A tensor-completion model with attention and a weighted loss is reported to fill missing DrugMatrix gene-expression values more accurately than 2D matrix factorization and CP decomposition.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Completion of the DrugMatrix Toxicogenomics Database using 3-Dimensional Tensors
A tensor-completion model with attention and a weighted loss is reported to fill missing DrugMatrix gene-expression values more accurately than 2D matrix factorization and CP decomposition.