The proceedings contain 535 papers. The topics discussed include: IMTCDF: a multi-module-based internet malicious traffic classification and detection framework;multi-behavior enhanced self-supervised graph learning f...
ISBN:
(纸本)9798350349184
The proceedings contain 535 papers. The topics discussed include: IMTCDF: a multi-module-based internet malicious traffic classification and detection framework;multi-behavior enhanced self-supervised graph learning for social recommendation;LFAS: an electricity load forecasting framework assisted by cooperative multi-task learning-based spike occurrence prediction;significance-aware medication recommendation with medication representation learning;facilitating pornographic text detection for open-domain dialogue systems via knowledge distillation of large language models;heart of betrayal: a PIN inference attack leveraging photoplethysmography on wearables;improving distantly-supervised relation extraction through label prompt;intelligent convergence: advancing 6G mmWave networks with RIS-enhanced ISAC for IoT applications;short-term portfolio optimization using doubly regularized exponential growth rate;an interpretable adversarial robustness evaluation method based on robust paths;and cooperative simulation of RDMA-based network and storage.
The proceedings contain 321 papers. The topics discussed include: an intelligent human-agent interaction support system in medicine;dynamic privacy-enhanced federated learning algorithm based on gradients perturbation...
ISBN:
(纸本)9798350331684
The proceedings contain 321 papers. The topics discussed include: an intelligent human-agent interaction support system in medicine;dynamic privacy-enhanced federated learning algorithm based on gradients perturbation;mapping tokenomics arrangements to expand the digital nomad ecosystem;intrusion detection based on sampling and improved ova technique on imbalanced data;logic event graph enhanced narrative generation;information diffusion prediction via exploiting cascade relationship diversity;multi-modal medical image classification method combining graph convolution neural networks;an edge server co-deployment method via joint optimization of communication delay and load balancing in edge collaborative environment;a surface defect detection method based on information entropy;using rough sets to improve the high-dimensional data anomaly detection method based on extended isolation forest;a multi-granularity knowledge-enhanced sequential recommendation;research on coarse granularity data sample completion method for district heating system;and an agent-based system architecture for automated guided vehicles in cloud-edge computing environments.
The proceedings contain 220 papers. The topics discussed include: zero-shot learning based on knowledge sharing;FenceBot: an elderly tracking app for mitigating health risk contacts;latent clues investigation in disci...
ISBN:
(纸本)9781728165974
The proceedings contain 220 papers. The topics discussed include: zero-shot learning based on knowledge sharing;FenceBot: an elderly tracking app for mitigating health risk contacts;latent clues investigation in discipline inspection and supervision based on knowledge graph analysis;study on typical objects three-dimensional modeling and classification technology based on UAV image sequence;a BERT based sentiment analysis and key entity detection approach for online financial texts;optimal open space cubicle assignment considering personal thermal preferences;MidiPGAN: a progressive GAN approach to MIDI generation;and an orthogonal learning design whale optimization algorithm with clustering mechanism.
Graphic Layout Generation focuses on providing layout references for various visual design tasks, such as advertisement design and poster design. Researchers have conducted exploratory studies from different dimension...
详细信息
ISBN:
(纸本)9798350349184;9798350349191
Graphic Layout Generation focuses on providing layout references for various visual design tasks, such as advertisement design and poster design. Researchers have conducted exploratory studies from different dimensions, such as human-computer interaction and layout representation. With the success of deep learning techniques, there has been a recent surge in research on graphic layout generation using deep generative models. Yet there are still no comprehensive studies on graphic layout generation. In this paper, we review methods for graphic layout generation from two perspectives: implementation and interactivity. We analyze the current status, the advantageous application scenarios, and the challenges of different methods. Based on the analysis results, we summarize the workflow of designers collaborating with graphic automatic layout systems and propose potential directions for future research.
In the realm of virtual object manipulation, multitouch solutions have been explored since the early days of tabletop gesture interactions. However, with the advent of virtual and mixed reality, the exploration of tou...
详细信息
ISBN:
(纸本)9798350349184;9798350349191
In the realm of virtual object manipulation, multitouch solutions have been explored since the early days of tabletop gesture interactions. However, with the advent of virtual and mixed reality, the exploration of touch-based solutions has predominantly shifted towards gesture design and input mediums like wearables, overlooking optimization for user experiences in immersive XR environments. In this paper, we introduce a novel approach by collaborating a smartwatch with AR glasses, considering both input device posture and the user's viewpoint, to implement five levels of exploration solutions. Through user experiments, we compare and identify the two most prominent ones and summarize factors influencing users' perception in manipulating virtual objects. These will promote more targeted interaction design and experimental research.
