Malicious network attacks are a serious source of concern, and machine learning techniques are widely used to build Attack Detectors with off-line training with real attack and non-attack data, and used online to moni...
详细信息
ISBN:
(纸本)9781665455800
Malicious network attacks are a serious source of concern, and machine learning techniques are widely used to build Attack Detectors with off-line training with real attack and non-attack data, and used online to monitor system entry points connected to networks. Many machine learning based Attack Detectors are typically trained to identify specific types attacks, and the training of such algorithms to cover several types of attacks may be excessively time consuming. This paper shows that G-networks, which are queueing networks with product form solution and special customers such as negative customers and triggers, can be trained just with non-attack traffic, can accurately detect several different attack types. This is established with a special case of G-networks with triggerred customer movement. A DARPA attack and non-attack traffic repository is used to train and test the the G-Network, yielding comparable or clearly better accuracy than most known attack detection techniques.
Federated Learning (FL) is an emerging privacy-preserving distributed machine learning paradigm that enables numerous clients to collaboratively train a global model without transmitting private datasets to the FL ser...
详细信息
Deep Learning (DL)-based solutions have shown promising performance in assessing neonatal pain. However, the occlusion of the visual modality (face and body) is common in clinical settings due to several factors, incl...
详细信息
ISBN:
(纸本)9798350312249
Deep Learning (DL)-based solutions have shown promising performance in assessing neonatal pain. However, the occlusion of the visual modality (face and body) is common in clinical settings due to several factors, including a prone sleeping position, low light, or swaddling. In such scenarios, other pain signals, such as audio signals, can be used as the major behavioral signs of pain. Although DL-based methods are proposed to assess pain from audio, these methods lack transparency and explainability (black box), which can decrease the user's trust in the automated decision. In this work, we visualize the neonate's audio signal as a spectrogram image to classify it as pain or no pain and present an instance-based approach for explaining the decision of the black-box model. Further, this work provides an analysis of the most helpful and harmful training instances using an influence score followed by assessing their impact on pain prediction. Experimental results demonstrate that the proposed approach can detect and remove harmful instances, eventually leading to a compressed dataset. Our results also show that the proposed work can add explainability to the current DL-based pain detection methods, which can enhance users' trust and provide a viable approach toward pain assessment in clinical settings.
Traffic congestion represents a daunting challenge for all facets of urban development, as well as represents a universal problem in all urban areas, to various extents. In recent years, many cities have adopted the u...
详细信息
ISBN:
(纸本)9798350368130
Traffic congestion represents a daunting challenge for all facets of urban development, as well as represents a universal problem in all urban areas, to various extents. In recent years, many cities have adopted the use of Intelligent Transport systems (ITS) to manage traffic congestion. These systems are indeed useful, but they are mainly geared to predict real-time traffic congestion, yielding to a certain short-sightedness in our prediction model. Through the use of Time Series Analysis, and Regression models for traffic congestion prediction, we posit that we can address the issue. The data that these models were trained on derives fro mtwo datasets, namely a dataset from Kaggle and another from the road traffic footage of Tirana. The data then points to Gated Recurrent Units (GRU) being a more accurate time series category, and Support Vector Regression (SVR) to be a better performer than linear regression.
We explore how geometric structures (or shapes) can be grown exponentially fast from a single node, through a sequence of centralized growth operations, and if collisions during growth are to be avoided. We identify a...
详细信息
The proceedings contain 142 papers. The topics discussed include: AI/ML model training in O-RAN: assessing datasets for hoarding and choosing strategies;DRL-based transmission design for distributed STAR-RIS-aided com...
