arXiv subjects

Antonia Wachter-Zeh

Publications and source records attributed to Antonia Wachter-Zeh.

At least 37 records · Page 2Linked to original sources

arXiv2025

On the List-Decodability of Random (Linear) Sum-Rank Metric Codes

In this paper, we establish the list-decoding capacity theorem for sum-rank metric codes. This theorem implies the list-decodability theorem for random general sum-rank metric codes: Any random general sum-rank metric code with a rate not exceeding the list-decoding capacity is $\left(ρ,O\left(1/ε\right)\right)$-list-decodable with high probability, where $ρ\in\left(0,1\right)$ represents the error fraction and $ε>0$ is referred to as the capacity gap. For random $\mathbb{F}_q$-linear sum-rank metric codes by using the same proof approach we demonstrate that any random $\mathbb{F}_q$-linear sum-rank metric code with a rate not exceeding the list-decoding capacity is $\left(ρ,\exp\left(O\left(1/ε\right)\right)\right)$-list-decodable with high probability, where the list size is exponential at this stage due to the high correlation among codewords in linear codes. To achieve an exponential improvement on the list size, we prove a limited correlation property between sum-rank metric balls and $\mathbb{F}_q$-subspaces. Ultimately, we establish the list-decodability theorem for random $\mathbb{F}_q$-linear sum-rank metric codes: Any random $\mathbb{F}_q$-linear sum-rank metric code with rate not exceeding the list decoding capacity is $\left(ρ, O\left(1/ε\right)\right)$-list-decodable with high probability. For the proof of the list-decodability theorem of random $\mathbb{F}_q$-linear sum-rank metric codes our proof idea is inspired by and aligns with that provided in the works \cite{Gur2010,Din2014,Gur2017} where the authors proved the list-decodability theorems for random $\mathbb{F}_q$-linear Hamming metric codes and random $\mathbb{F}_q$-linear rank metric codes, respectively.

cs.IT↗

arXiv2025

Self-Regulating Random Walks for Resilient Decentralized Learning on Graphs

Consider the setting of multiple random walks (RWs) on a graph executing a certain computational task. For instance, in decentralized learning via RWs, a model is updated at each iteration based on the local data of the visited node and then passed to a randomly chosen neighbor. RWs can fail due to node or link failures. The goal is to maintain a desired number of RWs to ensure failure resilience. Achieving this is challenging due to the lack of a central entity to track which RWs have failed to replace them with new ones by forking (duplicating) surviving ones. Without duplications, the number of RWs will eventually go to zero, causing a catastrophic failure of the system. We propose two decentralized algorithms called DecAFork and DecAFork+ that can maintain the number of RWs in the graph around a desired value even in the presence of arbitrary RW failures. Nodes continuously estimate the number of surviving RWs by estimating their return time distribution and fork the RWs when failures are likely to happen. DecAFork+ additionally allows terminations to avoid overloading the network by forking too many RWs. We present extensive numerical simulations that show the performance of DecAFork and DecAFork+ regarding fast detection and reaction to failures compared to a baseline, and establish theoretical guarantees on the performance of both algorithms.

cs.LG↗

arXiv2025

Reducing Ciphertext and Key Sizes for MLWE-Based Cryptosystems

The concatenation of encryption and decryption can be interpreted as data transmission over a noisy communication channel. In this work, we use finite blocklength methods (normal approximation and random coding union bound) as well as asymptotics to show that ciphertext and key sizes of the state-of-the-art post-quantum secure key encapsulation mechanism (KEM) Kyber can be reduced without compromising the security of the scheme. We show that in the asymptotic regime, it is possible to reduce the sizes of ciphertexts and secret keys by 25% for the parameter set Kyber1024 while keeping the bitrate at 1 as proposed in the original scheme. For a single Kyber encryption block used to share a 256-bit AES key, we furthermore show that reductions in ciphertext size of 39% and 33% are possible for Kyber1024 and Kyber512, respectively.

