scholarly journals Multi Similarity Metric Fusion in Graph-Based Semi-Supervised Learning

Computation ◽  
2019 ◽  
Vol 7 (1) ◽  
pp. 15 ◽  
Author(s):  
Saeedeh Bahrami ◽  
Alireza Bosaghzadeh ◽  
Fadi Dornaika
Keyword(s):  
Label Propagation ◽  
Unified Framework ◽  
Similarity Metric ◽  
Metric Graph ◽  
Manifold Embedding ◽  
Correlation Graph ◽  
Graph Estimation ◽  
Face Datasets ◽  
Two Phases ◽  
Fusion Approach

In semi-supervised label propagation (LP), the data manifold is approximated by a graph, which is considered as a similarity metric. Graph estimation is a crucial task, as it affects the further processes applied on the graph (e.g., LP, classification). As our knowledge of data is limited, a single approximation cannot easily find the appropriate graph, so in line with this, multiple graphs are constructed. Recently, multi-metric fusion techniques have been used to construct more accurate graphs which better represent the data manifold and, hence, improve the performance of LP. However, most of these algorithms disregard use of the information of label space in the LP process. In this article, we propose a new multi-metric graph-fusion method, based on the Flexible Manifold Embedding algorithm. Our proposed method represents a unified framework that merges two phases: graph fusion and LP. Based on one available view, different simple graphs were efficiently generated and used as input to our proposed fusion approach. Moreover, our method incorporated the label space information as a new form of graph, namely the Correlation Graph, with other similarity graphs. Furthermore, it updated the correlation graph to find a better representation of the data manifold. Our experimental results on four face datasets in face recognition demonstrated the superiority of the proposed method compared to other state-of-the-art algorithms.

Discriminative sparse flexible manifold embedding with novel graph for robust visual representation and label propagation

2017 ◽  
Vol 61 ◽  
pp. 492-510 ◽  
Author(s):  
Zhao Zhang ◽  
Yan Zhang ◽  
Fanzhang Li ◽  
Mingbo Zhao ◽  
Li Zhang ◽  
...  
Keyword(s):  
Visual Representation ◽  
Label Propagation ◽  
Manifold Embedding

scholarly journals Spatial-Temporal Person Re-Identification

2019 ◽  
Vol 33 ◽  
pp. 8933-8940 ◽  
Author(s):  
Guangcong Wang ◽  
Jianhuang Lai ◽  
Peigen Huang ◽  
Xiaohua Xie
Keyword(s):  
Probability Distribution ◽  
Semantic Information ◽  
Temporal Information ◽  
Unified Framework ◽  
Temporal Constraint ◽  
Query Image ◽  
Similarity Metric ◽  
Heterogeneous Information ◽  
Large Margin ◽  
Previous State

Most of current person re-identification (ReID) methods neglect a spatial-temporal constraint. Given a query image, conventional methods compute the feature distances between the query image and all the gallery images and return a similarity ranked table. When the gallery database is very large in practice, these approaches fail to obtain a good performance due to appearance ambiguity across different camera views. In this paper, we propose a novel two-stream spatial-temporal person ReID (st-ReID) framework that mines both visual semantic information and spatial-temporal information. To this end, a joint similarity metric with Logistic Smoothing (LS) is introduced to integrate two kinds of heterogeneous information into a unified framework. To approximate a complex spatial-temporal probability distribution, we develop a fast Histogram-Parzen (HP) method. With the help of the spatial-temporal constraint, the st-ReID model eliminates lots of irrelevant images and thus narrows the gallery database. Without bells and whistles, our st-ReID method achieves rank-1 accuracy of 98.1% on Market-1501 and 94.4% on DukeMTMC-reID, improving from the baselines 91.2% and 83.8%, respectively, outperforming all previous state-of-theart methods by a large margin.

scholarly journals Contrastive Learning and Neural Oscillations

1991 ◽  
Vol 3 (4) ◽  
pp. 526-545 ◽  
Author(s):  
Pierre Baldi ◽  
Fernando Pineda
Keyword(s):  
Gradient Descent ◽  
Learning Algorithms ◽  
Learning Rule ◽  
Unified Framework ◽  
Contrast Function ◽  
New Learning ◽  
Different Types ◽  
Free Network ◽  
Two Phases ◽  
The Brain

