Conference Papers

Permanent URI for this collectionhttps://idr.nitk.ac.in/handle/123456789/28506

Browse

Author: search.filters.author.Rudra, B.Subject: search.filters.subject.Machine Learning

Search Results

Now showing 1 - 3 of 3
  • No Thumbnail Available
    Item
    DNS tunneling detection using machine learning and cache miss properties
    (Institute of Electrical and Electronics Engineers Inc., 2021) Chowdhary, A.; Bhowmik, M.; Rudra, B.
    In a DNS Tunneling attack, data or other useful information is embedded within a DNS query and exfiltrated. Such attacks are difficult to detect because DNS is a fundamental protocol and blocking legitimate domain names can lead to an unpleasant experience for the users. Thus, detecting whether the DNS query is exfiltrating data or not is a challenging task. Mimicking genuine queries by the attacker makes this even more difficult. This research work presents two different methods for detecting the DNS Tunneling query and later they are combined to build a DNS Tunneling Attack Detector that can inform the client about a potential attack going on in real time. The first method uses cache misses in a DNS cache server and the second method utilizes machine learning techniques to classify a given DNS query. Overall, with around 93% accuracy of certain Machine Learning classifiers on classifying on a per packet basis along with extra validation from the cache-miss approach, a detector has been developed to accurately report DNS tunneling traffic © 2021 IEEE.
  • No Thumbnail Available
    Item
    An Intelligent Decision Support System for Bid Prediction of Undervalued Football Players
    (Institute of Electrical and Electronics Engineers Inc., 2022) Datta, M.; Rudra, B.
    The process of selecting football team players will determines a team's performance. An effective team is made up of a successful group of individual talented players. In general, a football team player selection is a decision made by the club based on the best available information. Club managers and scouts travel to different countries to watch matches and hire the best talent that can help their club to perform better. But for the lower leagues, it becomes difficult to hire the same talents because of strict budget. Here we devise a method so that we can leverage the undervalued players to get selected by the clubs. Clearly the benefit will be in two fold. First, the smaller clubs can get better players at an affordable cost. Second, the bigger clubs can get same performance players at a lower price helping them in cost cutting. We employ novelty detection methods to find out the undervalued players from our data and investigate our method by using five machine learning models. For performance evaluation, the five machine learning models used are support vector machine, Random Forest, Decision Tree, Linear Regression and XGBoost. Here XGboost performed best both for 10 fold cross-validation and external testing with a RMSE of 0.0122 and 0.0107 respectively. © 2022 IEEE.
  • No Thumbnail Available
    Item
    Binarization in DeepFake Audio Detection: A Comparative Study and Performance Analysis
    (Institute of Electrical and Electronics Engineers Inc., 2025) Gowhar, S.; Pandey, A.; Rudra, B.
    DeepFake audio, generated through advanced AI techniques, poses significant risks such as fraud, misinformation, and identity theft. As the quality of synthetic audio improves, detecting such fakes has become increasingly challenging. Traditional detection methods struggle to keep pace as AI-generated voices replicate speech patterns, tone, and pitch convincingly. While computationally intensive large-scale models can help detect DeepFakes generated by AI, their resource requirements make them impractical for deployment on mobile devices as well as on resource-constrained devices. This paper proposes a lightweight yet effective approach using binarized neural networks (BNNs) and further enhancements using additional dense layers and stacked modeling to overcome these challenges. We conduct a comprehensive performance analysis of the network and compare it with various machine learning and neural network methods to evaluate the tradeoff between detection accuracy and computational efficiency as an effect of binarization and precision loss in feature embeddings. © 2025 IEEE.