检索结果-内蒙古大学图书馆

Positive and unlabeled learning with controlled probability boundary fence 24

学校读者我要写书评

暂无评论

Positive and unlabeled learning with controlled probability ...

Proceedings of the 41st International Conference on Machine Learning

作者： Changchun Li Yuanchao Dai Lei Feng Ximing Li Bing Wang Jihong Ouyang College of Computer Science and Technology and Key Laboratory of Symbolic Computation and Knowledge Engineering of Ministry of Education Jilin University China Information Systems Technology and Design Pillar Singapore University of Technology and Design Singapore

Positive and Unlabeled (PU) learning refers to a special case of binary classification, and technically, it aims to induce a binary classifier from a few labeled positive training instances and loads of unlabeled instances. In this paper, we derive a theorem indicating that the probability boundary of the asymmetric disambiguation-free expected risk of PU learning is controlled by its asymmetric penalty, and we further empirically evaluated this theorem. Inspired by the theorem and its empirical evaluations, we propose an easy-to-implement two-stage PU learning method, namely Positive and Unlabeled Learning with Controlled Probability Boundary Fence (PUL-CPBF). In the first stage, we train a set of weak binary classifiers concerning different probability boundaries by minimizing the asymmetric disambiguation-free empirical risks with specific asymmetric penalty values. We can interpret these induced weak binary classifiers as a probability boundary fence. For each unlabeled instance, we can use the predictions to locate its class posterior probability and generate a stochastic label. In the second stage, we train a strong binary classifier over labeled positive training instances and all unlabeled instances with stochastic labels in a self-training manner. Extensive empirical results demonstrate that PUL-CPBF can achieve competitive performance compared with the existing PU learning baselines.

关键词：

Learning Transferable Negative Prompts for Out-of-Distribution Detection

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Li, Tianqi Pang, Guansong Bai, Xiao Miao, Wenjun Zheng, Jin School of Computer Science and Engineering State Key Laboratory of Complex & Critical Software Environment Jiangxi Research Institute Beihang University China School of Computing and Information Systems Singapore Management University Singapore

Existing prompt learning methods have shown certain capabilities in Out-of-Distribution (OOD) detection, but the lack of OOD images in the target dataset in their training can lead to mismatches between OOD images and In-Distribution (ID) categories, resulting in a high false positive rate. To address this issue, we introduce a novel OOD detection method, named ‘NegPrompt’, to learn a set of negative prompts, each representing a negative connotation of a given class label, for delineating the boundaries between ID and OOD images. It learns such negative prompts with ID data only, without any reliance on external outlier data. Further, current methods assume the availability of samples of all ID classes, rendering them ineffective in open-vocabulary learning scenarios where the inference stage can contain novel ID classes not present during training. In contrast, our learned negative prompts are transferable to novel class labels. Experiments on various ImageNet benchmarks show that NegPrompt surpasses state-of-the-art prompt-learning-based OOD detection methods and maintains a consistent lead in hard OOD detection in closed- and open-vocabulary classification scenarios. Code is available at https://***/mala-lab/negprompt. © 2024, CC BY.

关键词： Chemical detection

UR4NNV: Neural Network Verification, Under-approximation Reachability Works!

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Liang, Zhen Wu, Taoran Zhao, Ran Xue, Bai Wang, Ji Yang, Wenjing Deng, Shaojun Liu, Wanwei College of Computer Science and Technology National University of Defense Technology China Institute of Software Chinese Academy of Sciences China School of Computer Science and Technology University of Chinese Academy of Sciences China Key Laboratory of Software Engineering for Complex Systems National University of Defense Technology China

Recently, formal verification of deep neural networks (DNNs) has garnered considerable attention, and over-approximation based methods have become popular due to their effectiveness and efficiency. However, these strategies face challenges in addressing the "unknown dilemma" concerning whether the exact output region or the introduced approximation error violates the property in question. To address this, this paper introduces the UR4NNV verification framework, which utilizes under-approximation reachability analysis for DNN verification for the first time. UR4NNV focuses on DNNs with Rectified Linear Unit (ReLU) activations and employs a binary tree branch-based under-approximation algorithm. In each epoch, UR4NNV under-approximates a sub-polytope of the reachable set and verifies this polytope against the given property. Through a trial-and-error approach, UR4NNV effectively falsifies DNN properties while providing confidence levels when reaching verification epoch bounds and failing falsifying properties. Experimental comparisons with existing verification methods demonstrate the effectiveness and efficiency of UR4NNV, significantly reducing the impact of the "unknown dilemma".MSC Codes 68Q60, 68T07 Copyright © 2024, The Authors. All rights reserved.

