ISI Digital Commons (Indian Statistical Institute )
Not a member yet
7571 research outputs found
Sort by
Agent-Driven BFS Tree in Anonymous Graphs with Applications
No full textBreadth-First-Search (BFS) trees serve a pivotal role in designing efficient graph algorithms due to their efficacy in traversing and exploring graph structures with a systematic layer-by-layer approach. This paper introduces an agent-based novel approach for constructing a BFS tree on an arbitrary anonymous graph G with n nodes and m edges using k≥n autonomous mobile agents. In this paper, we provide algorithms for BFS tree construction for different starting configurations and demonstrate their applications. Our main result considers the dispersed starting configuration (i.e., each node is occupied by a single agent at the start) and takes O(DΔ) rounds to execute, where D is the diameter and Δ is the highest degree of G. The algorithm assumes the knowledge of a root node for the BFS tree. We further continue our investigation for BFS tree construction with two other classical configurations, namely, rooted configuration and arbitrary configuration (with and without the knowledge of root) of the agents with some other follow-up configurations for k\u3en. In addition, the paper demonstrates the application of the BFS tree construction methodology in tasks like - checking a graph for bipartiteness and gathering agents into a single node, a fundamental task in distributed robotics
An Efficient Neural Network Controller for Autonomous Lane-Keeping Assist System
No full textWith the advent of new technologies and hardware, the importance and demand for autonomous vehicles are also increasing nowadays. For the safe deployment of these autonomous vehicles, efficient and reliable control techniques are very crucial. Classical control techniques encounter difficulties in accommodating various intricate real-world driving situations, hence, fail to properly control them. Moreover, since a major part of the underlying feedback control computation includes heavy vision computation tasks, classical control techniques are falling short in controlling them properly due to the delay coming from this lengthy vision computation. To overcome this, data-driven approaches are being used in many applications. This research work particularly addresses the challenge of controlling an autonomous lane-keeping assist system leveraging the power of neural networks (NN). The aim of the proposed NN-based controller is to enhance the precision and robustness of the lane-keeping assist system. The NN-based controller demonstrates promising results with respect to some state-of-the-art methodologies, offering hope for safer and more effective autonomous driving in dynamic environments
Blockchain-Enabled Distributed Payment Card Tokenization System
No full textPayment card tokenization is a process that replaces sensitive card (Credit/Debit) information with a unique identifier or token. This token is used for transaction processing, reducing the risk of unauthorized access and fraud. However, the current landscape is dominated by a few centralized entities, such as Visa, Mastercard, functioning as Token Service Providers (TSP). To address this, we propose a blockchain-powered decentralized tokenization system where a smart contract generates tokens based on predefined criteria. Decentralization ensures the system\u27s robustness against attacks and offers transparent audit trails. Our prototype is built on the Ethereum platform, with smart contracts deployed on the Sepolia test network, allowing for performance evaluation and validation of the system\u27s efficacy
Exploring Green AI for Audio Deepfake Detection
No full textThe state-of-the-art audio deepfake detectors leveraging deep neural networks exhibit impressive recognition performance. Nonetheless, this advantage is accompanied by a significant carbon footprint. This is mainly due to the use of high-performance computing with accelerators and high training time. Studies show that average deep NLP model produces around 626k lbs of CO2 which is equivalent to five times of average US car emission at its lifetime. This is certainly a massive threat to the environment. To tackle this challenge, this study presents a novel framework for audio deepfake detection that can be seamlessly trained using standard CPU resources. Our proposed framework utilizes off-the-shelve self-supervised learning (SSL) based models which are pre-trained and available in public repositories. In contrast to existing methods that fine-tune SSL models and employ additional deep neural networks for downstream tasks, we exploit classical machine learning algorithms such as logistic regression and shallow neural networks using the SSL embeddings extracted using the pre-trained model. Our approach shows competitive results compared to the commonly used high-carbon footprint approaches. In experiments with the ASVspoof 2019 LA dataset, we achieve a 0.90% equal error rate (EER) with less than 1k trained model parameters. To encourage further research in this direction and support reproducible results, the Python code will be made publicly accessible following acceptance
Knowing the class distinguishing abilities of the features, to build better decision-making models
No full textExplainability allows end-users to have a transparent and humane reckoning of an ML scheme\u27s capability and utility. ML model\u27s modus opernadi can be explained via the features which trained it. To this end, we found no work explaining the features\u27 importance based on their class-distinguishing abilities. In a given dataset, a feature is not equally good at distinguishing between the data points\u27 possible categorizations (or classes). This work explains the features based on their class or category-distinguishing capabilities. We estimate the variables\u27 class-distinguishing capabilities (scores) for pair-wise class combinations, utilize them in a missing feature context, and propose a novel decision-making protocol. A key novelty of this work lies in the refusal to render a decision option when the missing feature (of the test point) has a high class-distinguishing potential for the likely classes. Two real-world datasets are used empirically to validate the explainability of our scheme
