We investigate privacy-preserving spectral clustering for community detection within stochastic block models (SBMs). Specifically, we focus on edge differential privacy (DP) and propose private algorithms for community recovery. Our work explores the fundamental trade-offs between the privacy budget and the accurate recovery of community labels. Furthermore, we establish information-theoretic conditions that guarantee the accuracy of our methods, providing theoretical assurances for successful community recovery under edge DP.

Hybrid Language Models (HLMs) are inference-time architectures that combine the low-latency efficiency of Small Language Models (SLMs) on clients (edge devices) with the high accuracy of Large Language Models (LLMs) in centralized servers. Unlike traditional end-to-end LLM inference, HLMs aim to reduce latency and communication by selectively invoking LLMs only when the local SLM’s predictions are uncertain—that is, when the model exhibits low confidence or high entropy in its token-level probability distribution. However, when the SLM encounters ambiguous or low-confidence predictions during inference, it must offload token-level probability distributions to the LLM for refinement. This frequent offloading leads to substantial communication overhead, particularly in bandwidth-constrained environments. To address this challenge, we propose FedHLM, a communication-efficient HLM framework that integrates uncertainty-aware inference with Federated Learning (FL). The key innovation lies in collaboratively learning token-level uncertainty thresholds that determine when SLM predictions require LLM assistance. Instead of relying on static or hand-tuned thresholds, FedHLM uses FL to enable distributed threshold optimization across clients while preserving data privacy. Additionally, embedding-based token representations are employed to facilitate semantic similarity comparisons during Peer-to-Peer (P2P) resolution, allowing clients to reuse tokens inferred by similar peers without efficiently involving the LLM. Moreover, we propose hierarchical model aggregation as a strategy to reduce redundant token transmissions. At the edge server level, client updates are aggregated to refine local routing policies, while global coordination across clusters further synchronizes decision boundaries. This layered approach ensures that repeated uncertainty patterns are captured and resolved locally, significantly reducing unnecessary LLM queries. Extensive simulations on large-scale news classification tasks demonstrate that FedHLM achieves over 95% reduction in LLM transmissions with negligible accuracy loss, highlighting its potential for scalable and efficient edge-Artificial Intelligence (AI) deployment.

Collaborative inference among multiple wireless edge devices has the potential to significantly enhance Artificial Intelligence (AI) applications, particularly for sensing and computer vision. This approach typically involves a three-stage process: a) data acquisition through sensing, b) feature extraction, and c) feature encoding for transmission. However, transmitting the extracted features poses a significant privacy risk, as sensitive personal data can be exposed during the process. To address this challenge, we propose a novel privacy-preserving collaborative inference mechanism, wherein each edge device in the network secures the privacy of extracted features before transmitting them to a central server for inference. Our approach is designed to achieve two primary objectives: 1) reducing communication overhead and 2) ensuring strict privacy guarantees during feature transmission, while maintaining effective inference performance. Additionally, we introduce an over-the-air pooling scheme specifically designed for classification tasks, which provides formal guarantees on the privacy of transmitted features and establishes a lower bound on classification accuracy.

We study the private online change detection problem for dynamic communities, using a censored block model (CBM). We consider edge differential privacy (DP) in both local and central settings, and propose joint change detection and community estimation procedures for both scenarios. We seek to understand the fundamental tradeoffs between the privacy budget, detection delay, and exact community recovery of community labels. Further, we provide theoretical guarantees for the effectiveness of our proposed method by showing necessary and sufficient conditions for change detection and exact recovery under edge DP. Simulation and real data examples are provided to validate the proposed methods.

In this paper, we study distributed training by majority vote with the sign stochastic gradient descent (signSGD) along with over-the-air computation (OAC) under local differential privacy constraints. In our approach, the users first clip the local stochastic gradients and inject a certain amount of noise as a privacy enhancement strategy. Subsequently, they activate the indices of OFDM subcarriers based on the signs of the perturbed local stochastic gradients to realize a frequency-shift-keying-based majority vote computation at the parameter server. We evaluate the privacy benefits of the proposed approach and characterize the per-user privacy leakage theoretically. Our results show that the proposed technique improves the privacy guarantees and limits the leakage to a scaling factor of$\mathcal{O}(1/\sqrt{K})$, where$K$is the number of users, thanks to the superposition property of the wireless channel. With numerical experiments, we show that the proposed non-coherent aggregation is superior to quadrature-phase-shift-keying-based coherent aggregation, namely, one-bit digital aggregation (OBDA), in learning accuracy under time synchronization errors when the same privacy enhancement strategy is introduced to both methods.

This work considers the problem of Distributed Mean Estimation (DME) over networks with intermittent connectivity, where the goal is to learn a global statistic over the data samples localized across distributed nodes with the help of a central server. To mitigate the impact of intermittent links, nodes can collaborate with their neighbors to compute local consensus which they forward to the central server. In such a setup, the communications between any pair of nodes must satisfy local differential privacy constraints. We study the tradeoff between collaborative relaying and privacy leakage due to the additional data sharing among nodes and, subsequently, propose a novel differentially private collaborative algorithm for DME to achieve the optimal tradeoff. Finally, we present numerical simulations to substantiate our theoretical findings.

