Learning Sample Selection Research Articles

Active learning sample selection based on multicriteria

In multivariate calibration problems, model performance is affected significantly by the calibration samples used during model building. In recent years, active learning methods have become one of the best methods for sample selection. However, most active learning methods only select instances from prediction uncertainty or sample space distance, and these single-criteria methods tend to select undesired samples. In addition, sample density characterizes the spatial information carried by the sample, but few studies in quantitative analysis utilize sample density alone to select calibration samples. Considering these issues, based on the k-means clustering algorithm, this paper proposes an active learning sample selection method (Diversity Informativeness Density Active Learning, DIDAL), which combines the three criteria of diversity, informativeness and sample density. The most representative sample is iteratively selected for - addition to the calibration set for modeling and estimating the chemical concentration of analytes. Soybean meal and soy sauce samples were analyzed by DIDAL and compared with existing sample selection methods. The prediction results show that the DIDAL algorithm significantly outperforms several existing algorithms and is close to the performance of full-sample modeling. A model with high prediction accuracy can be constructed by selecting only a few samples using the DIDAL method.

Semi-Supervised Deep Kernel Active Learning for Material Removal Rate Prediction in Chemical Mechanical Planarization.

The material removal rate (MRR) is an important variable but difficult to measure in the chemical-mechanical planarization (CMP) process. Most data-based virtual metrology (VM) methods ignore the large number of unlabeled samples, resulting in a waste of information. In this paper, the semi-supervised deep kernel active learning (SSDKAL) model is proposed. Clustering-based phase partition and phase-matching algorithms are used for the initial feature extraction, and a deep network is used to replace the kernel of Gaussian process regression so as to extract hidden deep features. Semi-supervised regression and active learning sample selection strategies are applied to make full use of information on the unlabeled samples. The experimental results of the CMP process dataset validate the effectiveness of the proposed method. Compared with supervised regression and co-training-based semi-supervised regression algorithms, the proposed model has a lower mean square error with different labeled sample proportions. Compared with other frameworks proposed in the literature, such as physics-based VM models, Gaussian-process-based regression models, and stacking models, the proposed method achieves better prediction results without using all the labeled samples.

Bio-inspired Active Learning method in spiking neural network

Spiking neural networks (SNNs) have gained a lot of attention and achievements recently because of their low-power advantages on neuromorphic hardware. However, training deep SNNs still requires a large number of labeled data which are expensive to obtain. To address this issue, we propose an effective Bio-inspired Active Learning (BAL) method in this paper to reduce the training cost of SNN models. Specifically, bio-inspired behavior patterns of spiking neurons are defined to represent the internal states of SNN models for active learning. Then, an active learning sample selection strategy is proposed by leveraging the empirical and generalization pattern divergence in SNNs. By labeling selected samples and adding them to training, behavioral patterns can be optimized to improve the performance of neural networks. Comprehensive experiments are conducted on the CIFAR-10, SVHN, and Fashion-MNIST datasets with various sample proportions. The experimental results demonstrate that the proposed BAL achieves state-of-the-arts performance in SNNs compared with the existing active learning methods.

Performance enhancement‐based active learning sample selection method

AbstractRepresentative samples are important for multivariate calibration. The highly efficient selection of representative samples to be labelled can save money and time. Existing methods, such as Kennard‐Stone and net analyte signal selection, are usually based on the distance between candidate samples and labelled calibration sets in feature space. However, these distances are influenced by the feature space, which is spanned by an information vector extracted from labelled samples. To overcome the negative effects of the distance‐based selection method, a model performance enhancement‐based sample selection method is proposed to select calibration samples efficiently. Based on loss function optimization, the samples that can improve model performance the most, as estimated by bootstrap, are sequentially selected and added to the calibration set. Due to the high representation of each sample, a few samples can build a model that has no significant loss of prediction ability when compared with a model built with the large number set of calibration samples. The performance enhancement‐based active learning (PEAL) sample selection method is both effective and efficient.

