Kinds Of Prior Knowledge Research Articles

Initiation of knowledge and technology transfer from academia to industry: opportunity recognition and transfer channel choice

Abstract Transferring knowledge and technology from academia to industry is usually understood as a process. While previous research focuses on phenomena along the process and its outcomes, the starting point of the process—the initiation of a transfer activity—remains unstudied. We conceptualize this initiation as a simultaneous recognition of a transfer opportunity and the choice of a transfer channel and provide first empirical insights into the initiation of the transfer process. We use survey data from 1149 scientists from the German state of Thuringia and their activities in science–industry collaboration, intellectual property rights, and spin-off creation. We employ seemingly unrelated regressions to account for selection and multiple channel choices in our econometric approach. Our results show a positive relationship between scientists’ different kinds of prior knowledge and the probability of recognizing a transfer opportunity. Contrary to our expectations, scientific impact reduces the likelihood of recognizing a transfer opportunity. For the choice of the transfer channel, the results show a positive relationship between choosing the spin-off channel and risk willingness, as well as basic research. Applied research increases the likelihood of choosing intellectual property rights as a channel. Furthermore, role models are positively associated with these two channels.

A knowledge graphs representation method based on IsA relation modeling

Knowledge graph representation learning, which is able to support further knowledge computation and reasoning, has drawn much research attention in the field of artificial intelligence in recent years. Due to data sparsity, traditional knowledge graph methods focusing on the observed facts have poor performance. Therefore, researchers introduce auxiliary information to tackle this issue. As a kind of prior knowledge containing rich semantics, IsA relation could be used to boost model performance. We propose a knowledge representation learning method combined with isA relation modeling. Specifically, we model two features of isA relation—transitivity and antisymmetry. The transitivity is modeled by the projection invariance of vectors, and antisymmetry is modeled by partial order constraints. We use RotatE as our base model which could also be replaced by other knowledge embedding model with higher dimension. Finally, experimental results of link prediction and triple classification task on two datasets verify the proposed model’s effectiveness outperforming baseline models.

RFI-GAN: A reference-guided fuzzy integral network for ultrasound image augmentation

The Generative Adversarial Network (GAN) is commonly used for medical image augmentation, a method to alleviate the data shortage for downstream tasks. However, traditional GANs is not effective in mining the intrinsic association between samples and data space. Also, the multiple sections and variable scaling ratios of ultrasound images lead to inconsistency among images. The current GAN-based image generation methods cannot obtain the overall anatomical semantic constraints of inconsistent images, thus illogical images are generated. Therefore, a Reference-guided Fuzzy Integral GAN (RFI-GAN) is proposed. First, RFI-GAN implicitly constrains the logicality of the generated images through feature fusion between samples and reference set. These references, a kind of prior knowledge, cover most of the anatomical structures in the dataset. Second, as the general linear fusion takes features as additive and cannot fit the inconsistency of features, we generalize the additivity of features to non-additivity through the Fuzzy Integral, which is used to fusion non-additive measures. Two Fuzzy Integral Modules are proposed to nonlinearly fusion the texture and structure features of ultrasound images under the guidance of references. Finally, we perform experiments on the thyroid ultrasound image dataset, demonstrating the effectiveness of RFI-GAN in terms of generation and downstream classification tasks.

Millimeter-wave radar object classification using knowledge-assisted neural network.

To improve the cognition and understanding capabilities of artificial intelligence (AI) technology, it is a tendency to explore the human brain learning processing and integrate brain mechanisms or knowledge into neural networks for inspiration and assistance. This paper concentrates on the application of AI technology in advanced driving assistance system. In this field, millimeter-wave radar is essential for elaborate environment perception due to its robustness to adverse conditions. However, it is still challenging for radar object classification in the complex traffic environment. In this paper, a knowledge-assisted neural network (KANN) is proposed for radar object classification. Inspired by the human brain cognition mechanism and algorithms based on human expertise, two kinds of prior knowledge are injected into the neural network to guide its training and improve its classification accuracy. Specifically, image knowledge provides spatial information about samples. It is integrated into an attention mechanism in the early stage of the network to help reassign attention precisely. In the late stage, object knowledge is combined with the deep features extracted from the network. It contains discriminant semantic information about samples. An attention-based injection method is proposed to adaptively allocate weights to the knowledge and deep features, generating more comprehensive and discriminative features. Experimental results on measured data demonstrate that KANN is superior to current methods and the performance is improved with knowledge assistance.