The concept of Contrastive Learning (CL) is developed as a family of possible learning algorithms for neural networks. CL is an extension of Deterministic Boltzmann Machines to more general dynamical systems. During learning, the network oscillates between two phases. One phase has a teacher signal and one phase has no teacher signal. The weights are updated using a learning rule that corresponds to gradient descent on a contrast function that measures the discrepancy between the free network and the network with a teacher signal. The CL approach provides a general unified framework for developing new learning algorithms. It also shows that many different types of clamping and teacher signals are possible. Several examples are given and an analysis of the landscape of the contrast function is proposed with some relevant predictions for the CL curves. An approach that may be suitable for collective analog implementations is described. Simulation results and possible extensions are briefly discussed together with a new conjecture regarding the function of certain oscillations in the brain. In the appendix, we also examine two extensions of contrastive learning to time-dependent trajectories.

scholarly journals Improving Multi-Label Learning by Correlation Embedding

2021 ◽  
Vol 11 (24) ◽  
pp. 12145
Author(s):  
Jun Huang ◽  
Qian Xu ◽  
Xiwen Qu ◽  
Yaojin Lin ◽  
Xiao Zheng
Keyword(s):  
Feature Space ◽  
Classification Model ◽  
Unified Framework ◽  
Single Instance ◽  
Nonlinear Version ◽  
Correlation Graph ◽  
Label Correlations ◽  
Class Labels ◽  
Low Dimensional ◽  
Label Correlation

In multi-label learning, each object is represented by a single instance and is associated with more than one class labels, where the labels might be correlated with each other. As we all know, exploiting label correlations can definitely improve the performance of a multi-label classification model. Existing methods mainly model label correlations in an indirect way, i.e., adding extra constraints on the coefficients or outputs of a model based on a pre-learned label correlation graph. Meanwhile, the high dimension of the feature space also poses great challenges to multi-label learning, such as high time and memory costs. To solve the above mentioned issues, in this paper, we propose a new approach for Multi-Label Learning by Correlation Embedding, namely MLLCE, where the feature space dimension reduction and the multi-label classification are integrated into a unified framework. Specifically, we project the original high-dimensional feature space to a low-dimensional latent space by a mapping matrix. To model label correlation, we learn an embedding matrix from the pre-defined label correlation graph by graph embedding. Then, we construct a multi-label classifier from the low-dimensional latent feature space to the label space, where the embedding matrix is utilized as the model coefficients. Finally, we extend the proposed method MLLCE to the nonlinear version, i.e., NL-MLLCE. The comparison experiment with the state-of-the-art approaches shows that the proposed method MLLCE has a competitive performance in multi-label learning.

scholarly journals Leveraging Latent Label Distributions for Partial Label Learning

2018 ◽  
Author(s):  
Lei Feng ◽  
Bo An
Keyword(s):  
Latent Variable ◽  
Ground Truth ◽  
Label Propagation ◽  
Unified Framework ◽  
Local Consistency ◽  
Confidence Levels ◽  
Model Training ◽  
Real World Datasets ◽  
Partial Label Learning ◽  
Learning Frameworks

In partial label learning, each training example is assigned a set of candidate labels, only one of which is the ground-truth label. Existing partial label learning frameworks either assume each candidate label of equal confidence or consider the ground-truth label as a latent variable hidden in the indiscriminate candidate label set, while the different labeling confidence levels of the candidate labels are regrettably ignored. In this paper, we formalize the different labeling confidence levels as the latent label distributions, and propose a novel unified framework to estimate the latent label distributions while training the model simultaneously. Specifically, we present a biconvex formulation with constrained local consistency and adopt an alternating method to solve this optimization problem. The process of alternating optimization exactly facilitates the mutual adaption of the model training and the constrained label propagation. Extensive experimental results on controlled UCI datasets as well as real-world datasets clearly show the effectiveness of the proposed approach.