关键词： Deep neural networks

Towards Efficient Deep Spiking Neural Networks Construction with Spiking Activity based Pruning

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Li, Yaxin Xu, Qi Shen, Jiangrong Xu, Hongming Chen, Long Pan, Gang School of Computer Science and Technology Dalian University of Technology Dalian China Guangdong Laboratory of Artificial Intelligence and Digital Economy Shenzhen China College of Computer Science and Technology Zhejiang University Hangzhou China Centre for Systems Engineering and Innovation Imperial College London London United Kingdom

The emergence of deep and large-scale spiking neural networks (SNNs) exhibiting high performance across diverse complex datasets has led to a need for compressing network models due to the presence of a significant number of redundant structural units, aiming to more effectively leverage their low-power consumption and biological interpretability advantages. Currently, most model compression techniques for SNNs are based on unstructured pruning of individual connections, which requires specific hardware support. Hence, we propose a structured pruning approach based on the activity levels of convolutional kernels named Spiking Channel Activity-based (SCA) network pruning framework. Inspired by synaptic plasticity mechanisms, our method dynamically adjusts the network’s structure by pruning and regenerating convolutional kernels during training, enhancing the model’s adaptation to the current target task. While maintaining model performance, this approach refines the network architecture, ultimately reducing computational load and accelerating the inference process. This indicates that structured dynamic sparse learning methods can better facilitate the application of deep SNNs in low-power and high-efficiency scenarios. Copyright © 2024, The Authors. All rights reserved.

关键词： Network architecture

2021 Monthly Rice Production in Chinese Coastal Provinces 3

学校读者我要写书评

暂无评论

2021 Monthly Rice Production in Chinese Coastal Provinces

3rd International Conference on Big Data Analytics and Practices, IBDAP 2022

作者： Parmar, Ajeet Munasinghe, Thilanka Tubbs, Heidi Anyamba, Assaf Rensselaer Polytechnic Institute Department of Electrical Computer and Systems Engineering Troy United States Rensselaer Polytechnic Institute Department of Information Technology and Web Science Troy United States Oak Ridge National Laboratory Geospatial Science and Human Security Division Oak Ridge United States

ISBN: (数字)9781665469388

ISBN: (纸本)9781665469388

This paper explores the dynamics of rice production in the Chinese provinces of Liaoning, Jilin, Heilongjiang, Shanghai, Jiangsu, and Zhejiang and seeks to predict monthly rice production in the months of April through October using precipitation and Normalized Difference Vegetation Index as the predictor variables available. We utilize ridge and lasso regression models to predict the rice yield. Results indicate that a lasso regression model with an R2 value of 0.9991501 with an adjusted R2 value of 0.9991502 and a ridge regression model with an R2 value of 0.9865443 with an adjusted R2 value of 0.9870637 are possible. The lasso regression model does not account for all predictor variables while the ridge regression model does. Both models could be expanded upon to include more observations. © 2022 IEEE.

关键词： Regression analysis

computer-Vision-Based Non-Contact Paste Concentration Measurement*

学校读者我要写书评

暂无评论

Computer-Vision-Based Non-Contact Paste Concentration Measur...

Automation in Manufacturing, Transportation and Logistics (ICaMaL), International Conference on

作者： Tailin Liang Zhaolin Yuan State Key Laboratory of Software Development Environment School of Computer Science and Engineering Beihang University Beijing China Department of Industrial and Systems Engineering Research Institute of Advanced Manufacturing The Hong Kong Polytechnic University Hong Kong China

ISBN: (数字)9798350378658

ISBN: (纸本)9798350378665

For monitoring the paste concentration, existing techniques, such as ultrasonic concentration meters and neutron meters, suffer from radiation hazards and low precision in high concentrations. This paper proposes a novel non-contact concentration measurement method based on deep learning. With the dataset collected by preparing the paste with different concentrations and taking photographs from the paste surface, a convolutional neural network is trained to extract the features from images and predict the concentration. The experiments indicate that the measurement accuracy is close to 88.79 % for the manual stirring paste dataset and 91.42 % for the automatic stirring paste dataset, which are sufficient in industrial applications. As a substitution of traditional concentration measurement instruments, the proposed vision-based concentration measurement method is non-contact and its accuracy is sufficient in industrial applications.