cs.CR↗

arXiv2025

BICompFL: Stochastic Federated Learning with Bi-Directional Compression

We address the prominent communication bottleneck in federated learning (FL). We specifically consider stochastic FL, in which models or compressed model updates are specified by distributions rather than deterministic parameters. Stochastic FL offers a principled approach to compression, and has been shown to reduce the communication load under perfect downlink transmission from the federator to the clients. However, in practice, both the uplink and downlink communications are constrained. We show that bi-directional compression for stochastic FL has inherent challenges, which we address by introducing BICompFL. Our BICompFL is experimentally shown to reduce the communication cost by an order of magnitude compared to multiple benchmarks, while maintaining state-of-the-art accuracies. Theoretically, we study the communication cost of BICompFL through a new analysis of an importance-sampling based technique, which exposes the interplay between uplink and downlink communication costs.

cs.LG↗

arXiv2024

Support-Guessing Decoding Algorithms in the Sum-Rank Metric

The sum-rank metric generalizes the Hamming and rank metric by partitioning vectors into blocks and defining the total weight as the sum of the rank weights of these blocks, based on their matrix representation. In this work, we explore support-guessing algorithms for decoding sum-rank-metric codes. Support-guessing involves randomly selecting candidate supports and attempting to decode the error under the assumption that it is confined to these supports. While previous works have focused on worst-case scenarios, we analyze the average case and derive an optimal support-guessing distribution in the asymptotic regime. We show that this distribution also performs well for finite code lengths. Our analysis provides exact complexity estimates for unique decoding scenarios and establishes tighter bounds beyond the unique decoding radius. Additionally, we introduce a randomized decoding algorithm for Linearized Reed--Solomon (LRS) codes. This algorithm extends decoding capabilities beyond the unique decoding radius by leveraging an efficient error-and-erasure decoder. Instead of requiring the entire error support to be confined to the guessed support, the algorithm succeeds as long as there is sufficient overlap between the guessed support and the actual error support. As a result, the proposed method improves the success probability and reduces computational complexity compared to generic decoding algorithms. Our contributions offer more accurate complexity estimates than previous works, which are essential for understanding the computational challenges involved in decoding sum-rank-metric codes. This improved complexity analysis, along with optimized support-guessing distributions, provides valuable insights for the design and evaluation of code-based cryptosystems using the sum-rank metric. This is particularly important in the context of quantum-resistant cryptography.

cs.IT↗

arXiv2024

Sequential Decoding of Multiple Traces Over the Syndrome Trellis for Synchronization Errors

Standard decoding approaches for convolutional codes, such as the Viterbi and BCJR algorithms, entail significant complexity when correcting synchronization errors. The situation worsens when multiple received sequences should be jointly decoded, as in DNA storage. Previous work has attempted to address this via separate-BCJR decoding, i.e., combining the results of decoding each received sequence separately. Another attempt to reduce complexity adapted sequential decoders for use over channels with insertion and deletion errors. However, these decoding alternatives remain prohibitively expensive for high-rate convolutional codes. To address this, we adapt sequential decoders to decode multiple received sequences jointly over the syndrome trellis. For the short blocklength regime, this decoding strategy can outperform separate-BCJR decoding under certain channel conditions, in addition to reducing decoding complexity. To mitigate the occurrence of a decoding timeout, formally called erasure, we also extend this approach to work bidirectionally, i.e., deploying two independent stack decoders that simultaneously operate in the forward and backward directions.

cs.IT↗

arXiv2024

Private Aggregation in Hierarchical Wireless Federated Learning with Partial and Full Collusion

In federated learning, a federator coordinates the training of a model, e.g., a neural network, on privately owned data held by several participating clients. The gradient descent algorithm, a well-known and popular iterative optimization procedure, is run to train the model. Every client computes partial gradients based on their local data and sends them to the federator, which aggregates the results and updates the model. Privacy of the clients' data is a major concern. In fact, it is shown that observing the partial gradients can be enough to reveal the clients' data. Existing literature focuses on private aggregation schemes that tackle the privacy problem in federated learning in settings where all users are connected to each other and to the federator. In this paper, we consider a hierarchical wireless system architecture in which the clients are connected to base stations; the base stations are connected to the federator either directly or through relays. We examine settings with and without relays, and derive fundamental limits on the communication cost under information-theoretic privacy with different collusion assumptions. We introduce suitable private aggregation schemes tailored for these settings whose communication costs are multiplicative factors away from the derived bounds.