Combining Graph Convolutional Neural Networks and Label Propagation

2022 ◽  
Vol 40 (4) ◽  
pp. 1-27
Author(s):  
Hongwei Wang ◽  
Jure Leskovec
Keyword(s):  
Neural Networks ◽  
Convolutional Neural Networks ◽  
Message Passing ◽  
Label Propagation ◽  
Unified Framework ◽  
Label Information ◽  
Message Passing Algorithms ◽  
Node Classification ◽  
Improved Performance ◽  
Node Labels

Label Propagation Algorithm (LPA) and Graph Convolutional Neural Networks (GCN) are both message passing algorithms on graphs. Both solve the task of node classification, but LPA propagates node label information across the edges of the graph, while GCN propagates and transforms node feature information. However, while conceptually similar, theoretical relationship between LPA and GCN has not yet been systematically investigated. Moreover, it is unclear how LPA and GCN can be combined under a unified framework to improve the performance. Here we study the relationship between LPA and GCN in terms of feature/label influence , in which we characterize how much the initial feature/label of one node influences the final feature/label of another node in GCN/LPA. Based on our theoretical analysis, we propose an end-to-end model that combines GCN and LPA. In our unified model, edge weights are learnable, and the LPA serves as regularization to assist the GCN in learning proper edge weights that lead to improved performance. Our model can also be seen as learning the weights of edges based on node labels, which is more direct and efficient than existing feature-based attention models or topology-based diffusion models. In a number of experiments for semi-supervised node classification and knowledge-graph-aware recommendation, our model shows superiority over state-of-the-art baselines.

scholarly journals Semi-Supervised Ridge Regression with Adaptive Graph-Based Label Propagation

2018 ◽  
Vol 8 (12) ◽  
pp. 2636 ◽  
Author(s):  
Yugen Yi ◽  
Yuqi Chen ◽  
Jiangyan Dai ◽  
Xiaolin Gui ◽  
Chunlei Chen ◽  
...  
Keyword(s):  
Convergence Analysis ◽  
Ridge Regression ◽  
Supervised Classification ◽  
Label Propagation ◽  
Classification Method ◽  
Sample Problem ◽  
Unified Framework ◽  
Graph Learning ◽  
Learning Scheme ◽  
Out Of Sample

In order to overcome the drawbacks of the ridge regression and label propagation algorithms, we propose a new semi-supervised classification method named semi-supervised ridge regression with adaptive graph-based label propagation (SSRR-AGLP). Firstly, we present a new adaptive graph-learning scheme and integrate it into the procedure of label propagation, in which the locality and sparsity of samples are considered simultaneously. Then, we introduce the ridge regression algorithm into label propagation to solve the “out of sample” problem. As a consequence, the proposed SSSRR-AGLP integrates adaptive graph learning, label propagation and ridge regression into a unified framework. Finally, an effective iterative updating algorithm is designed for solving the algorithm, and the convergence analysis is also provided. Extensive experiments are conducted on five databases. Through comparing the results with some well-known algorithms, the effectiveness and superiority of the proposed algorithm are demonstrated.

Precipitation of NiHfsi phase in NiAl single crystals containing Hf

1995 ◽  
Vol 53 ◽  
pp. 532-533
Author(s):  
A. Garg ◽  
R. D. Noebe ◽  
R. Darolia
Keyword(s):  
High Temperature ◽  
Single Crystals ◽  
High Temperature Strength ◽  
Bridgman Technique ◽  
Shell Mold ◽  
Third Phase ◽  
Unknown Phase ◽  
Transmission Electron ◽  
Two Phases ◽  
Nial Matrix

Small additions of Hf to NiAl produce a significant increase in the high-temperature strength of single crystals. Hf has a very limited solubility in NiAl and in the presence of Si, results in a high density of G-phase (Ni16Hf6Si7) cuboidal precipitates and some G-platelets in a NiAl matrix. These precipitates have a F.C.C structure and nucleate on {100}NiAl planes with almost perfect coherency and a cube-on-cube orientation-relationship (O.R.). However, G-phase is metastable and after prolonged aging at high temperature dissolves at the expense of a more stable Heusler (β'-Ni2AlHf) phase. In addition to these two phases, a third phase was shown to be present in a NiAl-0.3at. % Hf alloy, but was not previously identified (Fig. 4 of ref. 2 ). In this work, we report the morphology, crystal-structure, O.R., and stability of this unknown phase, which were determined using conventional and analytical transmission electron microscopy (TEM).Single crystals of NiAl containing 0.5at. % Hf were grown by a Bridgman technique. Chemical analysis indicated that these crystals also contained Si, which was not an intentional alloying addition but was picked up from the shell mold during directional solidification.

