Filters








846 Hits in 6.3 sec

Automatic Discovery and Optimization of Parts for Image Classification [article]

Sobhan Naderi Parizi, Andrea Vedaldi, Andrew Zisserman, Pedro Felzenszwalb
2015 arXiv   pre-print
We generate an initial pool of parts by randomly sampling part candidates and selecting a good subset using L1/L2 regularization.  ...  This lets us do away with engineered heuristics. We also introduce the notion of "negative parts", intended as parts that are negatively correlated with one or more classes.  ...  We do part selection and joint training using 60 PCA coefficients of the CNN features. We select parts from an initial pool of 1000 random parts.  ... 
arXiv:1412.6598v2 fatcat:mecqnldbnvevjnyxraaylvkqbu

Profitable neglect

Jerome A. Kroth
1981 Contemporary Psychology  
Gauss once said, “I have had my solutions for a long time; but I do not yet know how I am to arrive at them.” Let us all join together, Jerome. Let us, like Gauss, com- bine our mutual talents.  ...  The key element in Lazare’s approach is the eliciting of the patient’s requests. Not only is the request not ignored, it is actively evoked, so that the therapist can respond to it.  ... 
doi:10.1037/020405 fatcat:merljt2a4neonahnyzwqq7yjhq

Zero-shot Task Transfer for Invoice Extraction via Class-aware QA Ensemble [article]

Prithiviraj Damodaran, Prabhkaran Singh, Josemon Achankuju
2021 arXiv   pre-print
We present VESPA, an intentionally simple yet novel zero-shot system for layout, locale, and domain agnostic document extraction.  ...  In spite of the availability of large corpora of documents, the lack of labeled and validated datasets makes it a challenge to discriminatively train document extraction models for enterprises.  ...  Once scored we use a heuristic based approach based on the range of the scores 2 to denoise and/or sharpen the image.  ... 
arXiv:2108.06069v1 fatcat:43uxojmghfbevozxw2yjcfrj34

Computer-Aided Geometric Sensitivity Analysis of Plane Trusses with a Square Grid Topology

Péter Dóbé, Gábor Domokos
2014 Periodica Polytechnica Electrical Engineering and Computer Science  
If necessary, however, it is easy to convert a supported truss into an unsupported one with similar properties [5] . The topology of the truss is described by the undirected topology graph G(V, E).  ...  A truss is said to be minimally rigid if it is rigid, and by removing any bar, it becomes non-rigid. These concepts are illustrated in Figure 1 .  ...  For non-degenerate geometries, the bars in the influenced zone of a selected joint correspond to the edges in the rigid core of the vertex in the topology graph matching the joint.  ... 
doi:10.3311/ppee.2207 fatcat:s5ddtqh6nje6dfrloa5o4hwlei

A Moment-Matching Method to Study the Variability of Phenomena Described by Partial Differential Equations

Jean-Frédéric Gerbeau, Damiano Lombardi, Eliott Tixier
2018 SIAM Journal on Scientific Computing  
Given an adequate model, it is possible to account for this variability by allowing some parameters to adopt a stochastic behavior.  ...  This inverse procedure is improved by incorporating an algorithm that selects a subset of the model DOFs that both reduces its computational cost and increases its robustness.  ...  Least-squares moment-matching An alternative to the present approach is to directly minimize the moment difference using a least-squares method.  ... 
doi:10.1137/16m1103476 fatcat:2jn455pu7jcdhpz2tbvf4pivo4

EMBnet.journal 17.2

EMBnet Journal
2011 EMBnet journal  
Funding: I wish to thank the European Commission for its support to projects EGEE (INFSO-RI-031688) and EMBRACE (LHSG-CT-2004-512092), which made this work possible.  ...  We also would like to thank Erik Bongcam-Rudloff for organising the joint EMBnet-EMBRACE workshop on creating web services for Bioinformatics (Uppsala, Sweden, 2008 Acknowledgements The author wishes  ...  Superclusteroid also detects complexes that do not match any confirmed complex in MIPS database.  ... 
doi:10.14806/ej.17.2.257 fatcat:77vndhd5afdl5ggw3bgywwxipy

DiploCloud: Efficient and Scalable Management of RDF Data in the Cloud

Marcin Wylot, Philippe Cudre-Mauroux
2016 IEEE Transactions on Knowledge and Data Engineering  
In spite of its seemingly simple data model, RDF actually encodes rich and complex graphs mixing both instance and schema-level data.  ...  Contrary to previous approaches, DiploCloud runs a physiological analysis of both instance and schema information prior to partitioning the data.  ...  They divide a query to basic graph patterns [17] which are then matched to the stored RDF data. A number of further approaches propose to store RDF data by taking advantage of its graph structure.  ... 
doi:10.1109/tkde.2015.2499202 fatcat:mxxfq3jxwbbwllpbrjpmag3tau

Parallel Genomic Sequence-Searching on an Ad-Hoc Grid: Experiences, Lessons Learned, and Implications

Mark Gardner, Wu-chun Feng, Jeremy Archuleta, Heshan Lin, Xiaosong Ma
2006 ACM/IEEE SC 2006 Conference (SC'06)  
The Basic Local Alignment Search Tool (BLAST) allows bioinformaticists to characterize an unknown sequence by comparing it against a database of known sequences.  ...  Hence, we assembled GreenGene, an ad-hoc grid that was constructed "on the fly" from donated computational, network, and storage resources during last year's SC|05.  ...  This work is supported in part by a DOE ECPI Award (DEFG02-05ER25685), and Xiaosong Ma's joint appointment between NCSU and ORNL.  ... 
doi:10.1109/sc.2006.46 fatcat:odls4ycryvgebk5mjc3pkkiqzu

