Anurag Dutta
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Anurag Dutta has a total of 62 co-authors. Here are the top co-authors by number of shared publications:
- Athilingam Ramamoorthy
Department of Mathematics, Vel Tech Rangarajan Dr. Sagunthala R & D Institute of Science and Technology, Avadi, Chennai 600062, Tamil Nadu, India
13 shared publications - Pijush Kanti Kumar
Department of Information Technology, Government College of Engineering and Textile Technology, Calcutta, India
12 shared publications - Lakshmanan Kumarasankaralingam
Department of Mathematics, Kuwait American School of Education, Salmiya, Kuwait
9 shared publications - M. Gayathri Lakshmi
Department of Mathematics, Saveetha Engineering College, Chennai, India
7 shared publications - Rajat Subhra Chakraborty
Department of Computer Science and Engineering, Indian Institute of Technology Kharagpur, Kharagpur, India
7 shared publications Manan Roy Choudhury
Department of Computer Science and Engineering, Government College of Engineering and Textile Technology, Serampore, India
5 shared publications
Latest Publications
- Open Access
- 9 November 2026
- Published by Association for Computing Machinery (ACM)
- p. 60-68
Evaluating the Trust and Provenance Guarantees of Multimedia Authentication Codes: a Visualization-Driven Deep Cryptanalytic Assessment of SipHash
- 9 November 2026
- Published by Association for Computing Machinery (ACM)
- p. 60-68
Open AccessEvaluating the Trust and Provenance Guarantees of Multimedia Authentication Codes: a Visualization-Driven Deep Cryptanalytic Assessment of SipHash
- 1 November 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Biometrics, Behavior, and Identity Science
- Vol. 8 (6) , 1035-1047
Kol-3-Freq: Frequency-Aware Video Deepfake Detection Using Kolmogorov-Arnold Networks
- 1 November 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Biometrics, Behavior, and Identity Science
- Vol. 8 (6) , 1035-1047
Kol-3-Freq: Frequency-Aware Video Deepfake Detection Using Kolmogorov-Arnold Networks
- Open Access
- 1 September 2026
- Published by MDPI AG in AppliedMath
- Vol. 6 (9) , 143
Bench of Euler: A Benchmark for Evaluating the Problem-Solving Abilities of Large Language Models
Large language models (LLMs) have recently improved their problem-solving abilities and can solve complex mathematical problems with an increasing accuracy, necessitating the development of more challenging benchmarks. Over the years, the performance of LLMs on several benchmark datasets has also improved, motivating the development of evaluation frameworks that emphasize deeper algorithmic reasoning. This research presents Bench of Euler, a benchmark comprising 954 challenging mathematical and computational puzzles from the “Project Euler” collection, designed to evaluate complex multi-step mathematical reasoning. Solving these puzzles requires in-depth algorithmic deduction, as solutions based on brute force enumeration are mathematically infeasible. The benchmark is further classified into bands with respect to the problem difficulty to enable structured evaluation. Several existing open-source and proprietary, as well as base and reasoning, models have been evaluated on Bench of Euler, among which the highest-performing gpt-4o model by OpenAI achieved the highest accuracy of 43.21%. To evaluate the algorithmic efficiency, the models were prompted to generate Python code, which was subsequently executed in an isolated sandbox with strict timeout constraints. A failure analysis of the top-performing models revealed their inability to effectively optimize solutions, resulting in time-limit exceeded (TLE) errors during execution and outputs that often remained incomplete or approximate. Additionally, second-pass evaluations yielded an average performance improvement of approximately 2.09%, indicating limited gains from a second stochastic sample under the adopted decoding protocol.
- 1 September 2026
- Published by MDPI AG in AppliedMath
- Vol. 6 (9) , 143
Open AccessBench of Euler: A Benchmark for Evaluating the Problem-Solving Abilities of Large Language Models
Large language models (LLMs) have recently improved their problem-solving abilities and can solve complex mathematical problems with an increasing accuracy, necessitating the development of more challenging benchmarks. Over the years, the performance of LLMs on several benchmark datasets has also improved, motivating the development of evaluation frameworks that emphasize deeper algorithmic reasoning. This research presents Bench of Euler, a benchmark comprising 954 challenging mathematical and computational puzzles from the “Project Euler” collection, designed to evaluate complex multi-step mathematical reasoning. Solving these puzzles requires in-depth algorithmic deduction, as solutions based on brute force enumeration are mathematically infeasible. The benchmark is further classified into bands with respect to the problem difficulty to enable structured evaluation. Several existing open-source and proprietary, as well as base and reasoning, models have been evaluated on Bench of Euler, among which the highest-performing gpt-4o model by OpenAI achieved the highest accuracy of 43.21%. To evaluate the algorithmic efficiency, the models were prompted to generate Python code, which was subsequently executed in an isolated sandbox with strict timeout constraints. A failure analysis of the top-performing models revealed their inability to effectively optimize solutions, resulting in time-limit exceeded (TLE) errors during execution and outputs that often remained incomplete or approximate. Additionally, second-pass evaluations yielded an average performance improvement of approximately 2.09%, indicating limited gains from a second stochastic sample under the adopted decoding protocol.
