Zobrazeno 1 - 10
of 21
pro vyhledávání: '"Danieli, Federico"'
Autor:
Ramapuram, Jason, Danieli, Federico, Dhekane, Eeshan, Weers, Floris, Busbridge, Dan, Ablin, Pierre, Likhomanenko, Tatiana, Digani, Jagrit, Gu, Zijin, Shidani, Amitis, Webb, Russ
Attention is a key part of the transformer architecture. It is a sequence-to-sequence mapping that transforms each sequence element into a weighted sum of values. The weights are typically obtained as the softmax of dot products between keys and quer
Externí odkaz:
http://arxiv.org/abs/2409.04431
Parallelization techniques have become ubiquitous for accelerating inference and training of deep neural networks. Despite this, several operations are still performed in a sequential manner. For instance, the forward and backward passes are executed
Externí odkaz:
http://arxiv.org/abs/2309.16318
Fully-test-time adaptation (F-TTA) can mitigate performance loss due to distribution shifts between train and test data (1) without access to the training data, and (2) without knowledge of the model training procedure. In online F-TTA, a pre-trained
Externí odkaz:
http://arxiv.org/abs/2309.03964
This work develops a novel all-at-once space-time preconditioning approach for resistive magnetohydrodynamics (MHD), with a focus on model problems targeting fusion reactor design. We consider parallel-in-time due to the long time domains required to
Externí odkaz:
http://arxiv.org/abs/2309.00768
Autor:
Suau, Xavier, Danieli, Federico, Keller, T. Anderson, Blaas, Arno, Huang, Chen, Ramapuram, Jason, Busbridge, Dan, Zappella, Luca
Multiview Self-Supervised Learning (MSSL) is based on learning invariances with respect to a set of input transformations. However, invariance partially or totally removes transformation-related information from the representations, which might harm
Externí odkaz:
http://arxiv.org/abs/2306.16058
Autor:
Danieli, Federico, MacLachlan, Scott
Time-parallel algorithms seek greater concurrency by decomposing the temporal domain of a Partial Differential Equation (PDE), providing possibilities for accelerating the computation of its solution. While parallelisation in time has allowed remarka
Externí odkaz:
http://arxiv.org/abs/2104.09404
Parallel-in-time methods have become increasingly popular in the simulation of time-dependent numerical PDEs, allowing for the efficient use of additional MPI processes when spatial parallelism saturates. Most methods treat the solution and paralleli
Externí odkaz:
http://arxiv.org/abs/2101.07003
Two of the most popular parallel-in-time methods are Parareal and multigrid-reduction-in-time (MGRIT). Recently, a general convergence theory was developed in Southworth (2019) for linear two-level MGRIT/Parareal that provides necessary and sufficien
Externí odkaz:
http://arxiv.org/abs/2010.11879
Akademický článek
Tento výsledek nelze pro nepřihlášené uživatele zobrazit.
K zobrazení výsledku je třeba se přihlásit.
K zobrazení výsledku je třeba se přihlásit.
Autor:
Danieli, Federico, MacLachlan, Scott
Time-parallel algorithms seek greater concurrency by decomposing the temporal domain of a partial differential equation, providing possibilities for accelerating the computation of its solution. While parallelisation in time has allowed remarkable sp
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=od_______386::7276c3e6a0a271b029a65fa4d59bf350
http://epub.oeaw.ac.at/?arp=buecher/Organisationseinheiten/_id105092_/ETNA/etna_Vol_58/pp43-65.pdf
http://epub.oeaw.ac.at/?arp=buecher/Organisationseinheiten/_id105092_/ETNA/etna_Vol_58/pp43-65.pdf