Imaging of Al-Li-Cu alloys with a Scanning Ion Microprobe

1990 ◽  
Vol 48 (2) ◽  
pp. 314-315
Author(s):  
K.K. Soni ◽  
D.B. Williams ◽  
J.M. Chabala ◽  
R. Levi-Setti ◽  
D.E. Newbury
Keyword(s):  
Mass Spectrometry ◽  
Secondary Ion Mass Spectrometry ◽  
Equilibrium Phase ◽  
Icosahedral Symmetry ◽  
Ion Microprobe ◽  
X Ray ◽  
Cu Alloys ◽  
Two Phases ◽  
The University ◽  
Secondary Ion

In contrast to the inability of x-ray microanalysis to detect Li, secondary ion mass spectrometry (SIMS) generates a very strong Li+ signal. The latter’s potential was recently exploited by Williams et al. in the study of binary Al-Li alloys. The present study of Al-Li-Cu was done using the high resolution scanning ion microprobe (SIM) at the University of Chicago (UC). The UC SIM employs a 40 keV, ∼70 nm diameter Ga+ probe extracted from a liquid Ga source, which is scanned over areas smaller than 160×160 μm2 using a 512×512 raster. During this experiment, the sample was held at 2 × 10-8 torr.In the Al-Li-Cu system, two phases of major importance are T1 and T2, with nominal compositions of Al2LiCu and Al6Li3Cu respectively. In commercial alloys, T1 develops a plate-like structure with a thickness <∼2 nm and is therefore inaccessible to conventional microanalytical techniques. T2 is the equilibrium phase with apparent icosahedral symmetry and its presence is undesirable in industrial alloys.

Defect structures of Nb1+αS2 sulfidation scales

1992 ◽  
Vol 50 (1) ◽  
pp. 38-39
Author(s):  
Chuxin Zhou ◽  
L. W. Hobbs
Keyword(s):  
Stacking Faults ◽  
Rhombohedral Structure ◽  
Stacking Sequence ◽  
Defect Structures ◽  
X Ray Diffraction ◽  
X Ray ◽  
Plane Normal ◽  
Lower Case ◽  
Sulfur Layer ◽  
Two Phases

One of the major purposes in the present work is to study the high temperature sulfidation properties of Nb in severe sulfidizing environments. Kinetically, the sulfidation rate of Nb is satisfactorily slow, but the microstructures and non-stoichiometry of Nb1+αS2 challenge conventional oxidation/sulfidation theory and defect models of non-stoichiometric compounds. This challenge reflects our limited knowledge of the dependence of kinetics and atomic migration processes in solid state materials on their defect structures.Figure 1 shows a high resolution image of a platelet from the middle portion of the Nb1+αS2 scale. A thin lamellar heterogeneity (about 5nm) is observed. From X-ray diffraction results, we have shown that Nb1+αS2 scale is principally rhombohedral structure, but 2H-NbS2 can result locally due to stacking faults, because the only difference between these 2H and 3R phases is variation in the stacking sequence along the c axis. Following an ABC notation, we use capital letters A, B and C to represent the sulfur layer, and lower case letters a, b and c to refer to Nb layers. For example, the stacking sequence of 2H phase is AbACbCA, which is a ∼12Å period along the c axis; the stacking sequence of 3R phase is AbABcBCaCA to form an ∼18Å period along the c axis. Intergrowth of these two phases can take place at stacking faults or by a shear in the basal plane normal to the c axis.

Sign in / Sign up

Export Citation Format

Share Document