Characterization of hv-Convex Sequences

Reconstructing a discrete object by means of X-rays along a finite set U of (discrete) directions represents one of the main task in discrete tomography. Indeed, it is an ill-posed inverse problem, since different structures exist having the same projections along all lines whose directions range in U. Characteristic of ambiguous reconstructions are special configurations, called switching components, whose understanding represents a main issue in discrete tomography, and an independent interesting geometric problem as well. The investigation of switching component usually bases on some kind of prior knowledge that is incorporated in the tomographic problem. In this paper, we focus on switching components under the constraint of convexity along the horizontal and the vertical directions imposed to the unknown object. Moving from their geometric characterization in windows and curls, we provide a numerical description, by encoding them as special sequences of integers. A detailed study of these sequences leads to the complete understanding of their combinatorial structure, and to a polynomial-time algorithm that explicitly reconstructs any of them from a pair of integers arbitrarily given.

A 3D Tubular Flux Model for Centerline Extraction in Neuron Volumetric Images.

Digital morphology reconstruction from neuron volumetric images is essential for computational neuroscience. The centerline of the axonal and dendritic tree provides an effective shape representation and serves as a basis for further neuron reconstruction. However, it is still a challenge to directly extract the accurate centerline from the complex neuron structure with poor image quality. In this paper, we propose a neuron centerline extraction method based on a 3D tubular flux model via a two-stage CNN framework. In the first stage, a 3D CNN is used to learn the latent neuron structure features, namely flux features, from neuron images. In the second stage, a light-weight U-Net takes the learned flux features as input to extract the centerline with a spatial weighted average strategy to constrain the multi-voxel width response. Specifically, the labels of flux features in the first stage are generated by the 3D tubular model which calculates the geometric representations of the flux between each voxel in the tubular region and the nearest point on the centerline ground truth. Compared with self-learned features by networks, flux features, as a kind of prior knowledge, explicitly take advantage of the contextual distance and direction distribution information around the centerline, which is beneficial for the precise centerline extraction. Experiments on two challenging datasets demonstrate that the proposed method outperforms other state-of-the-art methods by 18% and 35.1% in F1-measurement and average distance scores at the most, and the extracted centerline is helpful to improve the neuron reconstruction performance.

An Algorithm for the Retrieval of High Temporal-Spatial Resolution Shortwave Albedo from Landsat-8 Surface Reflectance and MODIS BRDF

Variations in surface physicochemical properties and spatial structures can prominently transform surface albedo which conversely influence surface energy balances and global climate, making it crucial to continuously monitor and quantify surface dynamics at fine scales. Here, we made two improvements to propose an algorithm for the simultaneous retrieval of 30-m Landsat albedo, based on the coupling of Landsat-8 and MODIS BRDF. First, two kinds of prior knowledge were added to disaggregate BRDF, including the Anisotropic Flat Index (AFX) and the Albedo-to-Nadir reflectance ratio (AN ratio), from MODIS scales into Landsat scales. Second, a simplified data fusion method was used to simulate albedo for the same, subsequent, or antecedent dates. Finally, we validated the reliability and correlations of the algorithm at six sites of the Surface Radiation (SURFRAD) budget network and intercompared the results with another algorithm called the ‘concurrent approach’. The results showed that the proposed algorithm had favorable usability and robustness, with a root mean square error (RMSE) of 0.015 (8%) and a mean bias of −0.005; while the concurrent approach had a RMSE of 0.026 (14%) and a mean bias of −0.018. The results emphasized that the proposed algorithm has captured subtle changes in albedo over a 16-day period.

An intelligent system of pelvic lymph node detection

Computed tomography (CT) scanning is a fast and painless procedure that can capture clear imaging information beneath the abdomen and is widely used to help diagnose and monitor disease progress. The pelvic lymph node is a key indicator of colorectal cancer metastasis. In the traditional process, an experienced radiologist must read all the CT scanning images slice by slice to track the lymph nodes for future diagnosis. However, this process is time-consuming, exhausting, and subjective due to the complex pelvic structure, numerous blood vessels, and small lymph nodes. Therefore, automated methods are desirable to make this process easier. Currently, the available open-source CTLNDataset only contains large lymph nodes. Consequently, a new data set called PLNDataset, which is dedicated to lymph nodes within the pelvis, is constructed to solve this issue. A two-level annotation calibration method is proposed to guarantee the quality and correctness of pelvic lymph node annotation. Moreover, a novel system composed of a keyframe localization network and a lymph node detection network is proposed to detect pelvic lymph nodes in CT scanning images. The proposed method makes full use of two kinds of prior knowledge: spatial prior knowledge for keyframe localization and anchor prior knowledge for lymph node detection. A series of experiments are carried out to evaluate the proposed method, including ablation experiments, comparing other state-of-the-art methods, and visualization of results. The experimental results demonstrate that our proposed method outperforms other methods on PLNDataset and CTLNDataset. This system is expected to be applied in future clinical practice.

