Gain-loss ratio of storing intermediate data from workflows
Sequentially, the systematic processing of a significant amount of data can be necessary for input datasets to get desired outputs. In a workflow management system(WMS), usually, users build workflows by manually selecting and interconnecting various modules concerning some particular tasks. Thus, a system of automatically suggesting the appropriate intermediate datasets for modules and a suggestion for the decision of saving intermediate states will be helpful in a pipeline building process. This work investigates a technique for both automatically suggesting the intermediate datasets to use and store through mining and analyzing association rules from the previously developed workflows. Investigation on workflows shows that the association rule mining technique can help us to suggest subsequent modules for retrieving and storing data and explain them with gain-loss ratios.
Code (0)
등록된 구현이 없습니다.
Tasks
ManagementSimilar Papers 제목 키워드 기반
RNN-Transducer-based Losses for Speech Recognition on Noisy Targets
Training speech recognition systems on noisy transcripts is a significant challenge in industrial pipelines, where datasets are enormous and ensuring accurate transcription for every instance is difficult. In this work, …
speech-recognitionSpeech RecognitionIntermediate direct preference optimization
We propose the intermediate direct preference optimization (DPO) method to calculate the DPO loss at selected intermediate layers as an auxiliary loss for finetuning large language models (LLMs). The conventional DPO met…
InterMPL: Momentum Pseudo-Labeling with Intermediate CTC Loss
This paper presents InterMPL, a semi-supervised learning method of end-to-end automatic speech recognition (ASR) that performs pseudo-labeling (PL) with intermediate supervision. Momentum PL (MPL) trains a connectionist …
Automatic Speech RecognitionAutomatic Speech Recognition (ASR)Decoderspeech-recognition+1Unsupervised Training for Deep Speech Source Separation with Kullback-Leibler Divergence Based Probabilistic Loss Function
In this paper, we propose a multi-channel speech source separation with a deep neural network (DNN) which is trained under the condition that no clean signal is available. As an alternative to a clean signal, the propose…
Batchnorm Allows Unsupervised Radial Attacks
The construction of adversarial examples usually requires the existence of soft or hard labels for each instance, with respect to which a loss gradient provides the signal for construction of the example. We show that fo…