scholarly journals Guided Generative Models using Weak Supervision for Detecting Object Spatial Arrangement in Overhead Images

2021 ◽  
Author(s):  
Weiwei Duan ◽  
Yao-Yi Chiang ◽  
Stefan Leyk ◽  
Johannes H. Uhl ◽  
Craig A. Knoblock
Keyword(s):  
Spatial Arrangement ◽  
Generative Models ◽  
Weak Supervision

scholarly journals Weakly Supervised Disentanglement by Pairwise Similarities

2020 ◽  
Vol 34 (04) ◽  
pp. 3495-3502 ◽  
Author(s):  
Junxiang Chen ◽  
Kayhan Batmanghelich
Keyword(s):  
Real World ◽  
Latent Variables ◽  
Generative Models ◽  
Experimental Results ◽  
New Method ◽  
Weak Supervision ◽  
Real World Problem ◽  
Variational Autoencoder ◽  
Weakly Supervised

Recently, researches related to unsupervised disentanglement learning with deep generative models have gained substantial popularity. However, without introducing supervision, there is no guarantee that the factors of interest can be successfully recovered (Locatello et al. 2018). Motivated by a real-world problem, we propose a setting where the user introduces weak supervision by providing similarities between instances based on a factor to be disentangled. The similarity is provided as either a binary (yes/no) or real-valued label describing whether a pair of instances are similar or not. We propose a new method for weakly supervised disentanglement of latent variables within the framework of Variational Autoencoder. Experimental results demonstrate that utilizing weak supervision improves the performance of the disentanglement method substantially.

scholarly journals Weakly Supervised Sequence Tagging from Noisy Rules

2020 ◽  
Vol 34 (04) ◽  
pp. 5570-5578
Author(s):  
Esteban Safranchik ◽  
Shiying Luo ◽  
Stephen Bach
Keyword(s):  
Markov Models ◽  
Named Entity Recognition ◽  
Training Sequence ◽  
Generative Models ◽  
Training Data ◽  
Entity Recognition ◽  
Weak Supervision ◽  
Sequence Elements ◽  
New Type ◽  
Weakly Supervised

We propose a framework for training sequence tagging models with weak supervision consisting of multiple heuristic rules of unknown accuracy. In addition to supporting rules that vote on tags in the output sequence, we introduce a new type of weak supervision, called linking rules, that vote on how sequence elements should be grouped into spans with the same tag. These rules are an alternative to candidate span generators that require significantly more human effort. To estimate the accuracies of the rules and combine their conflicting outputs into training data, we introduce a new type of generative model, linked hidden Markov models (linked HMMs), and prove they are generically identifiable (up to a tag permutation) without any observed training labels. We find that linked HMMs provide an average 7 F1 point boost on benchmark named entity recognition tasks versus generative models that assume the tags are i.i.d. Further, neural sequence taggers trained with these structure-aware generative models outperform comparable state-of-the-art approaches to weak supervision by an average of 2.6 F1 points.

Conformation of Ribosomes from Escherichia Coli

1974 ◽  
Vol 32 ◽  
pp. 210-211
Author(s):  
M. Boublik ◽  
W. Hellmann ◽  
F. Jenkins
Keyword(s):  
Binding Sites ◽  
Ribosomal Proteins ◽  
Fluorescent Dyes ◽  
Protein Biosynthesis ◽  
Three Dimensional ◽  
Spatial Arrangement ◽  
Dimensional Structure ◽  
Complete Understanding ◽  
And Function

The present knowledge of the three-dimensional structure of ribosomes is far too limited to enable a complete understanding of the various roles which ribosomes play in protein biosynthesis. The spatial arrangement of proteins and ribonuclec acids in ribosomes can be analysed in many ways. Determination of binding sites for individual proteins on ribonuclec acid and locations of the mutual positions of proteins on the ribosome using labeling with fluorescent dyes, cross-linking reagents, neutron-diffraction or antibodies against ribosomal proteins seem to be most successful approaches. Structure and function of ribosomes can be correlated be depleting the complete ribosomes of some proteins to the functionally inactive core and by subsequent partial reconstitution in order to regain active ribosomal particles.