ISBN:
(纸本)9798350307672
The proceedings contain 142 papers. The topics discussed include: AI/ML model training in O-RAN: assessing datasets for hoarding and choosing strategies;DRL-based transmission design for distributed STAR-RIS-aided communications;digital twin modelling of cascaded amplifiers in the COSMOS testbed;communication-efficient federated learning for real-time applications in edge networks;LSTM-based channel estimator for optical IRS-assisted non-linear VLC systems;3GPP-based testbed for edge computing;architecture, implementation and application deployment;self-organized drone placement for the surveillance of uneven surface;on performance evaluation of channel codes over satellite channels;implementation and performance analysis of NOMA on software defined radio;and lock-in amplifier based alignment for free space optical communication.
Underwater wireless sensor networks can monitor ocean information, which provides a new approach to marine environmental monitoring, disaster warning and resource exploration. However, the development of underwater wi...
详细信息
The capacity to analyze the causes of poor decisions made by visual recognition models is becoming increasingly crucial as the security requirements of various real-world systems continue to escalate. However, the com...
详细信息
ISBN:
(纸本)9798350359329;9798350359312
The capacity to analyze the causes of poor decisions made by visual recognition models is becoming increasingly crucial as the security requirements of various real-world systems continue to escalate. However, the complex structure and blackbox nature within deep neural networks constrain the mining of their error causes. Based on this, we propose a conceptbased (e.g. a group of pixel blocks that contain leaves represents the concept of leaves) automated strong localization interpretation framework, called hierarchically optimized concept-sensitive interpretation (HOCS), to provide quantitative analysis of the semantics of wrong decisions in the classification network is provided from two directions of internal and external information interference of samples. HOCS was applied to models with spurious correlation and well-distributed data in the training set. The results showed that it provided concrete explanations in a way that was understandable to humans and demonstrated the significant advantages of HOCS in terms of efficiency and accuracy.
With the continuous progress of computer technology, distributed intelligent systems become more and more popular. Based on this background, this paper discusses the theory of image recognition based on distributed in...
详细信息
The proceedings contain 44 papers. The special focus in this conference is on internationalsymposium on Stabilization,Safety, and Security of distributedsystems. The topics include: Improved Paths to Stability ...
ISBN:
(纸本)9783031442735
The proceedings contain 44 papers. The special focus in this conference is on internationalsymposium on Stabilization,Safety, and Security of distributedsystems. The topics include: Improved Paths to Stability for the Stable Marriage Problem;lattice Linearity of Multiplication and Modulo;the Fagnano Triangle Patrolling Problem (Extended Abstract);invited Paper: Monotonicity and Opportunistically-Batched Actions in Derecho;robust Overlays Meet Blockchains: On Handling High Churn and Catastrophic Failures;disconnected Agreement in networks Prone to Link Failures;where Are the Constants? New Insights on the Role of Round Constant Addition in the SymSum Distinguisher;invited Paper: Detection of False Data Injection Attacks in Power systems Using a Secured-Sensors and Graph-Based Method;KerberSSIze Us: Providing Sovereignty to the People;model Checking of distributed Algorithms Using Synchronous Programs;hierarchical Identity-Based Inner Product Functional Encryption for Unbounded Hierarchical Depth;brief Announcement: Efficient Probabilistic Approximations for Sign and Compare;meeting Times of Non-atomic Random Walks;minimum Algorithm Sizes for Self-stabilizing Gathering and Related Problems of Autonomous Mobile Robots (Extended Abstract);separation of Unconscious Colored Robots;forbidden Patterns in Temporal Graphs Resulting from Encounters in a Corridor;uniform k-Circle Formation by Fat Robots;brief Announcement: Rendezvous on a Known Dynamic Point in a Finite Unoriented Grid;brief Announcement: Crash-Tolerant Exploration by Energy Sharing Mobile Agents;time-Optimal Geodesic Mutual Visibility of Robots on Grids Within Minimum Area;the Fence Complexity of Persistent Sets;privacy in Population Protocols with Probabilistic Scheduling;dispersion of Mobile Robots in Spite of Faults;brief Announcement: Asynchronous Gathering of Finite Memory Robots on a Circle Under Limited Visibility;wait-Free Updates and Range Search Using Uruv;stand-Up Indulgent Gathering on Lines;offl
暂无评论