Filters








4,147 Hits in 4.2 sec

Track Seeding and Labelling with Embedded-space Graph Neural Networks [article]

Nicholas Choma, Daniel Murnane, Xiangyang Ju, Paolo Calafiura, Sean Conlon, Steven Farrell, Prabhat, Giuseppe Cerati, Lindsey Gray, Thomas Klijnsma, Jim Kowalkowski, Panagiotis Spentzouris (+7 others)
2020 arXiv   pre-print
The most promising of these solutions, graph neural networks (GNN), process the event as a graph that connects track measurements (detector hits corresponding to nodes) with candidate line segments between  ...  The embedded graphs feed into high-accuracy doublet and triplet classifiers, or can be used as an end-to-end track classifier by clustering in an embedded space.  ...  Graph neural network (GNN) [9, 10] models were then proposed and demonstrated to be effective at identifying tracks in realistic data [8, 11] .  ... 
arXiv:2007.00149v1 fatcat:n2amxcym4rginarfv7r4epf5cq

CTD2020: Learned Representations from Lower-Order Interactions for Efficient Clustering

Nicholas Choma
2020 Zenodo  
approaches with graph neural networks developed in the context of the Exa.TrkX project.  ...  We demonstrate how this framework fits in with both traditional clustering pipelines, and more advanced approaches such as graph neural networks.  ...  Refinement • Graph neural networks are expensive and require large footprints in GPU memory • Far more information content in doublet than in single hit • Key idea: Filter doublets with small MLP that  ... 
doi:10.5281/zenodo.4034394 fatcat:dh3wmbcsyjb5jawkl7gpfhqnte

CTD2020: Graph Neural Networks for Track Finding

Daniel Murnane
2020 Zenodo  
The most promising of these solutions, graph neural networks (GNN), process the event as a graph that connects track measurements (detector hits corresponding to nodes) with candidate line segments between  ...  Detector information can be associated with nodes and edges, enabling a GNN to propagate the embedded parameters around the graph and predict node-, edge- and graph-level observables.  ...  The Exa.TrkX Project 3 TRACK FINDING GOAL Given that graphs are a natural way to represent tracks, use learned embeddings and graph neural networks for sub-second processing of HL-LHC event data into  ... 
doi:10.5281/zenodo.4034347 fatcat:ro5qkiannrahzgkny5ylr2xzeu

CTD2020: Graph Neural Networks for Track Finding

Daniel Murnane
2020 Zenodo  
The most promising of these solutions, graph neural networks (GNN), process the event as a graph that connects track measurements (detector hits corresponding to nodes) with candidate line segments between  ...  Detector information can be associated with nodes and edges, enabling a GNN to propagate the embedded parameters around the graph and predict node-, edge- and graph-level observables.  ...  Graph neural network (GNN) [9, 10] models were then proposed and demonstrated to be effective at identifying tracks in realistic data [8, 11] .  ... 
doi:10.5281/zenodo.4088460 fatcat:5qzdpyuvkfau7chjcxza3rpkgm

CTD2020: ACTS Vertexing and Deep Learning Vertex Finding

Bastian Schlag
2020 Zenodo  
Constructing undirected, edge-weighted graphs from these results allows the subsequent usage of classical graph algorithms or graph neural networks for clustering tracks to vertex candidates.  ...  Learning a track representation in an embedding space in such a way that tracks emerging from a common vertex are close together while tracks from neighboring vertices are further separated from one another  ...  Siamese Neural Network Track embedding network Track embedding network x i x j d(h i ,h j ) Score Track Similarity Siamese Neural Network • one instance of embedding network used twice  ... 
doi:10.5281/zenodo.4034364 fatcat:ahoegd54xvdcxnpgta6cfkzu7a

Instance Embedding Transfer to Unsupervised Video Object Segmentation

Siyang Li, Bryan Seybold, Alexey Vorobyov, Alireza Fathi, Qin Huang, C.-C. Jay Kuo
2018 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition  
Thus, we adapt the instance networks trained on static images to video object segmentation and incorporate the embeddings with objectness and optical flow features, without model retraining or online fine-tuning  ...  The instance embedding network produces an embedding vector for each pixel that enables identifying all pixels belonging to the same object.  ...  This work was started when Siyang Li was an intern at Google and later continued at USC with support from Ittiam Systems.  ... 
doi:10.1109/cvpr.2018.00683 dblp:conf/cvpr/LiSVFHK18 fatcat:kkm4lr74kze4xkvod2pz4dpcby

Unsupervised Video Object Segmentation with Motion-Based Bilateral Networks [chapter]

Siyang Li, Bryan Seybold, Alexey Vorobyov, Xuejing Lei, C.-C. Jay Kuo
2018 Lecture Notes in Computer Science  
Then, we integrate the background estimate from the bilateral network with instance embeddings into a graph, which allows multiple frame reasoning with graph edges linking pixels from different frames.  ...  We classify graph nodes by defining and minimizing a cost function, and segment the video frames based on the node labels.  ...  The variants of embedding graph cut are evaluated by the J Mean of the final segmentation with seed labels propagated to all pixels.  ... 
doi:10.1007/978-3-030-01219-9_13 fatcat:mi6u2taunncavfps2lipvk6rue

Instance Embedding Transfer to Unsupervised Video Object Segmentation [article]