Parking Problem by Oblivious Mobile Robots in Infinite Grids
No full textIn this paper, the parking problem of a swarm of mobile robots has been studied. The robots are deployed at the nodes of an infinite grid, which has a subset of prefixed nodes marked as parking nodes. Each parking node pi has a capacity of ki which is given as input and equals the maximum number of robots a parking node can accommodate. As a solution to the parking problem, robots need to partition themselves into groups so that each parking node contains a number of robots that are equal to the capacity of the node in the final configuration. It is assumed that the number of robots in the initial configuration represents the sum of the capacities of the parking nodes. The robots are assumed to be autonomous, anonymous, homogeneous, identical and oblivious. They operate under an asynchronous scheduler. They neither have any agreement on the coordinate axes nor do they agree on a common chirality. All the initial configurations for which the problem is unsolvable have been identified. A deterministic distributed algorithm has been proposed for the remaining configurations, ensuring the solvability of the problem
Testing exchangeability by pairwise betting
No full textIn this paper, we address the problem of testing exchangeability of a sequence of random variables, X1,X2,···. This problem has been studied under the recently popular framework of testing by betting. But the mapping of testing problems to game is not one to one: many games can be designed for the same test. Past work established that it is futile to play single game betting on every observation: test martingales in the data filtration are powerless. Two avenues have been explored to circumvent this impossibility: betting in a reduced filtration (wealth is a test martingale in a coarsened filtration), or playing many games in parallel (wealth is an e-process in the data filtration). The former has proved to be difficult to theoretically analyze, while the latter only works for binary or discrete observation spaces. Here, we introduce a different approach that circumvents both drawbacks. We design a new (yet simple) game in which we observe the data sequence in pairs. Even though betting on individual observations is futile, we show that betting on pairs of observations is not. To elaborate, we prove that our game leads to a nontrivial test martingale, which is interesting because it has been obtained by shrinking the filtration very slightly. We show that our test controls type-1 error despite continuous monitoring, and is consistent for both binary and continuous observations, under a broad class of alternatives. Due to the shrunk filtration, optional stopping is only allowed at even stopping times: a relatively minor price. We provide a variety of simulations that align with our theoretical findings
Tight Multi-user Security of Ascon and Its Large Key Extension
No full textThe Ascon cipher suite has recently become the preferred standard in the NIST Lightweight Cryptography standardization process. Despite its prominence, the initial dedicated security analysis for the Ascon mode was conducted quite recently. This analysis demonstrated that the Ascon AEAD mode offers superior security compared to the generic Duplex mode, but it was limited to a specific scenario: single-user nonce-respecting, with a capacity strictly larger than the key size. In this paper, we eliminate these constraints and provide a comprehensive security analysis of the Ascon AEAD mode in the multi-user setting, where the capacity need not be larger than the key size. Regarding data complexity D and time complexity T, our analysis reveals that Ascon achieves AEAD security when T is bounded by min{2κ/μ,2c} (where κ is the key size, and μ is the number of users), and DT is limited to 2b (with b denoting the size of the underlying permutation, set at 320 for Ascon). Our results align with NIST requirements, showing that Ascon allows for a tag size as small as 64 bits while supporting a higher rate of 192 bits, provided the number of users remains within recommended limits. However, this security becomes compromised as the number of users increases significantly. To address this issue, we propose a variant of the Ascon mode called LK-Ascon, which enables doubling the key size. This adjustment allows for a greater number of users without sacrificing security, while possibly offering additional resilience against quantum key recovery attacks. We establish tight bounds for LK-Ascon, and furthermore show that both Ascon and LK-Ascon maintain authenticity security even when facing nonce-misuse adversaries
Turbulence Features in a Wall-Wake Flow Downstream of Two Horizontal Cylinders – A Numerical Approach
No full textThis study presents experimental and numerical studies of flow past two horizontal cylinders. The study was performed using the experimental data that was recorded over a rough bed with bluff bodies. First, general features have been specified at various downstream distances from cylinders, including streamwise velocity, Reynolds shear stress, and turbulence intensity profiles. Eventually one numerical model has been implemented to replicate the experimental scenario and to perform simulation. Information regarding general turbulence properties along with dissipation rate have been extracted from simulation software and compared with that of experimental data measured at different locations, which are found to be fairly consistent