Prior Knowledge Driven Label Embedding for Slot Filling in Natural Language Understanding

Traditional slot filling in natural language understanding (NLU) predicts a one-hot vector for each word. This form of label representation lacks semantic correlation modelling, which leads to severe data sparsity problem, especially when adapting an NLU model to a new domain. To address this issue, a novel label embedding based slot filling framework is proposed in this paper. Here, distributed label embedding is constructed for each slot using prior knowledge. Three encoding methods are investigated to incorporate different kinds of prior knowledge about slots: atomic concepts, slot descriptions, and slot exemplars. The proposed label embeddings tend to share text patterns and reuses data with different slot labels. This makes it useful for adaptive NLU with limited data. Also, since label embedding is independent of NLU model, it is compatible with almost all deep learning based slot filling models. The proposed approaches are evaluated on three datasets. Experiments on single domain and domain adaptation tasks show that label embedding achieves significant performance improvement over traditional one-hot label representation as well as advanced zero-shot approaches.

Knowledge-Powered Deep Breast Tumor Classification With Multiple Medical Reports.

Breast tumor classification with multiple medical reports such as B-ultrasound, Mammography (X-ray) and Nuclear Magnetic Resonance Imaging (MRI) is crucial to the intelligent cancer diagnosis system. Unlike the other domain texts, the medical reports have latent hierarchical syntactic structures and have hidden rich semantic information about the entities and relationships, which poses a great challenge of breast cancer classification. In this article, we proposed a Knowledge-powered Deep Breast Tumor Classification model (KDBTC), which takes the semantic information as a kind of prior knowledge and incorporated it into deep neural networks. Specially, our proposed model first uses Hierarchical Attention Bidirectional Recurrent Neural Networks (HA-BiRNNs) to encode the syntax-aware representation of medical reports in a hierarchical way. In the HA-BiRNN, a hierarchical neural network structure, consisting in two encoder layers of BiRNN (Bidirectional Recurrent Neural Networks), mirrors the hierarchical structure of medical reports, and a hierarchical attention mechanism, consisting of two levels attentions, attends to important elements within clinical report with word-level attention and sentence-level attention. Secondly, our model obtains the semantic information relevant to the medical reports from the clinical domain semantic tree, and encodes the semantic representation of medical reports by using Tree Structured Recurrent Neural Network with gated recursive units (Tree-GRUs). Finally, we classify breast tumors by combining both the syntax and semantic representations of medical reports. We evaluate our method on the real-world breast cancer medical reports, and results show that our method achieves higher performance on breast cancer classification.

Can soot primary particle size distributions be determined using laser-induced incandescence?

Soot from combustion processes often takes the form of fractal-like aggregates, assembled of primary particles, both of which obey polydisperse size distributions. In this work, the possibility of determining the primary particle size distribution through time-resolved laser-induced incandescence (TiRe-LII) under the influence of thermal shielding of polydispersely distributed aggregates is critically investigated for two typical measurement situations: in-flame measurements at high temperature and a soot-laden aerosol at room temperature. The uncertainty attached to the quantities is evaluated through Bayesian inference. We show how different kinds of prior knowledge concerning the aggregation state of the aerosol affect the uncertainties of the recovered size distribution parameters of the primary particles. To obtain reliable estimates for the primary particle size distribution parameters, specific information about the aggregate size distribution is required. This is especially the case for cold bath gases, where thermal shielding has a large effect. Furthermore, it is crucial to use the full duration of the usable LII signal trace to recover the width of the size distribution with small uncertainties. The uncertainty attached to TiRe-LII inferred primary particle size parameters becomes considerably larger when additional model parameters are considered.

