The metaverse is a unified, persistent, and shared multi-user virtual environment with fully immersive, hyper-temporal, diverse interconnected network. When combined healthcare, it can effectively improve medical services has great potential for development in realizing training, enhanced teaching, remote surgical treatment. provides immersive users through massive multimodal data, its data scale growth rate are bound to show exponential growth. Blockchain-based distributed storage...

Evaluation of biases in language models is often limited to synthetically generated datasets. This dependence traces back the need prompt-style dataset trigger specific behaviors models. In this paper, we address gap by creating a prompt with respect occupations collected from real-world natural sentences present Wikipedia.We aim understand differences between using template-based prompts and sentence when studying gender-occupation We find bias evaluations are very sensitiveto design...

Multi-modal learning with both text and images benefits multiple applications, such as attribute extraction for e-commerce products. In this paper, we propose Cross-Modality Attention Contrastive Language-Image Pre-training (CMA-CLIP), a new multi-modal architecture to jointly learn the fine-grained inter-modality relationship. It fuses CLIP sequence-wise attention module modality-wise module. The network uses bridge gap at global level, capture alignment between images. Besides, it...

Content distribution networks (CDNs) improve the performance of content delivery by replicating popular on surrogate servers deployed at edge Internet. The CDN-P2P architecture, which combines complementary advantages both CDN and P2P networks, can quality service (QoS). In this paper, we propose a k-coordinated decentralized replica placement algorithm (DRPA) based gain formulation problem. Although is designed for different types focus robust ring-based architecture in study. our approach,...

Current blockchains are restricted by the low throughput. Aimed at this problem, we propose Txilm, a protocol that compresses size of transaction presentation in each block to save bandwidth network. In protocol, carries short hashes TXIDs instead complete transactions. Combined with sorted transactions based on TXIDs, Txilm realizes 80 times data reduction compared original blockchains. We also evaluate probability hash collisions, and provide methods resolving such collisions. Finally,...

This paper presents Natural Evolution Strategies (NES), a recent family of algorithms that constitute more principled approach to black-box optimization than established evolutionary algorithms. NES maintains parameterized distribution on the set solution candidates, and natural gradient is used update distribution's parameters in direction higher expected fitness. We introduce collection techniques address issues convergence, robustness, sample complexity, computational complexity...

As ML models have increased in capabilities and accuracy, so has the complexity of their deployments. Increasingly, model consumers are turning to service providers serve ML-as-a-service (MLaaS) paradigm. MLaaS proliferates, a critical requirement emerges: how can verify that correct predictions were served, face malicious, lazy, or buggy providers? In this work, we present first practical ImageNet-scale method inference non-interactively, i.e., after been done. To do so, leverage recent...

This paper presents a novel algorithm for gesture recognition and identity authentication based on continuous hidden Markov model (CHMM) optical flow method. study aims to solve the information security problems about ubiquitous electric power internet of things. In this system, method is used segment extract features preprocessed dynamic obtain motion track, CHMM chosen establish valid user model, which leads ensuring gestures are accurately recognised. The proposed test recognise result...

Scedosporium and Lomentospora infections in humans are generally chronic stubborn. The use of azoles alone cannot usually inhibit the growth these fungi. To further explore combined effect multiple drugs potential mechanisms action, we tested antifungal effects tacrolimus (FK506) everolimus combination with vitro vivo on 15 clinical strains / species detected level Rhodamine 6G, ROS activity, apoptosis. results showed that combinations itraconazole, voriconazole, posaconazole synergistic 9...

Cyber security visualization designers can benefit from human factors engineering concepts and principles to resolve key challenges in visual interface design. We survey that have been applied the past decade of research. highlight these relate them cybersecurity provide guidelines help address some context use ecological design approach present factors-based for visualization. will

Zero-knowledge (ZK) proof systems have emerged as a promising solution for building security-sensitive applications. However, bugs in ZK applications are extremely difficult to detect and can allow malicious party silently exploit the system without leaving any observable trace. This paper presents Coda, novel statically-typed language zero-knowledge Critically, Coda makes it possible formally specify statically check properties of application through rich refinement type system. One key...

In E-commerce search, spelling correction plays an important role to find desired products for customers in processing user-typed search queries. However, resolving phonetic errors is a critical but much overlooked area. The query with tends appear correct based on pronunciation nonetheless inaccurate (e.g., "bluetooth sound system" vs. "blutut sant sistam") numerous noisy forms and sparse occurrences. this work, we propose generalized system integrating phonetics address without additional...

Many bioimage analysis projects produce quantitative descriptors of regions interest in images. Associating these with visual characteristics the objects they describe is a key step understanding data at hand. However, as many and their workflows are moving to cloud, addressing interactive exploration remote environments has become pressing issue. To address it, we developed Image Data Explorer (IDE) web application that integrates linked visualization images derived points exploratory...

Over the past few years, AI methods of generating images have been increasing in capabilities, with recent breakthroughs enabling high-resolution, photorealistic "deepfakes" (artificially generated purpose misinformation or harm). The rise deepfakes has potential for social disruption. Recent work proposed using ZK-SNARKs (zero-knowledge succinct non-interactive argument knowledge) and attested cameras to verify that were taken by a camera. allow verification image transformations...

Autonomous ground robots autonomously are being used inthe places where it is very hazardous for human beings to reach and operate,such as nuclear power plants chemical industries. The aim of the researchpresented here develop a control system that enables such robotsnavigate with various sensors depth camera, 2D scanninglaser, 3D Lidar, GPS, IMU. controller uses current positionmeasured using on Husky A200, given waypoints thedestination. Then calculates best possible route based...

Project 6 during the 2023 BioHackathon Europe in Barcelona focused on "Enhancing image analysis community Galaxy." Despite Galaxy's strong presence genomics and proteomics, its tools workflows are currently scattered. This project aimed to gather efforts across fields build a robust interdisciplinary community.

Metaphor comprehension is an important component in the process of constructing metaphorical thinking. Therefore, topic “metaphor comprehension” should be a crucial part metaphor studies. This study uses Bibliometrix to conduct bibliometric research on from 2013 2023, focusing knowledge structure, hot spots, and fronts. The co-citation analysis shows that Contemporary Theory (CTM) Cognitive Linguistics has provided dominant foundation for comprehension. A large number hypotheses CTM not only...

The grasp generation of dexterous hand often requires a large number grasping annotations. Especially for functional grasp-requiring the pose to be convenient subsequent use object. However, annotating high DoF is rather challenging. This prompt us explore how people achieve manipulations on new objects based past experiences. We find that are adept at discovering and leveraging various similarities between when items, including shape, layout, type. In light this, we analyze collect...