Defending Against Neural Network Model Inversion Attacks via Data Poisoning [0.03%]
基于数据投毒的神经网络模型逆向攻击防御方法研究
Shuai Zhou,Dayong Ye,Tianqing Zhu et al.
Shuai Zhou et al.
Model inversion attacks pose a significant privacy threat to machine learning models by reconstructing sensitive data from their outputs. While various defenses have been proposed to counteract these attacks, they often come at the cost of ...
Universal Approximation Theorem and Error Bounds for Quantum Neural Networks and Quantum Reservoirs [0.03%]
量子神经网络和量子 reservoirs 的通用逼近定理及误差界限制
Lukas Gonon,Antoine Jacquier
Lukas Gonon
Universal approximation theorems are the foundations of classical neural networks, providing theoretical guarantees that the latter are able to approximate maps of interest. Recent results have shown that this can also be achieved in a quan...
RBFleX-NAS: Training-Free Neural Architecture Search Using Radial Basis Function Kernel and Hyperparameter Detection [0.03%]
基于径向基函数核和超参数检测的无需训练的神经架构搜索方法(RBFleX-NAS)
Tomomasa Yamasaki,Zhehui Wang,Tao Luo et al.
Tomomasa Yamasaki et al.
Neural architecture search (NAS) is an automated technique to design optimal neural network architectures for a specific workload. Conventionally, evaluating candidate networks in NAS involves extensive training, which requires significant ...
Dynamic Erasing Network With Adaptive Temporal Modeling for Weakly Supervised Video Anomaly Detection [0.03%]
自适应时序建模的动态擦除网络：弱监督视频异常检测
Chen Zhang,Guorong Li,Yuankai Qi et al.
Chen Zhang et al.
The weakly supervised video anomaly detection aims to learn a detection model using only video-level labeled data. Prior studies ignore the complexity or duration of anomalies present in abnormal videos during temporal modeling. Moreover, e...
Luca Savant Aira,Diego Valsesia,Enrico Magli
Luca Savant Aira
We present stochastic Gaussian splatting (SGS): the first framework for uncertainty estimation using Gaussian splatting (GS). GS recently advanced the novel-view synthesis field by achieving impressive reconstruction quality at a fraction o...
Probably Approximately Correct Bayes Meta-Learning With Parameterized-Bounded Guarantees [0.03%]
具有参数化有界保证的近似贝叶斯元学习方法
Zhewei Zhang,Yujun Cheng,Junyu Shen et al.
Zhewei Zhang et al.
In meta-learning, the learner extracts knowledge from the observed tasks and quickly adapts to unseen future tasks. We provide a novel and rigorous-analyzed probably approximately correct Bayes (PAC-Bayes) meta-learning method with paramete...
Broad Critic Deep Actor Reinforcement Learning for Continuous Control [0.03%]
基于广义评论家和深度行动人的连续控制的强化学习
Shiron Thalagala,Pak Kin Wong,Xiaozheng Wang et al.
Shiron Thalagala et al.
In the domain of continuous control, deep reinforcement learning (DRL) demonstrates promising results. However, the dependence of DRL on deep neural networks (DNNs) results in the demand for extensive data and increased computational cost. ...
Online Graph Models: Tackling the Challenges of Non-Gaussian Noise in Adaptive Filtering [0.03%]
在线图模型：适应滤波中非高斯噪声挑战的图模型方法研究
Shan Zhong,Gang Wang,Kah Chan Teh et al.
Shan Zhong et al.
Adaptive filtering faces significant challenges in handling complex non-Gaussian noise, while graph signal processing (GSP) excels at processing data with intricate structures. This brief introduces a novel method for solving non-Gaussian n...
CoLeCLIP: Open-Domain Continual Learning via Joint Task Prompt and Vocabulary Learning [0.03%]
通过联合任务提示和词汇学习实现开放式领域连续学习（CoLeCLIP）
Yukun Li,Guansong Pang,Wei Suo et al.
Yukun Li et al.
This article investigates the problem of continual learning (CL) of vision-language models (VLMs) in open domains, where models are required to perform continual updating and inference on a stream of datasets from diverse seen and unseen do...
Shenglun Chen,Xinzhu Ma,Hong Zhang et al.
Shenglun Chen et al.
As a key problem in computer vision, depth completion aims to recover dense depth maps from sparse ones [generally derived from light detection and ranging (LiDAR)]. Most methods introduce synchronous RGB images and leverage multimodal fusi...