Farzad Parvaresh
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Farzad Parvaresh has a total of 59 co-authors. Here are the top co-authors by number of shared publications:
- Maryam Khatami
Department of Pure Mathematics, Faculty of Mathematics and Statistics, University of Isfahan, Isfahan, Iran
7 shared publications - Reza Sobhani
Department of Applied Mathematics and Computer Science, Faculty of Mathematics and Statistics, University of Isfahan, Isfahan, Iran
7 shared publications - Alireza Abdollahi
Department of Pure Mathematics, Faculty of Mathematics and Statistics, University of Isfahan, Isfahan, Iran
6 shared publications - Javad Bagherian
Department of Pure Mathematics, Faculty of Mathematics and Statistics, University of Isfahan, Isfahan, Iran
5 shared publications - Mohamad Farzan Sabahi
Department of Electrical Engineering, University of Isfahan, Isfahan, Iran
5 shared publications - Fatemeh Jafari
Department of Pure Mathematics, Faculty of Mathematics and Statistics, University of Isfahan, Isfahan, Iran
4 shared publications
Latest Publications
- 29 May 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Information Theory
- Vol. PP (00189448) , 1
On Enumerating Feasible Permutations for Rank Modulation Codes in DNA storage via Hyperplane Arrangements
For a directed graph G with edge set {1, 2, . . . , k}, a feasible permutations for G is a permutation π on {1, 2, . . . , k} such that there exists an injective weight function p : {1, 2, . . . , k} → Z+ for which the sum of the weights of all incoming edges equals the sum of the weights of all outgoing edges at every vertex of G, provided that if p(i1) < p(i2) < · · · < p(ik), then π(it) = t for every t ∈ {1, . . . , k}. In this paper, we study the number of feasible permutations (denoted by Fq,ℓ) for the De Bruijn graph Gq,ℓ−1 , where q > 2 is the size of the alphabet. In the case q = 4, the corresponding De Bruijn graph is used in DNA data storage. To enumerate these feasible permutations, we establish a connection between feasible permutations and regions in a special hyperplane arrangement, denoted by A. Using Zaslavsky’s formula, we present some numerical results and obtain the exact number of F3,2 and F4,2. Since the formula becomes complicated for larger values of q and ℓ, we concentrate on the case ℓ = 2 and then by counting the regions in a hyperplane sub-arrangement of A, we provide a lower bound on Fq,2. We compare our bound with the latest lower bound on Fq,2 and show that our bound is Ω(q3J(q)) while the previous one is Ω(q2J(q)), where J(q) = (q2 −2q +1)!(q2)!/(q2 −q)!.
- 29 May 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Information Theory
- Vol. PP (00189448) , 1
On Enumerating Feasible Permutations for Rank Modulation Codes in DNA storage via Hyperplane Arrangements
For a directed graph G with edge set {1, 2, . . . , k}, a feasible permutations for G is a permutation π on {1, 2, . . . , k} such that there exists an injective weight function p : {1, 2, . . . , k} → Z+ for which the sum of the weights of all incoming edges equals the sum of the weights of all outgoing edges at every vertex of G, provided that if p(i1) < p(i2) < · · · < p(ik), then π(it) = t for every t ∈ {1, . . . , k}. In this paper, we study the number of feasible permutations (denoted by Fq,ℓ) for the De Bruijn graph Gq,ℓ−1 , where q > 2 is the size of the alphabet. In the case q = 4, the corresponding De Bruijn graph is used in DNA data storage. To enumerate these feasible permutations, we establish a connection between feasible permutations and regions in a special hyperplane arrangement, denoted by A. Using Zaslavsky’s formula, we present some numerical results and obtain the exact number of F3,2 and F4,2. Since the formula becomes complicated for larger values of q and ℓ, we concentrate on the case ℓ = 2 and then by counting the regions in a hyperplane sub-arrangement of A, we provide a lower bound on Fq,2. We compare our bound with the latest lower bound on Fq,2 and show that our bound is Ω(q3J(q)) while the previous one is Ω(q2J(q)), where J(q) = (q2 −2q +1)!(q2)!/(q2 −q)!.
