arXiv subjects

Mohammad Hossein Manshaei

Publications and source records attributed to Mohammad Hossein Manshaei.

At least 19 recordsLinked to original sources

arXiv2025

MAGE-ID: A Multimodal Generative Framework for Intrusion Detection Systems

Modern Intrusion Detection Systems (IDS) face severe challenges due to heterogeneous network traffic, evolving cyber threats, and pronounced data imbalance between benign and attack flows. While generative models have shown promise in data augmentation, existing approaches are limited to single modalities and fail to capture cross-domain dependencies. This paper introduces MAGE-ID (Multimodal Attack Generator for Intrusion Detection), a diffusion-based generative framework that couples tabular flow features with their transformed images through a unified latent prior. By jointly training Transformer and CNN-based variational encoders with an EDM style denoiser, MAGE-ID achieves balanced and coherent multimodal synthesis. Evaluations on CIC-IDS-2017 and NSL-KDD demonstrate significant improvements in fidelity, diversity, and downstream detection performance over TabSyn and TabDDPM, highlighting the effectiveness of MAGE-ID for multimodal IDS augmentation.

cs.LG↗

arXiv2025

Modeling Visual Hallucination: A Generative Adversarial Network Framework

Visual hallucination refers to the perception of recognizable things that are not present. These phenomena are commonly linked to a range of neurological/psychiatric disorders. Despite ongoing research, the mechanisms through which the visual system generates hallucinations from real-world environments are still not well understood. Abnormal interactions between different regions of the brain responsible for perception are known to contribute to the occurrence of visual hallucinations. In this study, we propose and extend a generative neural network-based framework to address challenges within the visual system, aiming to create goal-driven models inspired by neurobiological mechanisms of visual hallucinations. We focus on the adversarial interactions between the visual system and the frontal lobe regions, proposing the Hallu-GAN model to suggest how these interactions can give rise to visual hallucinations. The architecture of the Hallu-GAN model is based on generative adversarial networks. Our simulation results indicate that disturbances in the ventral stream can lead to visual hallucinations. To further analyze the impact of other brain regions on the visual system, we extend the Hallu-GAN model by adding EEG data from individuals. This extended model, referred to as Hallu-GAN+, enables the examination of both hallucinating and non-hallucinating states. By training the Hallu-GAN+ model with EEG data from an individual with Charles Bonnet syndrome, we demonstrated its utility in analyzing the behavior of those experiencing hallucinations. Our simulation results confirmed the capability of the proposed model in resembling the visual system in both healthy and hallucinating states.

q-bio.NC↗

arXiv2025

Empowering Persian LLMs for Instruction Following: A Novel Dataset and Training Approach

Instruction-tuned large language models have demonstrated remarkable capabilities in following human instructions across various domains. However, their proficiency remains notably deficient in many low-resource languages. To address this challenge, we begin by introducing FarsInstruct a comprehensive instruction dataset designed to enhance the instruction following ability of large language models specifically for the Persian language a significant yet underrepresented language globally. FarsInstruct encompasses a wide range of task types and datasets, each containing a mix of straightforward to complex manual written instructions, as well as translations from the Public Pool of Prompts, ensuring a rich linguistic and cultural representation. Furthermore, we introduce Co-CoLA, a framework designed to enhance the multi-task adaptability of LoRA-tuned models. Through extensive experimental analyses, our study showcases the effectiveness of the FarsInstruct dataset coupled with training by the Co-CoLA framework, in improving the performance of large language models within the Persian context. As of the current writing, FarsInstruct comprises 197 templates across 21 distinct datasets, and we intend to update it consistently, thus augmenting its applicability.

cs.CL↗

arXiv2023

RADIANCE: Radio-Frequency Adversarial Deep-learning Inference for Automated Network Coverage Estimation

Radio-frequency coverage maps (RF maps) are extensively utilized in wireless networks for capacity planning, placement of access points and base stations, localization, and coverage estimation. Conducting site surveys to obtain RF maps is labor-intensive and sometimes not feasible. In this paper, we propose radio-frequency adversarial deep-learning inference for automated network coverage estimation (RADIANCE), a generative adversarial network (GAN) based approach for synthesizing RF maps in indoor scenarios. RADIANCE utilizes a semantic map, a high-level representation of the indoor environment to encode spatial relationships and attributes of objects within the environment and guide the RF map generation process. We introduce a new gradient-based loss function that computes the magnitude and direction of change in received signal strength (RSS) values from a point within the environment. RADIANCE incorporates this loss function along with the antenna pattern to capture signal propagation within a given indoor configuration and generate new patterns under new configuration, antenna (beam) pattern, and center frequency. Extensive simulations are conducted to compare RADIANCE with ray-tracing simulations of RF maps. Our results show that RADIANCE achieves a mean average error (MAE) of 0.09, root-mean-squared error (RMSE) of 0.29, peak signal-to-noise ratio (PSNR) of 10.78, and multi-scale structural similarity index (MS-SSIM) of 0.80.

