Patent
Dynamic network configuration
Larsson Tony,Haraldson Johan,Isaksson Martin +2 more
- 10 Sep 2020
TL;DR: In this article, a method for dynamically configuring a network for training a machine learning model is proposed, which is performed at a computing device communicatively coupled to the network and comprises the following steps: selecting (202) client computing devices to participate in training the model; determining (204) a first value of an evaluation metric of the model based on the selected client Computing devices, determining (206) the presence of an adjustment trigger, adjusting (208) the number of client computing device used to determine the value of the evaluation metric in response to the adjustment trigger; and
read more
Abstract: A method for dynamically configuring a network (100) is proposed for training a machine learning model. The network comprises a server computing device (102) and a plurality of client computing devices (104). The method is performed at a computing device communicatively coupled to the network and comprises the following steps: selecting (202) client computing devices to participate in training the model; determining (204) a first value of an evaluation metric of the model based on the selected client computing devices; determining (206) the presence of an adjustment trigger; adjusting (208) the number of client computing devices used to determine the value of the evaluation metric in response to the adjustment trigger; determining (210) a second value of the evaluation metric based on the adjusted number of client computing devices; and setting (212) the number of client computing devices participating in training the model accordingly.
read more
Chat with Paper
AI Agents for this Paper
Find similar papers on Google Scholar, PubMed and Arxiv
Write a critical review of this paper
Analyze citations of this paper to find unaddressed research gaps
References
•Posted Content
Federated Learning of Deep Networks using Model Averaging
TL;DR: This work presents a practical method for the federated learning of deep networks that proves robust to the unbalanced and non-IID data distributions that naturally arise, and allows high-quality models to be trained in relatively few rounds of communication.
1.4K
Patent
Dynamic load balancing and scaling of allocated cloud resources in an enterprise network
Li Erran Li,Thomas Y. Woo +1 more
- 15 Sep 2010
TL;DR: In this paper, a workload distribution system for an enterprise network (101) extended in to a cloud network (102) and a related method is described, where the enterprise network employs one or more load balancers in both a private enterprise network and a public cloud network to distribute work among the servers in both networks based on criteria such as overall system performance and costs.
212
Similarity search on time series based on threshold queries
Johannes Aßfalg,Hans-Peter Kriegel,Peer Kröger,Peter Kunath,Alexey Pryakhin,Matthias Renz +5 more
- 26 Mar 2006
TL;DR: This paper presents a new efficient access method which uses the fact that only partial information of the time series is required at query time and introduces the novel concept of threshold queries in time series databases which report those time series exceeding a user-defined query threshold at similar time frames compared to the query time series.
Patent
Central processing unit resource allocation method and computing node
Shan Weihua,Liu Jintao,Li Houqing +2 more
- 17 Oct 2014
TL;DR: In this paper, a method for allocating a central processing unit resource to a virtual machine, including determining, according to a change in the number of virtual machines in an advanced resource pool, the amount of allocated physical cores in the default resource pool.
20
Patent
Machine for development and deployment of analytical models
Desmond Duggan,Qian Zhu,Teresa Tung,Kang Jaeyoung Christopher,Sun Wenjia +4 more
- 30 Jun 2016
TL;DR: In this paper, the authors propose a system and interface to allow domain experts and other users to develop, deploy, and iterate on analytical models by exposing analytical model configuration parameters to a user.
18