Entropic Statistical Description of Big Data Quality in Hotel Customer Relationship Management

Lydia González-Serrano, Pilar Talón-Ballestero, Sergio Muñoz-Romero, Cristina Soguero-Ruiz, José Luis Rojo-Álvarez
2019 Entropy  
of duplicated customers, which has not received much attention in recent literature, probably and partly because it is not an easy-to-state problem in statistical terms.  ...  development of theoretically principled approaches.  ...  The managerial criterion to select this test instance can be because it represents a relevant customer, or it is just a new customer to be introduced into the CRM and to be checked in advance that it is  ... 
doi:10.3390/e21040419 pmid:33267133 pmcid:PMC7514908 fatcat:lbm57vicfbdpbjqshvm6y35cym

SYSTEMATIC ERRORS IN LOW-LATENCY GRAVITATIONAL WAVE PARAMETER ESTIMATION IMPACT ELECTROMAGNETIC FOLLOW-UP OBSERVATIONS

Tyson B. Littenberg, Ben Farr, Scott Coughlin, Vicky Kalogera
2016 Astrophysical Journal  
This suggests that electromagnetic observing campaigns should not take a strict approach to selecting which LIGO/Virgo candidates warrant follow-up observations based on low-latency mass estimates.  ...  Including simple precession, though more computationally costly, mitigates biases except for signals with extreme precession effects.  ...  It is generally assumed that the gamma rays are emitted along a narrow, relativistic jet with an opening angle of order 10 • .  ... 
doi:10.3847/0004-637x/820/1/7 fatcat:mjd5g4z75zhwzpx3i65bvfpwxm

Scalable online first-order monitoring

Joshua Schneider, David Basin, Frederik Brix, Srđan Krstić, Dmitriy Traytel
2021 International Journal on Software Tools for Technology Transfer (STTT)  
We implement these techniques in an automatic data slicer based on Apache Flink and empirically evaluate its performance using two tools—MonPoly and DejaVu—to monitor the substreams.  ...  Because monitoring is not embarrassingly parallel in general, slicing can lead to data duplication.  ...  Note that the righthand side of the equation has the same structure as the one for the joint data slicerŜ f (e) in Definition 4 once we replace matches(q, e) with matches(ϕ, e).  ... 
doi:10.1007/s10009-021-00607-1 fatcat:uxlpc3zy6ncclntqfdjdlz7u6i

Exploiting Neuron and Synapse Filter Dynamics in Spatial Temporal Learning of Deep Spiking Neural Network

Haowen Fang, Amar Shrestha, Ziyi Zhao, Qinru Qiu
2020 Proceedings of the Twenty-Ninth International Joint Conference on Artificial Intelligence  
by its temporal structures.  ...  Their accuracy outperforms state-of-the-art approaches.  ...  Any opinions, findings, or recommendations are those of the authors and do not necessarily reflect the view of the DARPA or the US government.  ... 
doi:10.24963/ijcai.2020/384 dblp:conf/ijcai/XiongR20 fatcat:i6bh3gis35bxvinnp67wsk3rm4

Near-Optimal Budgeted Data Exchange for Distributed Loop Closure Detection

Yulun Tian, Kasra Khosoussi, Matthew Giamou, Jonathan How, Jonathan Kelly
2018 Robotics: Science and Systems XIV  
We approach this problem from different perspectives and leverage existing results on monotone submodular maximization to provide efficient approximation algorithms with performance guarantees.  ...  ., battery and bandwidth) to exchange sensory data.  ...  Recall that once an observation is broadcasted, other robots can collectively verify all potential loop closures involving that observation (i.e., incident to the selected vertex); see Figure 1 and  ... 
doi:10.15607/rss.2018.xiv.071 dblp:conf/rss/TianKGHK18 fatcat:bteb5pn3njaare7i7vnhctheme

Make it stand

Romain Prévost, Emily Whiting, Sylvain Lefebvre, Olga Sorkine-Hornung
2013 ACM Transactions on Graphics  
We propose to assist users in producing novel, properly balanced designs by interactively deforming an existing model.  ...  It is therefore not surprising that many 3D models stand in impossibly balanced configurations. As long as the models remain in a computer this is of no consequence: the laws of physics do not apply.  ...  This work was supported in part by an SNF award 200021_137879, ERC grant ShapeForge (StG-2012-307877), ERC grant iModel (StG-2012-306877) and a gift from Adobe Research.  ... 
doi:10.1145/2461912.2461957 fatcat:maj7jqi64zce3dmrxgldwfsgku

Near-Optimal Budgeted Data Exchange for Distributed Loop Closure Detection [article]

Yulun Tian, Kasra Khosoussi, Matthew Giamou, Jonathan P. How and Jonathan Kelly
2018 arXiv   pre-print
We approach this problem from different perspectives and leverage existing results on monotone submodular maximization to provide efficient approximation algorithms with performance guarantees.  ...  ., battery and bandwidth) to exchange sensory data.  ...  Recall that once an observation is broadcasted, other robots can collectively verify all potential loop closures involving that observation (i.e., incident to the selected vertex); see Figure 1 and  ... 
arXiv:1806.00188v1 fatcat:mx6ex6ysqrba3ht6znpilbwuom
« Previous Showing results 1 — 15 out of 846 results