cs.LG↗

arXiv2022

Game of GANs: Game-Theoretical Models for Generative Adversarial Networks

Generative Adversarial Networks (GANs) have recently attracted considerable attention in the AI community due to its ability to generate high-quality data of significant statistical resemblance to real data. Fundamentally, GAN is a game between two neural networks trained in an adversarial manner to reach a zero-sum Nash equilibrium profile. Despite the improvement accomplished in GANs in the last few years, several issues remain to be solved. This paper reviews the literature on the game theoretic aspects of GANs and addresses how game theory models can address specific challenges of generative model and improve the GAN's performance. We first present some preliminaries, including the basic GAN model and some game theory background. We then present taxonomy to classify state-of-the-art solutions into three main categories: modified game models, modified architectures, and modified learning methods. The classification is based on modifications made to the basic GAN model by proposed game-theoretic approaches in the literature. We then explore the objectives of each category and discuss recent works in each category. Finally, we discuss the remaining challenges in this field and present future research directions.

cs.LG↗

arXiv2021

Optimal Association Strategy of Multi-gateway Wireless Sensor Networks Against Smart Jammers

Engineers have numerous low-power wireless sensor devices in the current network setup for the Internet of Things, such as ZigBee, LoRaWAN, ANT, or Bluetooth. These low-power wireless sensors are the best candidates to transfer and collect data. But they are all vulnerable to the physical jamming attack since it is not costly for the attackers to run low power jammer sources in these networks. Having multiple gateways and providing alternative connections to sensors would help these networks to mitigate successful jamming. In this paper, we propose an analytical model to solve the problem of gateway selection and association based on a Stackelberg game, where the jammer is the follower. We first formulate the payoffs of both sensor network and attacker and then establish and prove the conditions leading to NASH equilibrium. With numerical investigation, we also present how our model can capture the performance of sensor networks under jamming with a varying number of gateways. Our results show that compared to the single gateway scenario, the network's throughput will improve by 26% and 60% when we deploy two and four gateways in the presence of a single jammer.

cs.GT↗

arXiv2021

Wide-AdGraph: Detecting Ad Trackers with a Wide Dependency Chain Graph

Websites use third-party ads and tracking services to deliver targeted ads and collect information about users that visit them. These services put users' privacy at risk, and that is why users' demand for blocking these services is growing. Most of the blocking solutions rely on crowd-sourced filter lists manually maintained by a large community of users. In this work, we seek to simplify the update of these filter lists by combining different websites through a large-scale graph connecting all resource requests made over a large set of sites. The features of this graph are extracted and used to train a machine learning algorithm with the aim of detecting ads and tracking resources. As our approach combines different information sources, it is more robust toward evasion techniques that use obfuscation or changing the usage patterns. We evaluate our work over the Alexa top-10K websites and find its accuracy to be 96.1% biased and 90.9% unbiased with high precision and recall. It can also block new ads and tracking services, which would necessitate being blocked by further crowd-sourced existing filter lists. Moreover, the approach followed in this paper sheds light on the ecosystem of third-party tracking and advertising.

cs.CR↗

arXiv2021

Efficient UAV Trajectory-Planning using Economic Reinforcement Learning

Advances in unmanned aerial vehicle (UAV) design have opened up applications as varied as surveillance, firefighting, cellular networks, and delivery applications. Additionally, due to decreases in cost, systems employing fleets of UAVs have become popular. The uniqueness of UAVs in systems creates a novel set of trajectory or path planning and coordination problems. Environments include many more points of interest (POIs) than UAVs, with obstacles and no-fly zones. We introduce REPlanner, a novel multi-agent reinforcement learning algorithm inspired by economic transactions to distribute tasks between UAVs. This system revolves around an economic theory, in particular an auction mechanism where UAVs trade assigned POIs. We formulate the path planning problem as a multi-agent economic game, where agents can cooperate and compete for resources. We then translate the problem into a Partially Observable Markov decision process (POMDP), which is solved using a reinforcement learning (RL) model deployed on each agent. As the system computes task distributions via UAV cooperation, it is highly resilient to any change in the swarm size. Our proposed network and economic game architecture can effectively coordinate the swarm as an emergent phenomenon while maintaining the swarm's operation. Evaluation results prove that REPlanner efficiently outperforms conventional RL-based trajectory search.