- 12 July 2026
- Published by Springer Nature in Iran Journal of Computer Science
- Vol. 9 (1) , 43
IMO-CoT: a benchmark from International Mathematics Olympiads for evaluating chain-of-thought reasoning in large language models
- 12 July 2026
- Published by Springer Nature in Iran Journal of Computer Science
- Vol. 9 (1) , 43
IMO-CoT: a benchmark from International Mathematics Olympiads for evaluating chain-of-thought reasoning in large language models
- 9 March 2026
- Published by Springer Nature in Journal of Computational Social Science
- Vol. 9 (2) , 34
C4Censor: A Lightweight Benchmark dataset for Inappropriate Content Detection
- 9 March 2026
- Published by Springer Nature in Journal of Computational Social Science
- Vol. 9 (2) , 34
C4Censor: A Lightweight Benchmark dataset for Inappropriate Content Detection
- 27 October 2025
- Published by Springer Nature in Iran Journal of Computer Science
- Vol. 8 (4) , 2893-2908
Radicalized family of hyperbolic secant (RadSech) as an activation function
- 27 October 2025
- Published by Springer Nature in Iran Journal of Computer Science
- Vol. 8 (4) , 2893-2908
Radicalized family of hyperbolic secant (RadSech) as an activation function
- 24 October 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Network and Service Management
- Vol. 23, 1076-1085
RISK-4-Auto: Residually Interconnected and Superimposed Kolmogorov-Arnold Networks for Automotive Network Traffic Classification
In modern automobiles, a Controller Area Network (CAN) bus facilitates communication among all electronic control units for critical safety functions, including steering, braking, and fuel injection. However, due to the lack of security features, it may be vulnerable to malicious bus traffic-based attacks that cause the automobile to malfunction. Such malicious bus traffic can be the result of either external fabricated messages or direct injection through the on-board diagnostic port, highlighting the need for an effective intrusion detection system to efficiently identify suspicious network flows and potential intrusions. This work introduces Residually Interconnected and Superimposed Kolmogorov-Arnold Networks (RISK-4-Auto), a set of four deep neural network architectures for intrusion detection targeting in-vehicle network traffic classification. RISK-4-Auto models, when applied on three hexadecimally identifiable sequence-based open-source datasets (collected through direct injection in the on-board diagnostic port), outperform six state-of-the-art vehicular network intrusion detection systems (as per their accuracies) by ≈1.0163% for all-class classification and ≈2.5535% on focused (single-class) malicious flow detection. Additionally, RISK-4-Auto enjoys a significantly lower overhead than existing state-of-the-art models, and is suitable for real-time deployment in resource-constrained automotive environments.
- 24 October 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Network and Service Management
- Vol. 23, 1076-1085
RISK-4-Auto: Residually Interconnected and Superimposed Kolmogorov-Arnold Networks for Automotive Network Traffic Classification
In modern automobiles, a Controller Area Network (CAN) bus facilitates communication among all electronic control units for critical safety functions, including steering, braking, and fuel injection. However, due to the lack of security features, it may be vulnerable to malicious bus traffic-based attacks that cause the automobile to malfunction. Such malicious bus traffic can be the result of either external fabricated messages or direct injection through the on-board diagnostic port, highlighting the need for an effective intrusion detection system to efficiently identify suspicious network flows and potential intrusions. This work introduces Residually Interconnected and Superimposed Kolmogorov-Arnold Networks (RISK-4-Auto), a set of four deep neural network architectures for intrusion detection targeting in-vehicle network traffic classification. RISK-4-Auto models, when applied on three hexadecimally identifiable sequence-based open-source datasets (collected through direct injection in the on-board diagnostic port), outperform six state-of-the-art vehicular network intrusion detection systems (as per their accuracies) by ≈1.0163% for all-class classification and ≈2.5535% on focused (single-class) malicious flow detection. Additionally, RISK-4-Auto enjoys a significantly lower overhead than existing state-of-the-art models, and is suitable for real-time deployment in resource-constrained automotive environments.