The exchange of knowledge is widely recognized as a crucial aspect of effective knowledge management. When it comes to sharing knowledge within Prison settings, things get complicated due to various challenges such as...
详细信息
Severe Doppler shift and the obstruction of the line-of-sight (LoS) path significantly decrease the communication performance. This article considers a downlink channel estimation problem for reconfigurable intelligen...
详细信息
Uncrewed Aerial Vehicles (UAVs) have enabled key duties in emergency preparedness, traffic monitoring, environmental monitoring, and public safety. Since the presence of GPS-enabled contexts is not always guaranteed, ...
详细信息
In this paper, we introduce a novel class of fast, beam search-based adversarial attack (BEAST) for Language Models (LMs). BEAST employs interpretable parameters, enabling attackers to balance between attack speed, su...
详细信息
In this paper, we introduce a novel class of fast, beam search-based adversarial attack (BEAST) for Language Models (LMs). BEAST employs interpretable parameters, enabling attackers to balance between attack speed, success rate, and the readability of adversarial prompts. The computational efficiency of BEAST facilitates us to investigate its applications on LMs for jailbreaking, eliciting hallucinations, and privacy attacks. Our gradient-free targeted attack can jailbreak aligned LMs with high attack success rates within one minute. For instance, BEAST can jailbreak Vicuna-7B-v1.5 under one minute with a success rate of 89% when compared to a gradient-based baseline that takes over an hour to achieve 70% success rate using a single Nvidia RTX A6000 48GB GPU. BEAST can also generate adversarial suffixes for successful jailbreaks that can transfer to unseen prompts and unseen models such as GPT-4-Turbo. Additionally, we discover a unique outcome wherein our untargeted attack induces hallucinations in LM chatbots. Through human evaluations, we find that our untargeted attack causes Vicuna-7B-v1.5 to produce ∼15% more incorrect outputs when compared to LM outputs in the absence of our attack. We also learn that 22% of the time, BEAST causes Vicuna to generate outputs that are not relevant to the original prompt. Further, we use BEAST to generate adversarial prompts in a few seconds that can boost the performance of existing membership inference attacks for LMs. We believe that our fast attack, BEAST, has the potential to accelerate research in LM security and privacy. Copyright 2024 by the author(s)
Controlling an active distribution network(ADN)from a single PCC has been advantageous for improving the performance of coordinated Intermittent RESs(IRESs).Recent studies have proposed a constant PQ regulation approa...
详细信息
Controlling an active distribution network(ADN)from a single PCC has been advantageous for improving the performance of coordinated Intermittent RESs(IRESs).Recent studies have proposed a constant PQ regulation approach at the PCC of ADNs using coordination of non-MPPT based ***,due to the intermittent nature of DGs coupled with PCC through uni-directional broadcast communication,the PCC becomes vulnerable to transient *** address this challenge,this study first presents a detailed mathematical model of an ADN from the perspective of PCC regulation to realize rigidness of PCC against ***,an H_(∞)controller is formulated and employed to achieve optimal performance against disturbances,consequently,ensuring the least oscillations during transients at ***,an eigenvalue analysis is presented to analyze convergence speed limitations of the newly derived system ***,simulation results show the proposed method offers superior performance as compared to the state-of-the-art methods.
Population growth in cities results in a demand for parking lots from an increasing number of automobiles, which frequently contributes to the global problem of traffic congestion. This study presents the smart parkin...
详细信息
In recent advancements within wireless sensor networks(WSN),the deployment of unmanned aerial vehicles(UAVs)has emerged as a groundbreaking strategy for enhancing routing efficiency and overall network *** research in...
详细信息
In recent advancements within wireless sensor networks(WSN),the deployment of unmanned aerial vehicles(UAVs)has emerged as a groundbreaking strategy for enhancing routing efficiency and overall network *** research introduces a sophisticated framework,driven by computational intelligence,that merges clustering techniques with UAV mobility to refine routing strategies in *** proposed approach divides the sensor field into distinct sectors and implements a novel weighting system for the selection of cluster heads(CHs).This system is primarily aimed at reducing energy consumption through meticulously planned routing and path *** a greedy algorithm for inter-cluster dialogue,our framework orchestrates CHs into an efficient communication *** comparative analysis,the proposed model demonstrates a marked improvement over traditional methods such as the cluster chain mobile agent routing(CCMAR)and the energy-efficient cluster-based dynamic algorithms(ECCRA).Specifically,it showcases an impressive 15%increase in energy conservation and a 20%reduction in data transmission time,highlighting its advanced ***,this paper investigates the impact of various network parameters on the efficiency and robustness of the WSN,emphasizing the vital role of sophisticated computational strategies in optimizing network operations.
Wireless communications are particularly vulnerable to eavesdropping attacks due to their broadcast nature. To effectively deal with eavesdroppers, existing security techniques usually require accurate channel state i...
详细信息
In order to address the convergence issues faced by traditional multi-objective optimization algorithms as the number of objectives increases, this paper proposes a new optimization algorithm, MaOEA-SPC, based on the ...
详细信息
Prenatal depression,which can affect pregnant women’s physical and psychological health and cause postpartum depression,is increasing ***,it is essential to detect prenatal depression early and conduct an attribution...
详细信息
Prenatal depression,which can affect pregnant women’s physical and psychological health and cause postpartum depression,is increasing ***,it is essential to detect prenatal depression early and conduct an attribution *** studies have used questionnaires to screen for prenatal depression,but the existing methods lack *** diagnose the early signs of prenatal depression and identify the key factors that may lead to prenatal depression from questionnaires,we present the semantically enhanced option embedding(SEOE)model to represent questionnaire *** can quantitatively determine the relationship and patterns between options and *** first quantifies options and resorts them,gathering options with little difference,since Word2Vec is highly dependent on *** resort task is transformed into an optimization problem involving the traveling salesman ***,all questionnaire samples are used to train the options’vector using ***,an LSTM and GRU fused model incorporating the cycle learning rate is constructed to detect whether a pregnant woman is suffering from *** verify the model,we compare it with other deep learning and traditional machine learning *** experiment results show that our proposed model can accurately identify pregnant women with depression and reach an F1 score of *** most relevant factors of depression found by SEOE are also verified in the *** addition,our model is of low computational complexity and strong generalization,which can be widely applied to other questionnaire analyses of psychiatric disorders.
暂无评论