Siyang Li, Bryan Seybold, Alexey Vorobyov, Alireza Fathi, Qin Huang, C.-C. Jay Kuo
2018 arXiv   pre-print
Thus, we adapt the instance networks trained on static images to video object segmentation and incorporate the embeddings with objectness and optical flow features, without model retraining or online fine-tuning  ...  The instance embedding network produces an embedding vector for each pixel that enables identifying all pixels belonging to the same object.  ...  each seed in the embedding graph.  ... 
arXiv:1801.00908v2 fatcat:672yimc3ubb5jbu5uxeez3q2de

Neural Graph Machines: Learning Neural Networks Using Graphs [article]

Thang D. Bui, Sujith Ravi, Vivek Ramavajjala
2017 arXiv   pre-print
Label propagation is a powerful and flexible semi-supervised learning technique on graphs. Neural networks, on the other hand, have proven track records in many supervised learning tasks.  ...  In this work, we propose a training framework with a graph-regularised objective, namely "Neural Graph Machines", that can combine the power of neural networks and label propagation.  ...  Here, we advocate a training objective that uses graphs to augment neural network learning, and works with many forms of graphs and with any type of neural network.  ... 
arXiv:1703.04818v1 fatcat:3v4chn6e5nb7hnwlyzaucw4rmu

Performance of a geometric deep learning pipeline for HL-LHC particle tracking

Xiangyang Ju, Daniel Murnane, Paolo Calafiura, Nicholas Choma, Sean Conlon, Steven Farrell, Yaoyuan Xu, Maria Spiropulu, Jean-Roch Vlimant, Adam Aurisano, Jeremy Hewes, Giuseppe Cerati (+12 others)
2021 European Physical Journal C: Particles and Fields  
AbstractThe Exa.TrkX project has applied geometric learning concepts such as metric learning and graph neural networks to HEP particle tracking.  ...  Exa.TrkX's tracking pipeline groups detector measurements to form track candidates and filters them.  ...  The fourth stage of the pipeline is the training and inference of the graph neural network.  ... 
doi:10.1140/epjc/s10052-021-09675-8 fatcat:vy3cu4tzknc7nkufcctfclz3oe

Novel deep learning methods for track reconstruction [article]

Steven Farrell, Paolo Calafiura, Mayur Mudigonda, Prabhat, Dustin Anderson, Jean-Roch Vlimant, Stephan Zheng, Josh Bendavid, Maria Spiropulu, Giuseppe Cerati, Lindsey Gray, Jim Kowalkowski, Panagiotis Spentzouris, Aristeidis Tsaris
2018 arXiv   pre-print
The second set of models use Graph Neural Networks (GNNs) for the tasks of hit classification and segment classification.  ...  In the first set of models, Recurrent Neural Networks (RNNs) are used to extrapolate, build, and evaluate track candidates akin to Kalman Filter algorithms.  ...  We acknowledge NVIDIA, SuperMicro and the Kavli Foundation for their support of "iBanks".  ... 
arXiv:1810.06111v1 fatcat:7manc44gerh25kcufkbgwpavc4

Performance of a Geometric Deep Learning Pipeline for HL-LHC Particle Tracking [article]

Xiangyang Ju
2021 arXiv   pre-print
The Exa.TrkX project has applied geometric learning concepts such as metric learning and graph neural networks to HEP particle tracking.  ...  Exa.TrkX's tracking pipeline groups detector measurements to form track candidates and filters them.  ...  The fourth stage of the pipeline is the training and inference of the graph neural network.  ... 
arXiv:2103.06995v2 fatcat:d5jsnbocijhc7izr54ugbjskxe

Deep Learning and Its Application to LHC Physics

Dan Guest, Kyle Cranmer, Daniel Whiteson
2018 Annual Review of Nuclear and Particle Science  
The connections between machine learning and high energy physics data analysis are explored, followed by an introduction to the core concepts of neural networks, examples of the key results demonstrating  ...  This review is aimed at the reader who is familiar with high energy physics but not machine learning.  ...  K.C. is supported by the National Science Foundation (ACI-1450310 and PHY-1505463) and by the Moore-Sloan Data Science Environment at at NYU. LITERATURE CITED  ... 
doi:10.1146/annurev-nucl-101917-021019 fatcat:4ll2ex624jcutgimi5w7wya2bq

Bayesian Prior Learning via Neural Networks for Next-item Recommendation [article]

Manoj Reddy Dareddy, Zijun Xue, Nicholas Lin, Junghoo Cho
2022 arXiv   pre-print
Our novel approach of combining black-box style neural networks, known to be suitable for function approximation with Bayesian estimation methods have resulted in an innovative method that outperforms  ...  We train two neural networks to accurately predict the alpha & beta parameter values of the Beta distribution.  ...  five seeds. 4.5.2 Nearest Embedding.  ... 
arXiv:2205.05209v1 fatcat:xrxguyeprbgh3nldxa2tvn3kvy

Learning Geometric Equivalence between Patterns Using Embedding Neural Networks

Olga Moskvyak, Frederic Maire
2017 2017 International Conference on Digital Image Computing: Techniques and Applications (DICTA)  
Convolutional neural networks are used to learn a mapping from a face image space to an Euclidean space of a smaller dimension.  ...  The input of a Siamese network is a pair of images (P i , P j ) and a label y ij .  ... 
doi:10.1109/dicta.2017.8227457 dblp:conf/dicta/MoskvyakM17 fatcat:c25r2qgp3rg4rajbqxfzbcsbay
« Previous Showing results 1 — 15 out of 4,147 results