Improving Gradient-based DAG Learning by Structural Asymmetry

Wang, Yujie; Yang, Shuai; Guo, Xianjie; Yu, Kui

doi:10.1109/ickg52313.2021.00022

Cited by 1 publication

(2 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Two datasets were prepared in advance: a full dataset of 392 cases from which records containing missing values were deleted, and a short version of the dataset from which 100 cases were randomly selected. For the descriptive statistics of these two data sets, the median, the first quartile, and the third quartile for each variable were as follows: for the full version of the data, 'age,' 27 [23][24][25][26][27][28][29][30][31][32][33][34][35][36] The results of the experiment showed that both qLiNGAM and DirectLiNGAM with Gaussian kernel identified the causal relationship in Fig 6 . for both the full example dataset (392 examples) and the short version dataset (100 randomly selected examples). As shown in Fig 6, the left-hand path shows that 'age' affects 'insulin,' which, in turn, affects 'glucose,' and the right-hand path shows that 'age' directly affects 'glucose'.…”

Section: Experiments With Real-world Medical Data: Partmentioning

confidence: 99%

“…[ 32 ]. On the other hand, compared to scale-invariant methods such as DirectLiNGAM, continuous optimization-based approaches are susceptible to rescaling of variables, such as standardization, and the risk of estimating different causal graphs depending on the scale of variables [ 33 , 34 ] and the risk of detecting reversed edges between certain variables in causal graphs depending on the data have been pointed out [ 35 ]. Therefore, there could be room for improvement in the practical application of the constrained-based approach because it is difficult to know the true scale of the variables in advance, potentially limiting its application to real-world data.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Application of quantum computing to a linear non-Gaussian acyclic model for novel medical knowledge discovery

Kawaguchi

2023

PLoS ONE

View full text Add to dashboard Cite

Recently, the utilization of real-world medical data collected from clinical sites has been attracting attention. Especially as the number of variables in real-world medical data increases, causal discovery becomes more and more effective. On the other hand, it is necessary to develop new causal discovery algorithms suitable for small data sets for situations where sample sizes are insufficient to detect reasonable causal relationships, such as rare diseases and emerging infectious diseases. This study aims to develop a new causal discovery algorithm suitable for a small number of real-world medical data using quantum computing, one of the emerging information technologies attracting attention for its application in machine learning. In this study, a new algorithm that applies the quantum kernel to a linear non-Gaussian acyclic model, one of the causal discovery algorithms, is developed. Experiments on several artificial data sets showed that the new algorithm proposed in this study was more accurate than existing methods with the Gaussian kernel under various conditions in the low-data regime. When the new algorithm was applied to real-world medical data, a case was confirmed in which the causal structure could be correctly estimated even when the amount of data was small, which was not possible with existing methods. Furthermore, the possibility of implementing the new algorithm on real quantum hardware was discussed. This study suggests that the new proposed algorithm using quantum computing might be a good choice among the causal discovery algorithms in the low-data regime for novel medical knowledge discovery.

show abstract