Endogenous sample selection: A laboratory study

Accounting for sample selection is a challenge not only for empirical researchers, but also for the agents populating our models. Yet most models abstract from these issues and assume that agents successfully tackle selection problems. We design an experiment where a person who understands selection observes all the data required to account for it. Subjects make choices under uncertainty and their choices reveal valuable information that is biased due to the presence of unobservables. We find that almost no subjects optimally account for endogenous selection. On the other hand, behavior is far from random, but actually quite amenable to analysis: Subjects follow simple heuristics that result in a partial accounting of selection and mitigate mistakes. Contingent thinking learning sample selection C91 D83

Statistical Rare-Event Analysis and Parameter Guidance by Elite Learning Sample Selection

Accurately estimating the failure region of rare events for memory-cell and analog circuit blocks under process variations is a challenging task. In this article, we propose a new statistical method, called EliteScope , to estimate the circuit failure rates in rare-event regions and to provide conditions of parameters to achieve targeted performance. The new method is based on the iterative blockade framework to reduce the number of samples, but consists of two new techniques to improve existing methods. First, the new approach employs an elite-learning sample-selection scheme, which can consider the effectiveness of samples and well coverage for the parameter space. As a result, it can reduce additional simulation costs by pruning less effective samples while keeping the accuracy of failure estimation. Second, the EliteScope identifies the failure regions in terms of parameter spaces to provide a good design guidance to accomplish the performance target. It applies variance-based feature selection to find the dominant parameters and then determine the in-spec boundaries of those parameters. We demonstrate the advantage of our proposed method using several memory and analog circuits with different numbers of process parameters. Experiments on four circuit examples show that EliteScope achieves a significant improvement on failure-region estimation in terms of accuracy and simulation cost over traditional approaches. The 16b 6T-SRAM column example also demonstrates that the new method is scalable for handling large problems with large numbers of process variables.

Efficient optimization of fully-integrated inductive DC–DC converters comprising tapered inductor layout synthesis and temperature effects

The evolution of computer-aided design tools has extended the capabilities of a designer by pushing the optimality of complex circuits beyond the ad hoc manual implementation. This work presents a framework to co-optimize the circuit and the layout parameters of fully integrated inductive DC–DC converters. The framework comprises expensive optimization that is speeded up by active learning sample selection and evolutionary techniques to acquire an optimal converter. A tapered inductor topology is used to increase the quality of the on-chip inductor and to improve the efficiency of the overall monolithic DC–DC converter. The optimization framework is validated by co-optimizing the design parameters and the tapered inductor layout for a fully-integrated DC–DC boost converter in a 0.13 μm CMOS technology. The power loss in the circuit is reduced with 27 % resulting in a 7 % efficiency improvement, compared to a fully-integrated DC–DC boost converter with a regular inductor topology.

Hyperspectral image classification using an unsupervised neuro-fuzzy system

An unsupervised neuro-fuzzy system, Gaussian fuzzy self-organizing map (GFSOM), is proposed for hyperspectral image classification. This algorithm operates by integrating an unsupervised neural network with a Gaussian function-based fuzzy system. We also explore the potential for hyperspectral image analysis of three other artificial intelligence (AI)-based unsupervised techniques popular for multispectral image analysis: self-organizing map (SOM), fuzzy c-mean (FCM), and descending fuzzy learning vector quantization (DFLVQ). To apply these methods effectively and efficiently to hyperspectral imagery, an optimal learning sample selection strategy and a prototype initialization system are developed. An experimental study on classifying an EO-1/Hyperion hyperspectral image illustrates that GFSOM achieves the best accuracy, since it can model both the central tendency characteristics of input samples and capture the dispersion characteristics of data within a cluster. By adopting the system initialization approach developed here, all the AI-based techniques have the capability to classify hyperspectral images and can deliver acceptable accuracy, which could consequently accelerate their transitions from the multispectral to the hyperspectral field.

Seismic pattern recognition

This paper examines the fundamental problems in automatic recognition of seismic events. Particular emphasis is placed on feature extraction and digital signal processing with the discussion of geophysical features, comparison of feature extraction criteria, comparison of orthogonal transforms, autocovariance features, short-time spectral features, and signal modelling. Extensive computer results are provided. Other problem areas considered are cluster analysis, learning sample selection and the statistical and structural mixed model.