cs.LG↗

arXiv2024

Scalable and Reliable Over-the-Air Federated Edge Learning

Federated edge learning (FEEL) has emerged as a core paradigm for large-scale optimization. However, FEEL still suffers from a communication bottleneck due to the transmission of high-dimensional model updates from the clients to the federator. Over-the-air computation (AirComp) leverages the additive property of multiple-access channels by aggregating the clients' updates over the channel to save communication resources. While analog uncoded transmission can benefit from the increased signal-to-noise ratio (SNR) due to the simultaneous transmission of many clients, potential errors may severely harm the learning process for small SNRs. To alleviate this problem, channel coding approaches were recently proposed for AirComp in FEEL. However, their error-correction capability degrades with an increasing number of clients. We propose a digital lattice-based code construction with constant error-correction capabilities in the number of clients, and compare to nested-lattice codes, well-known for their optimal rate and power efficiency in the point-to-point AWGN channel.

cs.IT↗

arXiv2024

Linearized Reed-Solomon Codes with Support-Constrained Generator Matrix and Applications in Multi-Source Network Coding

Linearized Reed-Solomon (LRS) codes are evaluation codes based on skew polynomials. They achieve the Singleton bound in the sum-rank metric and therefore are known as maximum sum-rank distance (MSRD) codes. In this work, we give necessary and sufficient conditions for the existence of MSRD codes with a support-constrained generator matrix. The conditions on the support constraints are identical to those for MDS codes and MRD codes. The required field size for an $[n,k]_{q^m}$ LRS codes with support-constrained generator matrix is $q\geq \ell+1$ and $m\geq \max_{l\in[\ell]}\{k-1+\log_qk, n_l\}$, where $\ell$ is the number of blocks and $n_l$ is the size of the $l$-th block. The special cases of the result coincide with the known results for Reed-Solomon codes and Gabidulin codes. For the support constraints that do not satisfy the necessary conditions, we derive the maximum sum-rank distance of a code whose generator matrix fulfills the constraints. Such a code can be constructed from a subcode of an LRS code with a sufficiently large field size. Moreover, as an application in network coding, the conditions can be used as constraints in an integer programming problem to design distributed LRS codes for a distributed multi-source network.

cs.IT↗

arXiv2024

FedGT: Identification of Malicious Clients in Federated Learning with Secure Aggregation

We propose FedGT, a novel framework for identifying malicious clients in federated learning with secure aggregation. Inspired by group testing, the framework leverages overlapping groups of clients to identify the presence of malicious clients in the groups via a decoding operation. The clients identified as malicious are then removed from the model training, which is performed over the remaining clients. By choosing the size, number, and overlap between groups, FedGT strikes a balance between privacy and security. Specifically, the server learns the aggregated model of the clients in each group - vanilla federated learning and secure aggregation correspond to the extreme cases of FedGT with group size equal to one and the total number of clients, respectively. The effectiveness of FedGT is demonstrated through extensive experiments on the MNIST, CIFAR-10, and ISIC2019 datasets in a cross-silo setting under different data-poisoning attacks. These experiments showcase FedGT's ability to identify malicious clients, resulting in high model utility. We further show that FedGT significantly outperforms the private robust aggregation approach based on the geometric median recently proposed by Pillutla et al. in multiple settings.