- 19 May 2026
- Published by Springer Nature in Cryptography and Communications
- Vol. 18 (5) , 1-16
The sequence reconstruction of permutations under hamming metric with small errors
- 19 May 2026
- Published by Springer Nature in Cryptography and Communications
- Vol. 18 (5) , 1-16
The sequence reconstruction of permutations under hamming metric with small errors
- Open Access
Exact Recovery in the Data Block Model
Community detection in networks is a fundamental problem in machine learning and statistical inference, with applications in social networks, biological systems, and communication networks. The stochastic block model (SBM) serves as a canonical framework for studying community structure, and exact recovery, identifying the true communities with high probability, is a central theoretical question. While classical results characterize the phase transition for exact recovery based solely on graph connectivity, many real-world networks contain additional data, such as node attributes or labels. In this work, we study exact recovery in the Data Block Model (DBM), an SBM augmented with node-associated data, as formalized by Asadi, Abbe, and Verdú (2017). We introduce the Chernoff--TV divergence and use it to characterize a sharp exact recovery threshold for the DBM. We further provide an efficient algorithm that achieves this threshold, along with a matching converse result showing impossibility below the threshold. Finally, simulations validate our findings and demonstrate the benefits of incorporating vertex data as side information in community detection.
Open AccessExact Recovery in the Data Block Model
Community detection in networks is a fundamental problem in machine learning and statistical inference, with applications in social networks, biological systems, and communication networks. The stochastic block model (SBM) serves as a canonical framework for studying community structure, and exact recovery, identifying the true communities with high probability, is a central theoretical question. While classical results characterize the phase transition for exact recovery based solely on graph connectivity, many real-world networks contain additional data, such as node attributes or labels. In this work, we study exact recovery in the Data Block Model (DBM), an SBM augmented with node-associated data, as formalized by Asadi, Abbe, and Verdú (2017). We introduce the Chernoff--TV divergence and use it to characterize a sharp exact recovery threshold for the DBM. We further provide an efficient algorithm that achieves this threshold, along with a matching converse result showing impossibility below the threshold. Finally, simulations validate our findings and demonstrate the benefits of incorporating vertex data as side information in community detection.
- Open Access
The Sequence Reconstruction of Permutations under Hamming Metric with Small Errors
The sequence reconstruction problem asks for the recovery of a sequence from multiple noisy copies, where each copy may contain up to $r$ errors. In the case of permutations on \(n\) letters under the Hamming metric, this problem is closely related to the parameter $N(n,r)$, the maximum intersection size of two Hamming balls of radius $r$. While previous work has resolved \(N(n,r)\) for small radii (\(r \leq 4\)) and established asymptotic bounds for larger \(r\), we present new exact formulas for \(r \in \{5,6,7\}\) using group action techniques. In addition, we develop a formula for \(N(n,r)\) based on the irreducible characters of the symmetric group \(S_n\), along with an algorithm that enables computation of \(N(n,r)\) for larger parameters, including cases such as \(N(43,8)\) and \(N(24,14)\).Open AccessThe Sequence Reconstruction of Permutations under Hamming Metric with Small Errors
The sequence reconstruction problem asks for the recovery of a sequence from multiple noisy copies, where each copy may contain up to $r$ errors. In the case of permutations on \(n\) letters under the Hamming metric, this problem is closely related to the parameter $N(n,r)$, the maximum intersection size of two Hamming balls of radius $r$. While previous work has resolved \(N(n,r)\) for small radii (\(r \leq 4\)) and established asymptotic bounds for larger \(r\), we present new exact formulas for \(r \in \{5,6,7\}\) using group action techniques. In addition, we develop a formula for \(N(n,r)\) based on the irreducible characters of the symmetric group \(S_n\), along with an algorithm that enables computation of \(N(n,r)\) for larger parameters, including cases such as \(N(43,8)\) and \(N(24,14)\). - 1 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Wireless Communications
- Vol. 25, 25569-25579
Channel Estimation for Sparse mMIMO Base Stations With Hybrid Analog–Digital Architecture
- 1 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Wireless Communications
- Vol. 25, 25569-25579
Channel Estimation for Sparse mMIMO Base Stations With Hybrid Analog–Digital Architecture
- 24 May 2025
- Published by Springer Nature in Cryptography and Communications
- Vol. 17 (4) , 1075-1091
Upper bounds on the size of permutation codes with a Hamming distance of five
- 24 May 2025
- Published by Springer Nature in Cryptography and Communications
- Vol. 17 (4) , 1075-1091
Upper bounds on the size of permutation codes with a Hamming distance of five
- 15 April 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Information Theory
- Vol. 71 (6) , 4156-4166
Improved Bounds on the Size of Permutation Codes Under Kendall τ -Metric
In order to overcome the challenges caused by flash memories and also to protect against errors related to reading information stored in DNA molecules in the shotgun sequencing method, the rank modulation method has been proposed. In the rank modulation framework, codewords are permutations. In this paper, we study the largest size P(n, d) of permutation codes of length n, i.e., subsets of the set Sn of all permutations on {1, ..., n} with the minimum distance at least d ∈ {1, ..., (n/2)} under the Kendall τ-metric. By presenting an algorithm and two theorems, we improve the known lower and upper bounds for P(n, d). In particular, we show that P(n, d) = 4 for all n ≥ 6 and 3/5 (n/2) < d ≤ 2/3 (n/2). Additionally, we prove that for any prime number n and integer r ≤ n/6, P(n, 3) ≤ (n − 1)! − n − 6r/√n2 − 8rn + 20r2 √(n − 1)!/n(n − r)!. This result greatly improves the upper bound of P(n, 3) for all primes n ≥ 37.