Stable Recovery of Structured Signals From Corrupted Sub-Gaussian Measurements

This paper studies the problem of accurately recovering a structured signal from a small number of corrupted sub-Gaussian measurements. We consider three different procedures to reconstruct signal and corruption when different kinds of prior knowledge are available. In each case, we provide conditions (in terms of the number of measurements) for stable signal recovery from structured corruption with added unstructured noise. Our results theoretically demonstrate how to choose the regularization parameters in both partially and fully penalized recovery procedures and shed some light on the relationships among the three procedures. The key ingredient in our analysis is an extended matrix deviation inequality for isotropic sub-Gaussian matrices, which implies a tight lower bound for the restricted singular value of the extended sensing matrix. Numerical experiments are presented to verify our theoretical results.

A novel variational Bayesian method for variable selection in logistic regression models

With high-dimensional data emerging in various domains, sparse logistic regression models have gained much interest of researchers. Variable selection plays a key role in both improving the prediction accuracy and enhancing the interpretability of built models. Bayesian variable selection approaches enjoy many advantages such as high selection accuracy, easily incorporating many kinds of prior knowledge and so on. Because Bayesian methods generally make inference from the posterior distribution with Markov Chain Monte Carlo (MCMC) techniques, however, they become intractable in high-dimensional situations due to the large searching space. To address this issue, a novel variational Bayesian method for variable selection in high-dimensional logistic regression models is presented. The proposed method is based on the indicator model in which each covariate is equipped with a binary latent variable indicating whether it is important. The Bernoulli-type prior is adopted for the latent indicator variable. As for the specification of the hyperparameter in the Bernoulli prior, we provide two schemes to determine its optimal value so that the novel model can achieve sparsity adaptively. To identify important variables and make predictions, one efficient variational Bayesian approach is employed to make inference from the posterior distribution. The experiments conducted with both synthetic and some publicly available data show that the new method outperforms or is very competitive with some other popular counterparts.

Automatized spleen segmentation in non-contrast-enhanced MR volume data using subject-specific shape priors

To develop the first fully automated 3D spleen segmentation framework derived from T1-weighted magnetic resonance (MR) imaging data and to verify its performance for spleen delineation and volumetry. This approach considers the issue of low contrast between spleen and adjacent tissue in non-contrast-enhanced MR images. Native T1-weighted MR volume data was performed on a 1.5 T MR system in an epidemiological study. We analyzed random subsamples of MR examinations without pathologies to develop and verify the spleen segmentation framework. The framework is modularized to include different kinds of prior knowledge into the segmentation pipeline. Classification by support vector machines differentiates between five different shape types in computed foreground probability maps and recognizes characteristic spleen regions in axial slices of MR volume data. A spleen-shape space generated by training produces subject-specific prior shape knowledge that is then incorporated into a final 3D level set segmentation method. Individually adapted shape-driven forces as well as image-driven forces resulting from refined foreground probability maps steer the level set successfully to the segment the spleen. The framework achieves promising segmentation results with mean Dice coefficients of nearly 0.91 and low volumetric mean errors of 6.3%. The presented spleen segmentation approach can delineate spleen tissue in native MR volume data. Several kinds of prior shape knowledge including subject-specific 3D prior shape knowledge can be used to guide segmentation processes achieving promising results.

类别归纳推理的贝叶斯模型

贝叶斯模型为解释类别归纳推理的实验现象提供了一个统一的可计算框架。在该框架下，用不同的类别结构和随机过程表示不同的先验知识，并基于贝叶斯公式预测不同场景下的归纳力度。与其它模型相比，贝叶斯模型有较强的预测力度和更广的应用范围。文章总结了该模型的发展历史及现状，并首次系统阐述了其建模过程。未来研究可结合功能磁共振实验和计算语言学等方法，进一步拓展该模型的推理能力，提高其实际可用性。 The Bayesian model (BM) of category-based induction provides a unified computable framework for explaining the experimental phenomena (including the premise-conclusion similarity effect, the premise diversity effect, the premise monotonic effect and the premise-conclusion asymmetric effect, etc.) in category-based induction. Within this framework, the inductive reasoning in different contexts (such as induction about the generic biological properties or the causally transmitted properties) requires the constraint of different kinds of prior knowledge. Different kinds of prior knowledge can be represented by different kinds of category structures (i.e., the relationship between categories) and the corresponding stochastic process (i.e., the distribution of features/ properties in the category structure). Thus, BM can get the prior probability distributions for the Bayesian inference engine, and finally, the strength of an inductive argument can be calculated. As compared to the similarity coverage model (SCM) and feature-based inductive model (FBIM), BM can reflect the interaction of categories and properties, and has a clear mathematical basis, and also shows a better ability of prediction. This paper firstly reviews the research history and state of the art of the BM, and summarizes the process of computational cognitive modeling using BM. Secondly, BM is compared with the other models, and then the advantages and disadvantages of the BM are commented in details. Finally, some potential research directions are proposed: 1) further improving the ability of BM to deal with the common sense knowledge (e.g., the predatory behavior of animal), which may help to expand its application scope; 2) further increasing the power of BM to handle multiple objects and features/properties (if we learn that the animal A has the property X, what’s the possibility of the animal B having the property Y?); 3) that in combination with other methodologies (e.g., functional magnetic resonance imaging (fMRI) and computational linguistics, such as corpora), BM may improve its practical availability and reasoning abilities.