cs.AI↗

arXiv2020

Evaluating Sparse Interpretable Word Embeddings for Biomedical Domain

Word embeddings have found their way into a wide range of natural language processing tasks including those in the biomedical domain. While these vector representations successfully capture semantic and syntactic word relations, hidden patterns and trends in the data, they fail to offer interpretability. Interpretability is a key means to justification which is an integral part when it comes to biomedical applications. We present an inclusive study on interpretability of word embeddings in the medical domain, focusing on the role of sparse methods. Qualitative and quantitative measurements and metrics for interpretability of word vector representations are provided. For the quantitative evaluation, we introduce an extensive categorized dataset that can be used to quantify interpretability based on category theory. Intrinsic and extrinsic evaluation of the studied methods are also presented. As for the latter, we propose datasets which can be utilized for effective extrinsic evaluation of word vectors in the biomedical domain. Based on our experiments, it is seen that sparse word vectors show far more interpretability while preserving the performance of their original vectors in downstream tasks.

cs.CL↗

arXiv2020

On Coordination of Smart Grid and Cooperative Cloud Providers

Cooperative cloud providers in the form of cloud federations can potentially reduce their energy costs by exploiting electricity price fluctuations across different locations. In this environment, on the one hand, the electricity price has a significant influence on the federations formed, and, thus, on the profit earned by the cloud providers, and on the other hand, the cloud cooperation has an inevitable impact on the performance of the smart grid. In this regard, the interaction between independent cloud providers and the smart grid is modeled as a two-stage Stackelberg game interleaved with a coalitional game in this paper. In this game, in the first stage the smart grid, as a leader chooses a proper electricity pricing mechanism to maximize its own profit. In the second stage, cloud providers cooperatively manage their workload to minimize their electricity costs. Given the dynamic of cloud providers in the federation formation process, an optimization model based on a constrained Markov decision process (CMDP) has been used by the smart grid to achieve the optimal policy. Numerical results show that the proposed solution yields around 28% and 29% profit improvement on average for the smart grid, and the cloud providers, respectively, compared to the noncooperative scheme

cs.NI↗

arXiv2020

On the Feasibility of Sybil Attacks in Shard-Based Permissionless Blockchains

Bitcoin's single leader consensus protocol (Nakamoto consensus) suffers from significant transaction throughput and network scalability issues due to the computational requirements of it Proof-of-Work (PoW) based leader selection strategy. To overcome this, committee-based approaches (e.g., Elastico) that partition the outstanding transaction set into shards and (randomly) select multiple committees to process these transactions in parallel have been proposed and have become very popular. However, by design these committee or shard-based blockchain solutions are easily vulnerable to the Sybil attacks, where an adversary can easily compromise/manipulate the consensus protocol if it has enough computational power to generate multiple Sybil committee members (by generating multiple valid node identifiers). Despite the straightforward nature of these attacks, they have not been systematically analyzed. In this paper, we fill this research gap by modelling and analyzing Sybil attacks in a representative and popular shard-based protocol called Elastico. We show that the PoW technique used for identifier or ID generation in the initial phase of the protocol is vulnerable to Sybil attacks, and a node with high hash-power can generate enough Sybil IDs to successfully compromise Elastico. We analytically derive conditions for two different categories of Sybil attacks and perform numerical simulations to validate our theoretical results under different network and protocol parameters.

cs.CR↗

arXiv2019

Strategic Defense against Stealthy Link Flooding Attacks: A Signaling Game Approach

With the increasing diversity of Distributed Denial-of-Service (DDoS) attacks, it is becoming extremely challenging to design a fully protected network. For instance, Stealthy Link Flooding Attack (SLFA) is a variant of DDoS attacks that strives to block access to a target area by flooding a small set of links, and it is shown that it can bypass traditional DDoS defense mechanisms. One potential solution to tackle such SLFAs is to apply Moving Target Defense (MTD) techniques in which network settings are dynamically changed to confuse/deceive attackers, thus making it highly expensive to launch a successful attack. However, since MTD comes with some overhead to the network, to find the best strategy (i.e., when and/or to what extent) of applying it has been a major challenge. The strategy is significantly influenced by the attacker's behavior that is often difficult to guess. In this work, we address the challenge of obtaining the optimal MTD strategy that effectively mitigates SLFAs while incurs a minimal overhead. We design the problem as a signaling game considering the network defender and the attacker as players. A belief function is established throughout the engagement of the attacker and the defender during this SLFA campaign, which is utilized to pick the best response/action for each player. We analyze the game model and derive a defense mechanism based on the equilibria of the game. We evaluate the technique on a Mininet-based network environment where an attacker is performing SLFAs and a defender applies MTD based on equilibria of the game. The results show that our signaling game-based dynamic defense mechanism can provide a similar level of protection against SLFAs like the extensive MTD solution, however, causing a significantly reduced overhead.