- 15 April 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Information Theory
- Vol. 71 (6) , 4156-4166
Improved Bounds on the Size of Permutation Codes Under Kendall τ -Metric
In order to overcome the challenges caused by flash memories and also to protect against errors related to reading information stored in DNA molecules in the shotgun sequencing method, the rank modulation method has been proposed. In the rank modulation framework, codewords are permutations. In this paper, we study the largest size P(n, d) of permutation codes of length n, i.e., subsets of the set Sn of all permutations on {1, ..., n} with the minimum distance at least d ∈ {1, ..., (n/2)} under the Kendall τ-metric. By presenting an algorithm and two theorems, we improve the known lower and upper bounds for P(n, d). In particular, we show that P(n, d) = 4 for all n ≥ 6 and 3/5 (n/2) < d ≤ 2/3 (n/2). Additionally, we prove that for any prime number n and integer r ≤ n/6, P(n, 3) ≤ (n − 1)! − n − 6r/√n2 − 8rn + 20r2 √(n − 1)!/n(n − r)!. This result greatly improves the upper bound of P(n, 3) for all primes n ≥ 37.
- 23 September 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Intelligent Transportation Systems
- Vol. 25 (12) , 20435-20445
SeqNet: Sequential Networks for One-Shot Traffic Sign Recognition With Transfer Learning
In traffic sign recognition tasks, recognition of road signs by observing synthetic reference images is a human-like ability that can be performed by one-shot learning algorithms. One-shot object recognition is a challenging task for deep neural networks in which a deep model classifies query examples based on support images. It becomes more difficult when there is a domain shift between support and query samples. The generalization of a deep model on an unknown domain with different distributions is another problematic task in on-shot recognition. This work introduces a novel deep network named SeqNet to overcome the aforementioned problems. To the best of our knowledge, this work outperforms all state-of-the-art models in one-shot traffic sign recognition and one-shot logo identification by superior results. Our proposed SeqNet model generalizes to unseen domains without further model fine-tuning on the test data. Also, we show how using transferred knowledge from an irrelevant but large domain could reduce the network parameters that result in model size reduction. By utilizing the power of transferred knowledge from a large deep model the SeqNet becomes smaller and has about 6X fewer parameters than its competitors. The smaller size of the SeqNet architecture enables it to be used in resource-constrained devices in many applications such as smart vehicles. The experimental results depict that our proposed SeqNet performance is ameliorated by large margins, with up to 20% accuracy for one-shot classification and 30% area under the curve (AUC) for image retrieval tasks.
- 23 September 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Intelligent Transportation Systems
- Vol. 25 (12) , 20435-20445
SeqNet: Sequential Networks for One-Shot Traffic Sign Recognition With Transfer Learning
In traffic sign recognition tasks, recognition of road signs by observing synthetic reference images is a human-like ability that can be performed by one-shot learning algorithms. One-shot object recognition is a challenging task for deep neural networks in which a deep model classifies query examples based on support images. It becomes more difficult when there is a domain shift between support and query samples. The generalization of a deep model on an unknown domain with different distributions is another problematic task in on-shot recognition. This work introduces a novel deep network named SeqNet to overcome the aforementioned problems. To the best of our knowledge, this work outperforms all state-of-the-art models in one-shot traffic sign recognition and one-shot logo identification by superior results. Our proposed SeqNet model generalizes to unseen domains without further model fine-tuning on the test data. Also, we show how using transferred knowledge from an irrelevant but large domain could reduce the network parameters that result in model size reduction. By utilizing the power of transferred knowledge from a large deep model the SeqNet becomes smaller and has about 6X fewer parameters than its competitors. The smaller size of the SeqNet architecture enables it to be used in resource-constrained devices in many applications such as smart vehicles. The experimental results depict that our proposed SeqNet performance is ameliorated by large margins, with up to 20% accuracy for one-shot classification and 30% area under the curve (AUC) for image retrieval tasks.