In this paper, we consider the problem of answering count queries for genomic data subject to perfect privacy constraints. Count queries are often used in applications that collect aggregate (population-wide) information from biomedical Databases (DBs) for analysis, such as Genome-wide association studies. Our goal is to design mechanisms for answering count queries of the following form:How many users in the database have a specific set of genotypes at certain locations in their genome?At the same time, we aim to achieve perfect privacy (zero information leakage) of the sensitive genotypes at a pre-specified set of secret locations. The sensitive genotypes could indicate rare diseases and/or other health traits one may want to keep private. We present both local and central count-query mechanisms for the above problem that achieves perfect information-theoretic privacy for sensitive genotypes while minimizing the expected absolute error (or per-user error probability, depending on the setting) of the query answer. We also derived a lower bound of the per-user probability of error for an arbitrary query-answering mechanism that satisfies perfect privacy. We show that our mechanisms achieve error close to the lower bound, and match the lower bound for some special cases. We numerically show that the performance of each mechanism depends on the data prior distribution, the intersection between the queried and sensitive genotypes, and the strength of the correlation in the genomic data sequence.

In this paper, we study the problem of federated learning over a wireless channel with user sampling, modeled by a Gaussian multiple access channel, subject to central and local differential privacy (DP/LDP) constraints. It has been shown that the superposition nature of the wireless channel provides a dual benefit of bandwidth efficient gradient aggregation, in conjunction with strong DP guarantees for the users. Specifically, the central DP privacy leakage has been shown to scale as$\mathcal{O}(1/\sqrt{K})$, where$K$is the number of users. It has also been shown that user sampling coupled with orthogonal transmission can enhance the central DP privacy leakage with the same scaling behavior. In this work, we show that, by jointly incorporating both wireless aggregation and user sampling, one can obtain even stronger central DP guarantees. We propose a private wireless gradient aggregation scheme, which relies on independently randomized participation decisions by each user. The central DP leakage of our proposed scheme scales as$\mathcal{O}(1/K^{3/4})$. In addition, we show that LDP is also boosted by user sampling.

In this paper, we study Local Information Privacy (LIP). As a context-aware privacy notion, LIP relaxes the de facto standard privacy notion of local differential privacy (LDP) by incorporating prior knowledge and therefore achieving better utility. We study the relationships between LIP and some of the representative privacy notions including LDP, mutual information and maximal leakage. We show that LIP provides strong instance-wise privacy protection compared to other context-aware privacy notions. Moreover, we present some useful properties of LIP, including post-processing, linkage, composability, transferability and robustness to imperfect prior knowledge. Then we study a general utility-privacy tradeoff framework, under which we derive LIP based privacy-preserving mechanisms for both discrete and continuous-valued data. Three types of perturbation mechanisms are studied in this paper: 1) randomized response (RR), 2) random sampling (RS) and 3) additive noise (AN) (e.g., Gaussian mechanism). Our privacy mechanisms incorporate the prior knowledge into the perturbation parameters so as to enhance utility. Finally, we present a comprehensive set of experiments on real datasets to illustrate the advantage of context-awareness and compare the utility-privacy tradeoffs provided by different mechanisms.

In this paper, we study the problem of federated learning (FL) over a wireless channel, modeled by a Gaussian multiple access channel (MAC), subject to local differential privacy (LDP) constraints. We show that the superposition nature of the wireless channel provides a dual benefit of bandwidth efficient gradient aggregation, in conjunction with strong LDP guarantees for the users. We propose a private wireless gradient aggregation scheme, which shows that when aggregating gradients from K users, the privacy leakage per user scales as O(1/√K) compared to orthogonal transmission in which the privacy leakage scales as a constant. We also present analysis for the convergence rate of the proposed private FL aggregation algorithm and study the tradeoffs between wireless resources, convergence, and privacy.

In this paper, we consider the problem of designing additive noise mechanisms for data release subject to a local information privacy constraint. While there has been significant prior work on devising additive noise mechanisms for differential privacy (such as Laplacian and Gaussian mechanisms), for the notion of information privacy, which accounts for prior-knowledge about the data, there are no such general purpose additive noise mechanisms. To this end, we devise a prior-aware Laplacian noise mechanism, which satisfies local information privacy. We show that adding context awareness (i.e., via the knowledge of prior of the data) improves the tradeoff between utility and privacy when compared to context-unaware mechanisms.