cs.NI↗

arXiv2019

On Incentive Compatible Role-based Reward Distribution in Algorand

Algorand is a recent, open-source public or permissionless blockchain system that employs a novel proof-of-stake byzantine consensus protocol to efficiently scale the distributed transaction agreement problem to billions of users. In addition to being more democratic and energy-efficient, compared to popular protocols such as Bitcoin, Algorand also touts a much high transaction throughput. This paper is the first attempt in the literature to study and address this problem. By carefully modeling the participation costs and rewards received within a strategic interaction scenario, we first empirically show that even a small number of nodes defecting to participate in the protocol tasks due to insufficiency of the available incentives can result in the Algorand network failing to compute and add new blocks of transactions. We further show that this effect can be formalized by means of a mathematical model of interaction in Algorand given its participation costs and the current (or planned) reward distribution/sharing approach envisioned by the Algorand Foundation. Specifically, on analyzing this game model we observed that mutual cooperation under the currently proposed reward sharing approach is not a Nash equilibrium. This is a significant result which could threaten the success of an otherwise robust distributed consensus mechanism. We propose a novel reward sharing approach for Algorand and formally show that it is incentive-compatible, i.e., it can guarantee cooperation within a group of selfish Algorand users. Extensive numerical and Algorand simulation results further confirm our analytical findings. Moreover, these results show that for a given distribution of stakes in the network, our reward sharing approach can guarantee cooperation with a significantly smaller reward per round.

cs.CR↗

arXiv2019

Fostering Peer Learning through a New Game-Theoretical Approach in a Blended Learning Environment

Obtaining knowledge and skill achievement through peer learning can lead to higher academic achievement. However, peer learning implementation is not just about putting students together and hoping for the best. At its worst-designed, peer learning may result in one person doing all the effort for instance, or may fail to encourage the students to interact enough with the task and so enhance the task in hand. This study proposes a mechanism as well as an instructional design to foster well-organized peer learning based on game theory $(PD\_PL)$. The proposed mechanism uses prisoner's dilemma and maps the strategy and payoff concepts found in prisoner's dilemma onto a peer learning atmosphere. PD\_PL was implemented during several sessions of four university courses and with 142 computer engineering students. %The results of the pre-test and post-test exams of all the sessions were compared with R software through Paired Hotelling's T-Square analysis in order to investigate the impacts of $PD\_PL$ and the proposed instructional design on students' personal learning. The study results indicated that PD\_PL was beneficial and favourable to the students. Further analysis showed that the $PD\_PL$ had sometimes even enhanced learning by up to $47.2\%$. %The results of a subjective evaluation showed that the majority of respondents found $PD\_PL$ to be an attractive and efficient tool for learning enhancement. %Everybody who is interested in designing peer learning programs and tools will find this study interesting.

cs.CY↗

arXiv2019

Analyzing Defense Strategies Against Mobile Information Leakages: A Game-Theoretic Approach

Abuse of zero-permission sensors on-board mobile and wearable devices to infer users' personal context and information is a well-known privacy threat that has received significant attention. Efforts towards protection mechanisms that prevent or limit the success of such threats, however, have been ad-hoc so far and have primarily focused on designing threat-specific customized defense mechanisms. Such approaches are not very practical, as evident from their limited adoption within major mobile/wearable operating systems. In the end, it is clear that all privacy threats that take advantage of unrestricted access to zero-permission sensors can be prevented if access to these sensors is regulated. However, due to the dynamic nature of sensor usage and requirements of different mobile applications, design of such access control mechanisms is not trivial. To effectively design an automated mobile defense mechanism that can dynamically measure the threat level of different sensor access requests from different applications and appropriately block suspicious requests, the problem of zero-permission sensor access needs to be first formally defined and analyzed. This paper accomplishes the above objective by employing game theory, specifically, signaling games, to analytically model the sensor access scenario for mobile applications, including, formalizing sensor access strategies of mobile applications and defense strategies of the on-board defense mechanism and the associated costs and benefits. Within the confines of a formal and practical game model, the paper then outlines conditions under which equilibria can be achieved between entities (applications and defense mechanism) with conflicting goals. The game model is further analyzed using numerical simulations, and also extended in the form of a repeated signaling game.