With the wide application of deep neural networks, the security problem of the model is becoming more prominent, adversarial attack is an important tool for evaluating the robustness and security of the model, adversa...
详细信息
ISBN:
(纸本)9798350349184;9798350349191
With the wide application of deep neural networks, the security problem of the model is becoming more prominent, adversarial attack is an important tool for evaluating the robustness and security of the model, adversarial attack can be categorized into white-box and black-box attacks. Aiming at the problem of huge perturbation and the low success rate of the adversarial example created in the transfer attack in the black box, a local region approach for randomly segmented channels is given. By randomly segmenting individual dimensions in order to improve the transferability of the adversarial example, the ScoreCAM method is introduced to extract localized focus regions of the image to generate the adversarial example. Experiments show that the performance of the method is better than the baseline algorithm;the fooling rate improved by up to 17%;and the average 2-norm module length decreased by 54.2%.
In the realm of intelligent transportation systems, ensuring the safety of vulnerable road users (VRU) has always been a critical concern. To further enhance the safety of VRU, this paper proposes a bi-directional col...
详细信息
ISBN:
(纸本)9798350349184;9798350349191
In the realm of intelligent transportation systems, ensuring the safety of vulnerable road users (VRU) has always been a critical concern. To further enhance the safety of VRU, this paper proposes a bi-directional collaborative perception method for VRU and vehicles, which utilizes Road Side Units (RSU) as intermediaries to transmit intention information between them. By sharing intention information that represents future behavioral intentions of objects, network members calculate the intention conflicts between vehicles and VRUs. While vehicles actively avoid conflicts by utilizing intention information, RSUs transmit conflict warning messages to VRUs to assist them in actively avoiding conflicts, thus protecting the safety of VRU. Finally, the proposed scheme is implemented and evaluated on a real dataset, and finds that the proposed scheme has improved the VRU perception rate, perception efficiency, and safety protection by 44%, 19%, and 37%, respectively, compared to traditional schemes.
The variation of Link quality in underwater sensor networks often lead to link interruption and void-hole. To address the above issues, some protocols chose nodes with good link quality as relay nodes without consider...
详细信息
ISBN:
(纸本)9798350349184;9798350349191
The variation of Link quality in underwater sensor networks often lead to link interruption and void-hole. To address the above issues, some protocols chose nodes with good link quality as relay nodes without considering fairness, while others used multipath and greedy forwarding which cause high energy consumption. Therefore, this article proposes a fair and energy efficient cooperative routing protocol based On link quality for underwater acoustic sensor network (FECRP). This protocol divides underwater sensor networks into clusters, selects cluster heads and relay nodes based on perceived link quality, it cooperatives with MAC protocol in the data transmission within the cluster, scheduling the repeated transmission of data packets to overcome link interruptions. The simulation results show that this protocol improves fairness and reduces energy consumption while ensuring throughput.
With the increasing popularity of e -commerce, product poster design plays a vital role. To meet the demand for fast, efficient, and diversified design needs, automated poster design has become one of the important re...
详细信息
ISBN:
(纸本)9798350349184;9798350349191
With the increasing popularity of e -commerce, product poster design plays a vital role. To meet the demand for fast, efficient, and diversified design needs, automated poster design has become one of the important research directions. This work investigates the stylized generation of background graphic design based on light elements. Particularly, the research summarizes the characteristics and functions of light elements used in poster design, determines the generation parameters and rules of light elements referring to the principles of photographic lighting, then matches the color of light elements with the theme color of the product, and identifies the generated backgrounds best matching users' semantic needs. Particularly, BP neural network was employed to realize the selection of the high-quality background designs. Based on the above procedures, a platform of poster generative design system based on light elements was developed. To validate the performance of the design platform, usability tests and comparative experiments were conducted, and the results show that the generation approach can better improve the poster design quality and efficiency.
The rise of artificial intelligence agents and information technology platforms governed by data-driven algorithms has had profound impacts on the way knowledge work is managed and carried out. Concerning the novel us...
详细信息
ISBN:
(纸本)9798350349184;9798350349191
The rise of artificial intelligence agents and information technology platforms governed by data-driven algorithms has had profound impacts on the way knowledge work is managed and carried out. Concerning the novel uses of digital work platforms, the arising tensions might suggest that the socio-technical design frameworks adopted by major tech firms are becoming less viable and in some cases ineffective. We propose a conceptual architecture to support a socio-technical framework for exploring Personal Knowledge Ecologies in the context of the digital knowledge economy. This framework aims to increase the degree of individual freedom for the interoperation of these knowledge information infrastructures.
暂无评论