- 8 May 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Communications Letters
- Vol. 28 (7) , 1479-1483
Four-Cycle Free Spatially Coupled LDPC Codes With an Explicit Construction
Spatially coupled low-density parity-check convolutional codes (SC-LDPC-CCs) are a class of capacity approaching LDPC codes with low message recovery latency when a sliding window decoding is used. Recently, an edge spreading (unwrapping) approach has been proposed to generate a class of array-based SC-LDPC-CCs by partitioning a spreading matrix into a number of components. Applying the spreading matrix E as the exponent matrix of a QC-LDPC code, the relationship of 4-cycles between E and the corresponding SC-LDPC-CCs is investigated in two cases, in this paper. Then, by defining a new class of integer finite sequences, called good sequences , a class of 4-cycle free SC-LDPC-CCs is constructed by an explicit simple method that can achieve (in most cases) the minimum coupling widths. The constructed codes enjoy a relative advantage in flexibility in rate and lengths, simple constructions, and minimal short-cycle distributions, which can show themselves in improving the bit-error-rate performances.
- 8 May 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Communications Letters
- Vol. 28 (7) , 1479-1483
Four-Cycle Free Spatially Coupled LDPC Codes With an Explicit Construction
Spatially coupled low-density parity-check convolutional codes (SC-LDPC-CCs) are a class of capacity approaching LDPC codes with low message recovery latency when a sliding window decoding is used. Recently, an edge spreading (unwrapping) approach has been proposed to generate a class of array-based SC-LDPC-CCs by partitioning a spreading matrix into a number of components. Applying the spreading matrix E as the exponent matrix of a QC-LDPC code, the relationship of 4-cycles between E and the corresponding SC-LDPC-CCs is investigated in two cases, in this paper. Then, by defining a new class of integer finite sequences, called good sequences , a class of 4-cycle free SC-LDPC-CCs is constructed by an explicit simple method that can achieve (in most cases) the minimum coupling widths. The constructed codes enjoy a relative advantage in flexibility in rate and lengths, simple constructions, and minimal short-cycle distributions, which can show themselves in improving the bit-error-rate performances.
- 28 February 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 73 (00189456) , 1-9
High Impedance Faults Detection in Distribution Systems Using Signal Correlations
High impedance fault (HIF) detection is a problem in distribution power systems protection. Detection of these faults by overcurrent protection relays is difficult or impossible. Therefore, system recovery is postponed, and individuals may be exposed to the risk of electrical shock. This paper defines a family of HIF current-time patterns with different ignition angles. To detect the HIF current, the proposed detection method searches in its family of HIF patterns and finds the pattern that has the highest correlation to the HIF current. Based on the measured correlations, a time-domain algorithm is proposed for the HIF current. If the highest correlation between the family of HIF patterns and the measured current signal is above a given threshold value, then the fault is detected as an HIF. Checked by simulations, the proposed method can correctly detect HIF within one power system cycle.
- 28 February 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 73 (00189456) , 1-9
High Impedance Faults Detection in Distribution Systems Using Signal Correlations
High impedance fault (HIF) detection is a problem in distribution power systems protection. Detection of these faults by overcurrent protection relays is difficult or impossible. Therefore, system recovery is postponed, and individuals may be exposed to the risk of electrical shock. This paper defines a family of HIF current-time patterns with different ignition angles. To detect the HIF current, the proposed detection method searches in its family of HIF patterns and finds the pattern that has the highest correlation to the HIF current. Based on the measured correlations, a time-domain algorithm is proposed for the HIF current. If the highest correlation between the family of HIF patterns and the measured current signal is above a given threshold value, then the fault is detected as an HIF. Checked by simulations, the proposed method can correctly detect HIF within one power system cycle.