关键词： Meters Deep learning Accuracy Ultrasonic variables measurement Transportation Manuals Feature extraction Surface texture Convolutional neural networks Surface treatment

Outlier-Robust Training of Machine Learning Models

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Talak, Rajat Georgiou, Charis Shi, Jingnan Carlone, Luca Laboratory of Information & Decision Systems Massachusetts Institute of Technology CambridgeMA02139 United States Department of Electrical Engineering and Computer Science Massachusetts Institute of Technology CambridgeMA02139 United States

Robust training of machine learning models in the presence of outliers has garnered attention across various domains. The use of robust losses is a popular approach and is known to mitigate the impact of outliers. We bring to light two literatures that have diverged in their ways of designing robust losses: one using M-estimation, which is popular in robotics and computer vision, and another using a risk-minimization framework, which is popular in deep learning. We first show that a simple modification of the Black-Rangarajan duality provides a unifying view. The modified duality brings out a definition of a robust loss kernel σ that is satisfied by robust losses in both the literatures. Secondly, using the modified duality, we propose an Adaptive Alternation Algorithm (AAA) for training machine learning models with outliers. The algorithm iteratively trains the model by using a weighted version of the non-robust loss, while updating the weights at each iteration. The algorithm is augmented with a novel parameter update rule by interpreting the weights as inlier probabilities, and obviates the need for complex parameter tuning. Thirdly, we investigate convergence of the adaptive alternation algorithm to outlier-free optima. Considering arbitrary outliers (i.e., with no distributional assumption on the outliers), we show that the use of robust loss kernels σ increases the region of convergence. We experimentally show the efficacy of our algorithm on regression, classification, and neural scene reconstruction problems. © 2024, CC BY.

关键词： Contrastive Learning

Feature Correlation Aggregation: on the Path to Better Graph Neural Networks

学校读者我要写书评

暂无评论

Feature Correlation Aggregation: on the Path to Better Graph...

Proceedings of the Digital Image Computing: Technqiues and Applications (DICTA)

作者： Jieming Zhou Tong Zhang Pengfei Fang Lars Petersson Mehrtash Harandi College of Engineering and Computer Science Australian National University Canberra Australia Image and Visual Representation Laboratory EPFL Lausanne Switzerland Imaging and Computer Vision Group CSIRO Canberra Australia Department of Electrical and Computer Systems Engineering Monash University Melbourne Australia

Prior to the introduction of Graph Neural Networks (GNNs), modelling and analyzing irregular data, particularly graphs, was thought to be the Achilles’ heel of deep learning. The core concept of GNNs is to find a representation by recursively aggregating the representations of a central node and those of its neighbor, and many GNNs’ designs have demonstrated its success. However, most only focus on using the first-order information between a node and its neighbors. In this paper, we introduce a central node permutation variant function through a frustratingly simple and innocent-looking modification to the core operation of a GNN, namely the Feature cOrrelation aGgregation (FOG) module, which learns the second-order information from feature correlation between a node and its neighbors in the pipeline. By adding FOG into existing variants of GNNs, we empirically verify this second-order information complements the features generated by original GNNs across a broad set of benchmarks. A tangible boost in the model’s performance is observed where the model surpasses previous state-of-the-art results by a significant margin while employing fewer parameters. e.g., 26.202% improvement on a real-world molecular dataset using graph convolutional networks.

关键词：

A New Compact Cantor Set Fractal MIMO Antenna for 5G Millimeter-Wave Applications

学校读者我要写书评

暂无评论

A New Compact Cantor Set Fractal MIMO Antenna for 5G Millime...

Microwave and Antenna Symposium (IMAS), International

作者： Omaima Benkhadda Abdelati Reha Mohamed Saih Meryama Harrou Laboratory of Mathematics Computer Science Electrical Engineering and Physics (LAMIGEP) EMSI-Marrakech Marrakech Morocco Team of Microelectronics Embedded Systems and Telecommunication FST Sultan Moulay Slimane University Beni Mellal Morocco

ISBN: (数字)9798350356359

ISBN: (纸本)9798350356366

This article outlines the development of a compact MIMO antenna with two ports crafted for millimeter-wave band operations, specifically tailored to support fifth-generation (5G) wireless applications. The suggested antenna is constructed on a Rogers RT/duroid 5880 substrate, measuring ${12} \times {6}\times {0.508} \text{mm}{3}$ . The proposed antenna has a slotted ground on the bottom and two Fractal-shaped radiating elements inspired by the Cantor Set pattern, positioned atop the dielectric material. The suggested antenna covers the millimeter-wave spectrum utilized in 5G, showcasing a wide bandwidth extending from 26.6 GHz to 29.08 GHz and 38.5 to 39.2 GHz. The antenna attains a maximum gain of 7.64 dBi, robust isolation exceeding 26 dB and demonstrates exceptional diversity performance, with an envelope correlation coefficient (ECC) below 0.003 and a diversity gain (DG) surpassing 10 dB. The obtained results Confirm the validity of the designed MIMO antenna's robustness and make it a viable option for 5G wireless equipment.

关键词： Microwave antennas Wireless communication Correlation coefficient Antenna measurements Slot antennas 5G mobile communication Millimeter wave measurements Bandwidth Fractals Substrates