cs.GT↗

arXiv2018

A Game-Theoretic Analysis of Shard-Based Permissionless Blockchains

Low transaction throughput and poor scalability are significant issues in public blockchain consensus protocols such as Bitcoins. Recent research efforts in this direction have proposed shard-based consensus protocols where the key idea is to split the transactions among multiple committees (or shards), which then process these shards or set of transactions in parallel. Such a parallel processing of disjoint sets of transactions or shards by multiple committees significantly improves the overall scalability and transaction throughout of the system. However, one significant research gap is a lack of understanding of the strategic behavior of rational processors within committees in such shard-based consensus protocols. Such an understanding is critical for designing appropriate incentives that will foster cooperation within committees and prevent free-riding. In this paper, we address this research gap by analyzing the behavior of processors using a game-theoretic model, where each processor aims at maximizing its reward at a minimum cost of participating in the protocol. We first analyze the Nash equilibria in an N-player static game model of the sharding protocol. We show that depending on the reward sharing approach employed, processors can potentially increase their payoff by unilaterally behaving in a defective fashion, thus resulting in a social dilemma. In order to overcome this social dilemma, we propose a novel incentive-compatible reward sharing mechanism to promote cooperation among processors. Our numerical results show that achieving a majority of cooperating processors (required to ensure a healthy state of the blockchain network) is easier to achieve with the proposed incentive-compatible reward sharing mechanism than with other reward sharing mechanisms.

cs.GT↗

arXiv2018

Social Community-Aware Content Placement in Wireless Device-to-Device Communication Networks

In this paper, a novel framework for optimizing the caching of popular user content at the level of wireless user equipments (UEs) is proposed. The goal is to improve content offloading over wireless device-to-device (D2D) communication links. In the considered network, users belong to different social communities while their UEs form a single multi-hop D2D network. The proposed framework allows to exploit the multi-community social context of users for improving the local offloading of cached content in a multihop D2D network. To model the collaborative effect of a set of UEs on content offloading, a cooperative game between the UEs is formulated. For this game, it is shown that the Shapley value (SV) of each UE effectively captures the impact of this UE on the overall content offloading process. To capture the presence of multiple social communities that connect the UEs, a hypergraph model is proposed. Two line graphs, an influence-weighted graph, and a connectivity-weighted graph, are developed for analyzing the proposed hypergaph model. Using the developed line graphs along with the SV of the cooperative game, a precise offloading power metric is derived for each UE within a multi-community, multi-hop D2D network. Then, UEs with high offloading power are chosen as the optimal locations for caching the popular content. Simulation results show that, on the average, the proposed cache placement framework achieves 12%, 19%, and 21% improvements in terms of the number of UEs that received offloaded popular content compared to the schemes based on betweenness, degree, and closeness centrality, respectively.

cs.GT↗

arXiv2017

Stochastic Coalitional Games for Cooperative Random Access in M2M Communications

In this paper, the problem of random access contention between machine type devices (MTDs) in the uplink of a wireless cellular network is studied. In particular, the possibility of forming cooperative groups to coordinate the MTDs' requests for the random access channel (RACH) is analyzed. The problem is formulated as a stochastic coalition formation game in which the MTDs are the players that seek to form cooperative coalitions to optimize a utility function that captures each MTD's energy consumption and time-varying queue length. Within each coalition, an MTD acts as a coalition head that sends the access requests of the coalition members over the RACH. One key feature of this game is its ability to cope with stochastic environments in which the arrival requests of MTDs and the packet success rate over RACH are dynamically time-varying. The proposed stochastic coalitional is composed of multiple stages, each of which corresponds to a coalitional game in stochastic characteristic form that is played by the MTDs at each time step. To solve this game, a novel distributed coalition formation algorithm is proposed and shown to converge to a stable MTD partition. Simulation results show that, on the average, the proposed stochastic coalition formation algorithm can reduce the average fail ratio and energy consumption of up to 36% and 31% for a cluster-based distribution of MTDs, respectively, compared with a noncooperative case. Moreover, when the MTDs are more sensitive to the energy consumption (queue length), the coalitions' size will increase (decrease).

cs.GT↗