Semantic world modeling using probabilistic multiple hypothesis anchoring

In order to successfully perform typical household tasks such as manipulation or navigation, domestic robots need an accurate description of the world they are operating in. Creating and maintaining such a description, in this work referred to as world model, is a non-trivial task in a domestic environment that typically has a high number of objects, and is unstructured and dynamically changing. This work introduces probabilistic multiple hypothesis anchoring to create and maintain a semantically rich world model using probabilistic anchoring. Multiple hypothesis tracking-based data association is included to be able to deal with ambiguous scenarios. Multiple model tracking is included to be able to easily incorporate different kinds of prior knowledge.

Incorporating prior knowledge into learning by dividing training data

In most large-scale real-world pattern classification problems, there is always some explicit information besides given training data, namely prior knowledge, with which the training data are organized. In this paper, we proposed a framework for incorporating this kind of prior knowledge into the training of min-max modular (M3) classifier to improve learning performance. In order to evaluate the proposed method, we perform experiments on a large-scale Japanese patent classification problem and consider two kinds of prior knowledge included in patent documents: patent’s publishing date and the hierarchical structure of patent classification system. In the experiments, traditional support vector machine (SVM) and M3-SVM without prior knowledge are adopted as baseline classifiers. Experimental results demonstrate that the proposed method is superior to the baseline classifiers in terms of training cost and generalization accuracy. Moreover, M3-SVM with prior knowledge is found to be much more robust than traditional support vector machine to noisy dated patent samples, which is crucial for incremental learning.

Towards effective document clustering: A constrained K-means based approach

Document clustering is an important tool for document collection organization and browsing. In real applications, some limited knowledge about cluster membership of a small number of documents is often available, such as some pairs of documents belonging to the same cluster. This kind of prior knowledge can be served as constraints for the clustering process. We integrate the constraints into the trace formulation of the sum of square Euclidean distance function of K- means. Then,the combined criterion function is transformed into trace maximization, which is further optimized by eigen-decomposition. Our experimental evaluation shows that the proposed semi-supervised clustering method can achieve better performance, compared to three existing methods.

A Markov model for blind image separation by a mean-field EM algorithm

This paper deals with blind separation of images from noisy linear mixtures with unknown coefficients, formulated as a Bayesian estimation problem. This is a flexible framework, where any kind of prior knowledge about the source images and the mixing matrix can be accounted for. In particular, we describe local correlation within the individual images through the use of Markov random field (MRF) image models. These are naturally suited to express the joint pdf of the sources in a factorized form, so that the statistical independence requirements of most independent component analysis approaches to blind source separation are retained. Our model also includes edge variables to preserve intensity discontinuities. MRF models have been proved to be very efficient in many visual reconstruction problems, such as blind image restoration, and allow separation and edge detection to be performed simultaneously. We propose an expectation-maximization algorithm with the mean field approximation to derive a procedure for estimating the mixing matrix, the sources, and their edge maps. We tested this procedure on both synthetic and real images, in the fully blind case (i.e., no prior information on mixing is exploited) and found that a source model accounting for local autocorrelation is able to increase robustness against noise, even space variant. Furthermore, when the model closely fits the source characteristics, independence is no longer a strict requirement, and cross-correlated sources can be separated, as well.

Multi-tuple interpolation using Fourier descriptors

This paper presents a new object interpolation method called multi-tuple interpolation. Its principal feature is the use of Fourier descriptors to represent a contour and to obtain its four attributes, namely, shape, scale, orientation and position. Each attribute is interpolated in the frequency domain under the control of its individual parameter. A suitable combination of these parameters, which can be regarded as a kind of prior knowledge, can generate a result adaptable to the real 3D object. Several experiments are performed on both synthetic and real medical images to demonstrate this new approach. It is shown that the proposed method can be applied to other similar domains, e.g. shape blending.