Structural organization of escherichia coli ribosomes as determined by immuno electron microscopy

1978 ◽  
Vol 36 (2) ◽  
pp. 166-167 ◽  
Author(s):  
G. Stöffler ◽  
R.W. Bald ◽  
J. Dieckhoff ◽  
H. Eckhard ◽  
R. Lührmann ◽  
...  
Keyword(s):  
Electron Microscopy ◽  
Escherichia Coli ◽  
Ribosomal Proteins ◽  
Structural Organization ◽  
Three Dimensional ◽  
Ribosomal Subunit ◽  
Spatial Arrangement ◽  
Small Subunit ◽  
Antibody Binding ◽  
Immuno Electron Microscopy

A central step towards an understanding of the structure and function of the Escherichia coli ribosome, a large multicomponent assembly, is the elucidation of the spatial arrangement of its 54 proteins and its three rRNA molecules. The structural organization of ribosomal components has been investigated by a number of experimental approaches. Specific antibodies directed against each of the 54 ribosomal proteins of Escherichia coli have been performed to examine antibody-subunit complexes by electron microscopy. The position of the bound antibody, specific for a particular protein, can be determined; it indicates the location of the corresponding protein on the ribosomal surface.The three-dimensional distribution of each of the 21 small subunit proteins on the ribosomal surface has been determined by immuno electron microscopy: the 21 proteins have been found exposed with altogether 43 antibody binding sites. Each one of 12 proteins showed antibody binding at remote positions on the subunit surface, indicating highly extended conformations of the proteins concerned within the 30S ribosomal subunit; the remaining proteins are, however, not necessarily globular in shape (Fig. 1).

Colloidal systems in soils

1994 ◽  
Vol 52 ◽  
pp. 64-65
Author(s):  
J. Thieme ◽  
J. Niemeyer ◽  
P. Guttman
Keyword(s):  
Clay Minerals ◽  
Polymeric Materials ◽  
Spatial Arrangement ◽  
High Specific Surface Area ◽  
Turnover Rates ◽  
Colloidal Systems ◽  
Chemical Conditions ◽  
Aggregation Phenomena ◽  
Iron And Manganese ◽  
Soil Colloids

In soil science the fraction of colloids in soils is understood as particles with diameters smaller than 2μm. Clay minerals, aquoxides of iron and manganese, humic substances, and other polymeric materials are found in this fraction. The spatial arrangement (microstructure) is controlled by the substantial structure of the colloids, by the chemical composition of the soil solution, and by thesoil biota. This microstructure determines among other things the diffusive mass flow within the soils and as a result the availability of substances for chemical and microbiological reactions. The turnover of nutrients, the adsorption of toxicants and the weathering of soil clay minerals are examples of these surface mediated reactions. Due to their high specific surface area, the soil colloids are the most reactive species in this respect. Under the chemical conditions in soils, these minerals are associated in larger aggregates. The accessibility of reactive sites for these reactions on the surface of the colloids is reduced by this aggregation. To determine the turnover rates of chemicals within these aggregates it is highly desirable to visualize directly these aggregation phenomena.

Computational Micrograph Registration with Sieve Processes

1996 ◽  
Vol 54 ◽  
pp. 440-441
Author(s):  
P.J. Phillips ◽  
J. Huang ◽  
S. M. Dunn
Keyword(s):  
Planar Graph ◽  
Efficient Algorithm ◽  
Spatial Organization ◽  
Spatial Arrangement ◽  
Stereo Image ◽  
Image Model ◽  
Geometric Invariants ◽  
Point Set

In this paper we present an efficient algorithm for automatically finding the correspondence between pairs of stereo micrographs, the key step in forming a stereo image. The computation burden in this problem is solving for the optimal mapping and transformation between the two micrographs. In this paper, we present a sieve algorithm for efficiently estimating the transformation and correspondence.In a sieve algorithm, a sequence of stages gradually reduce the number of transformations and correspondences that need to be examined, i.e., the analogy of sieving through the set of mappings with gradually finer meshes until the answer is found. The set of sieves is derived from an image model, here a planar graph that encodes the spatial organization of the features. In the sieve algorithm, the graph represents the spatial arrangement of objects in the image. The algorithm for finding the correspondence restricts its attention to the graph, with the correspondence being found by a combination of graph matchings, point set matching and geometric invariants.

