT). But CRIoT sufferers from the Physical layer attack in cognitive radio, which affects the spectrum sensing accuracy and reduces the spectrum utilization. There are various attacks at the physical layer of cognitive radio among jamming attacks resulting in a denial of cognitive radio services and make spectrum underutilization. Continuous jamming can be detected quickly based on time delay on spectrum access, but discrete random jamming detection is challenging. This article proposes an autoencoder deep learning architecture-based jamming attack detection in cognitive radio. The jamming detection problem is modeled as anomaly detection. The autoencoder architecture is used to detect the jammer anomaly of the jammer. The proposed system involves the simulation of a random jamming attack and detecting it at a particular time instant information that may help mitigate the jammer attack .the proposed mechanism able to detect the jammer with 89% of accuracy.