We study the impact of untrusted relays on the degrees of freedom of multi-antenna multi-hop networks. In par- ticular, we consider the two user two-hop interference network, where two source nodes want to send independent messages securely to their designated receivers through the help of two untrusted relays. The relays are considered untrusted in terms of eavesdropping the messages sent by the sources. Moreover, we also assume that the messages are confidential, i.e., each receiver must not be able to decode the information meant for the other receiver. We assume that all the terminals (i.e., sources, relays, and the receivers) are equipped with multiple number of antennas. The goal of this work is to understand the secure degrees of freedom (SDoF) region of this multi-hop MIMO network under the two constraints of a) untrusted relays; and b) confidential messages. To cope with the untrusted nature of relays, we present achievable schemes in which both sources mix their information symbols with artificial noises so that the signals at each relay are completely immersed in the artificial noises space. However, this mixing must be done carefully, so as to ensure the feasibility of interference neutralization in the second hop to allow successful decoding at the respective destination. To this end, we devise transmission schemes based on interference alignment and interference neutralization techniques. The main contributions of this work are as follows: a) we present an upper bound on the SDoF region as a function of the number of antennas at the terminals, b) we present two achievable schemes, the first scheme is based on secure interference alignment and neutralization and is shown to be information theoretically optimal when all terminals have the same number of antennas; and a second scheme, based on secure sub-space alignment and neutralization, which is shown to be optimal for another specific antenna configuration. To the best of our knowledge, these are the first results on multi-hop MIMO relay networks with untrusted relays and confidential messages.

In this paper, the K-user interference channel with confidential messages is considered with delayed channel state information at transmitters (CSIT). We propose a novel secure transmission scheme in which the transmitters carefully mix information symbols with artificial noises to ensure confidentiality. Achieving confidentiality is challenging due to the delayed nature of CSIT, and the distributed nature of the transmitters. Our scheme works over two phases: phase one in which each transmitter sends information symbols mixed with artificial noises, and repeats such transmission over multiple rounds. In the next phase, each transmitter uses delayed CSIT of the previous phase and sends a function of the net interference and artificial noises (generated in previous phase), which is simultaneously useful for all receivers. These phases are designed to ensure the decodability of the desired messages while satisfying the confidentiality constraints. The proposed scheme achieves a sum secure degrees of freedom (SDoF) of at least [1/2](√K-6). To the best of our knowledge, this is the first result on the K-user interference channel with confidential messages and delayed CSIT that achieves a SDoF which scales with K -.

Device-to-Device (D2D) communications have been highlighted as one of the promising solutions to enhance spectrum utilization of LTE-Advanced networks. In this paper, we consider a D2D transmitter cooperating with a cellular network by acting as a relay to serve one of the cellular users. We consider the case in which the D2D transmitter is equipped with an energy harvesting capability. We investigate the trade-off between the amount of energy used for relaying and the energy used for decoding the cellular user data at the relaying node. We formulate an optimization problem to maximize the cellular user rate subject to a minimum rate requirement constraint for the D2D link. Moreover, we consider the case when receiving nodes are equipped with successive interference cancellation (SIC) capability and investigate the effect of using SIC on our proposed system performance. Finally, we show via numerical simulations the benefits of our cooperation-based system as compared to the non-cooperative scenario.

Compressive Sensing has been utilized in Cognitive Radio Networks (CRNs) to exploit the sparse nature of the occupation of the primary users. Also, distributed spectrum sensing has been proposed to tackle the wireless channel problems, like node or link failures, rather than the common “centralized approach” for spectrum sensing. In this paper, we propose a distributed spectrum sensing framework based on consensus algorithms where SU nodes exchange their binary decisions to take global decisions without a fusion center to coordinate the sensing process. Each SU will share its decision with its neighbors, and at every new iteration each SU will take a new decision based on its current decision and the decisions it receives from its neighbors; in the next iteration, each SU will share its new decision with its neighbors. We show via simulations that the detection performance can tend to the performance of majority-rule Fusion Center based CRNs.

Channel state information at the transmitter affects the degrees of freedom of the wireless networks. In this paper, we analyze the DoF for the K-user multiple-input single-output (MISO) broadcast channel (BC) with synergistic alternating channel state information at the transmitter (CSIT). Specifically, the CSIT of each user alternates between three states, namely, perfect CSIT (P), delayed CSIT (D) and no CSIT (N) among different time slots. For the K- user MISO BC, we show that the total achievable degrees of freedom (DoF) are given by K2/(2K-1) through utilizing the synergistic benefits of CSIT patterns. We compare the achievable DoF with results reported previously in the literature in the case of delayed CSIT and hybrid CSIT models.

Censoring has been proposed to be utilized in wireless distributed detection networks with a fusion center to enhance network performance in terms of error probability in addition to the well- established energy saving gains. In this paper, we further examine the employment of censoring in infrastructure-less cognitive radio networks, where nodes employ binary consensus algorithms to take global decisions regarding a binary hypothesis test without a fusion center to coordinate such a process. We show analytically - and verify by simulations - that censoring enhances the performance of such networks in terms of error probability and convergence times. Our protocol shows performance gains up to 46.6\% in terms of average error probability over its conventional counterpart, in addition to performance gains of about 48.7% in terms of average energy expenditure and savings up to 50% in incurred transmission overhead.