ЛАНДШАФТНАЯ ОРГАНИЗАЦИЯ БЕРЕГОВОЙ ГЕОСТРУКТУРЫ ОСТРОВА ШКОТА (ЗАЛИВ ПЕТРА ВЕЛИКОГО)

2019 ◽  
Author(s):  
K.S. Ganzei ◽  
V.V. Zharikov ◽  
N.F. Pshenichnikova ◽  
A.M. Lebedev ◽  
A.G. Kiselyova ◽  
...  
Keyword(s):  
Spatial Organization ◽  
Peter The Great Bay ◽  
Spatial Arrangement ◽  
Water Area ◽  
Marine Spatial Planning ◽  
Landscape Approach ◽  
Important Condition ◽  
Statistical Comparison ◽  
Peter The Great ◽  
Comprehensive Study

Важнейшим условием достижения устойчивого развития прибрежноморского природопользования в заливе Петра Великого системы является морское пространственное планирование. Основой для этого является информация о природных комплексах территории и акватории, полученная на основе ландшафтного подхода. Ключевым районом для изучения пространственной организации ландшафтов прибрежных геоструктур стала территория острова Шкота и его подводных склонов. Для наземных ландшафтов было описано 49 наблюдательных пунктов, 4 профиля были заложены для подводных ландшафтов описано 64 наблюдательных пункта, проложено 18 профилей. Выделено 22 вида ландшафтов, из них 16 наземных, 6 подводных. Берега острова сформированы преобладанием абразивноденудационного и абразивного типов. В результате всестороннего изучения показаны особенности пространственной организации воздушных и водных природных комплексов. Особенностью исследуемой территории является экспозиция дифференциации ландшафтов между юговосточной и северозападной частями острова, обусловленная муссонной природой климата. Результаты полевых и картографических работ послужили основой для выбора зон интенсивного, умеренного и ослабленного взаимодействия наземных и подводных ландшафтов. Пространственное расположение зон взаимодействия четко иллюстрируется значительными различиями экспозиции. Результаты статистического сравнения ландшафтов суши и мелководья, окружающего остров, на основе картометрических характеристик указывают на неоднородность геоструктуры острова, обусловленную, прежде всего, сочетанием ландшафтообразующих факторов. The most important condition for achieving sustainable development of coastalmarine environmental management in Peter the Great Bay is marine spatial planning. The basis for this is information about the natural complexes of the territory and water area, obtained based on the landscape approach. The main area for studying the spatial organization of landscapes of coastal geostructures was the territory of the island of Shkota and its underwater slopes. For terrestrial landscapes, 49 observation points were described, 4 profiles were laid 64 observation points were described for underwater landscapes, 18 profiles were laid. 22 species of landscapes have been identified, of which 16 are terrestrial, 6 are underwater. The shores of the island are formed by the predominance of abrasivedenudation and abrasive types. Because of a comprehensive study, features of the spatial organization of air and aquatic natural complexes are shown. A special feature of the study area is the exposure of the differentiation of landscapes between the southeastern and northwestern parts of the island, due to the monsoon nature of the climate. The results of field and cartographic works served as the basis for selecting areas of intense, moderate and weakened interaction of land and underwater landscapes. The spatial arrangement of the interaction zones is clearly illustrated by significant differences in exposure. The results of a statistical comparison of the land and shallow water landscapes surrounding the island, based on the cartometric characteristics, indicate the heterogeneity of the islands geostructure, primarily due to the combination of landscapeforming factors.

Sign in / Sign up

Export Citation Format

Share Document