cs.LG↗

arXiv2024

Capacity-Maximizing Input Symbol Selection for Discrete Memoryless Channels

Motivated by communication systems with constrained complexity, we consider the problem of input symbol selection for discrete memoryless channels (DMCs). Given a DMC, the goal is to find a subset of its input alphabet, so that the optimal input distribution that is only supported on these symbols maximizes the capacity among all other subsets of the same size (or smaller). We observe that the resulting optimization problem is non-concave and non-submodular, and so generic methods for such cases do not have theoretical guarantees. We derive an analytical upper bound on the capacity loss when selecting a subset of input symbols based only on the properties of the transition matrix of the channel. We propose a selection algorithm that is based on input-symbols clustering, and an appropriate choice of representatives for each cluster, which uses the theoretical bound as a surrogate objective function. We provide numerical experiments to support the findings.

cs.IT↗

arXiv2024

An End-to-End Coding Scheme for DNA-Based Data Storage With Nanopore-Sequenced Reads

We consider error-correcting coding for deoxyribonucleic acid (DNA)-based storage using nanopore sequencing. We model the DNA storage channel as a sampling noise channel where the input data is chunked into $M$ short DNA strands, which are copied a random number of times, and the channel outputs a random selection of $N$ noisy DNA strands. The retrieved DNA reads are prone to strand-dependent insertion, deletion, and substitution (IDS) errors. We construct an index-based concatenated coding scheme consisting of the concatenation of an outer code, an index code, and an inner code. We further propose a low-complexity (linear in $N$) maximum a posteriori probability decoder that takes into account the strand-dependent IDS errors and the randomness of the drawing to infer symbolwise a posteriori probabilities for the outer decoder. We present Monte-Carlo simulations for information-outage probabilities and frame error rates for different channel setups on experimental data. We finally evaluate the overall system performance using the read/write cost trade-off. A powerful combination of tailored channel modeling and soft information processing allows us to achieve excellent performance even with error-prone nanopore-sequenced reads outperforming state-of-the-art schemes.%

cs.IT↗

arXiv2024

Correcting a Single Deletion in Reads from a Nanopore Sequencer

Owing to its several merits over other DNA sequencing technologies, nanopore sequencers hold an immense potential to revolutionize the efficiency of DNA storage systems. However, their higher error rates necessitate further research to devise practical and efficient coding schemes that would allow accurate retrieval of the data stored. Our work takes a step in this direction by adopting a simplified model of the nanopore sequencer inspired by Mao \emph{et al.}, which incorporates some of its physical aspects. This channel model can be viewed as a sliding window of length $\ell$ that passes over the incoming input sequence and produces the Hamming weight of the enclosed $\ell$ bits, while shifting by one position at each time step. The resulting $(\ell+1)$-ary vector, referred to as the $\ell$-\emph{read vector}, is susceptible to deletion errors due to imperfections inherent in the sequencing process. We establish that at least $\log n - \ell$ bits of redundancy are needed to correct a single deletion. An error-correcting code that is optimal up to an additive constant, is also proposed. Furthermore, we find that for $\ell \geq 2$, reconstruction from two distinct noisy $\ell$-read vectors can be accomplished without any redundancy, and provide a suitable reconstruction algorithm to this effect.

cs.IT↗

arXiv2024

Coding for Composite DNA to Correct Substitutions, Strand Losses, and Deletions

Composite DNA is a recent method to increase the base alphabet size in DNA-based data storage.This paper models synthesizing and sequencing of composite DNA and introduces coding techniques to correct substitutions, losses of entire strands, and symbol deletion errors. Non-asymptotic upper bounds on the size of codes with $t$ occurrences of these error types are derived. Explicit constructions are presented which can achieve the bounds.

cs.IT↗

arXiv2024

Maximal-Capacity Discrete Memoryless Channel Identification

The problem of identifying the channel with the highest capacity among several discrete memoryless channels (DMCs) is considered. The problem is cast as a pure-exploration multi-armed bandit problem, which follows the practical use of training sequences to sense the communication channel statistics. A capacity estimator is proposed and tight confidence bounds on the estimator error are derived. Based on this capacity estimator, a gap-elimination algorithm termed BestChanID is proposed, which is oblivious to the capacity-achieving input distribution and is guaranteed to output the DMC with the largest capacity, with a desired confidence. Furthermore, two additional algorithms NaiveChanSel and MedianChanEl, that output with certain confidence a DMC with capacity close to the maximal, are introduced. Each of those algorithms is beneficial in a different regime and can be used as a subroutine in BestChanID. The sample complexity of all algorithms is analyzed as a function of the desired confidence parameter, the number of channels, and the channels' input and output alphabet sizes. The cost of best channel identification is shown to scale quadratically with the alphabet size, and a fundamental lower bound for the required number of channel senses to identify the best channel with a certain confidence is derived.

cs.IT↗

arXiv2023

Error-Correcting Codes for Nanopore Sequencing

Nanopore sequencing, superior to other sequencing technologies for DNA storage in multiple aspects, has recently attracted considerable attention. Its high error rates, however, demand thorough research on practical and efficient coding schemes to enable accurate recovery of stored data. To this end, we consider a simplified model of a nanopore sequencer inspired by Mao \emph{et al.}, incorporating intersymbol interference and measurement noise. Essentially, our channel model passes a sliding window of length $\ell$ over a $q$-ary input sequence that outputs the \textit{composition} of the enclosed $\ell$ bits and shifts by $δ$ positions with each time step. In this context, the composition of a $q$-ary vector $\bfx$ specifies the number of occurrences in $\bfx$ of each symbol in $\lbrace 0,1,\ldots, q-1\rbrace$. The resulting compositions vector, termed the \emph{read vector}, may also be corrupted by $t$ substitution errors. By employing graph-theoretic techniques, we deduce that for $δ=1$, at least $\log \log n$ symbols of redundancy are required to correct a single ($t=1$) substitution. Finally, for $\ell \geq 3$, we exploit some inherent characteristics of read vectors to arrive at an error-correcting code that is of optimal redundancy up to a (small) additive constant for this setting. This construction is also found to be optimal for the case of reconstruction from two noisy read vectors.

cs.IT↗

arXiv2023

Sequential Decoding of Convolutional Codes for Synchronization Errors

Sequential decoding, commonly applied to substitution channels, is a sub-optimal alternative to Viterbi decoding with significantly reduced memory costs. In this work, a sequential decoder for convolutional codes over channels that are prone to insertion, deletion, and substitution errors, is described and analyzed. Our decoder expands the code trellis by a new channel-state variable, called drift state, as proposed by Davey and MacKay. A suitable decoding metric on that trellis for sequential decoding is derived, generalizing the original Fano metric. The decoder is also extended to facilitate the simultaneous decoding of multiple received sequences that arise from a single transmitted sequence. Under low-noise environments, our decoding approach reduces the decoding complexity by a couple orders of magnitude in comparison to Viterbi's algorithm, albeit at slightly higher bit error rates. An analytical method to determine the computational cutoff rate is also suggested. This analysis is supported with numerical evaluations of bit error rates and computational complexity, which are compared with respect to optimal Viterbi decoding.

cs.IT↗

arXiv2023

Sparsity and Privacy in Secret Sharing: A Fundamental Trade-Off

This work investigates the design of sparse secret sharing schemes that encode a sparse private matrix into sparse shares. This investigation is motivated by distributed computing, where the multiplication of sparse and private matrices is moved from a computationally weak main node to untrusted worker machines. Classical secret-sharing schemes produce dense shares. However, sparsity can help speed up the computation. We show that, for matrices with i.i.d. entries, sparsity in the shares comes at a fundamental cost of weaker privacy. We derive a fundamental tradeoff between sparsity and privacy and construct optimal sparse secret sharing schemes that produce shares that leak the minimum amount of information for a desired sparsity of the shares. We apply our schemes to distributed sparse and private matrix multiplication schemes with no colluding workers while tolerating stragglers. For the setting of two non-communicating clusters of workers, we design a sparse one-time pad so that no private information is leaked to a cluster of untrusted and colluding workers, and the shares with bounded but non-zero leakage are assigned to a cluster of partially trusted workers. We conclude by discussing the necessity of using permutations for matrices with correlated entries.

cs.CR↗