Learning Implicit Tasks for Patient-Specific Risk Modeling in ICU
Authors: Nozomi Nori, Hisashi Kashima, Kazuto Yamashita, Susumu Kunisawa, Yuichi Imanaka
AAAI 2017 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We demonstrate the effectiveness of our proposed method using a dataset collected from a hospital. Our method achieved higher predictive performance compared with a single-task learning method, the de facto standard, and several multi-task learning methods including a recently proposed method for ICU mortality risk prediction. |
| Researcher Affiliation | Academia | Nozomi Nori Graduate School of Informatics, Kyoto University Hisashi Kashima Graduate School of Informatics, Kyoto University Kazuto Yamashita Graduate School of Medicine, Kyoto University Susumu Kunisawa Graduate School of Medicine, Kyoto University Yuichi Imanaka Graduate School of Medicine, Kyoto University |
| Pseudocode | No | The paper describes the optimization problem and derivatives but does not include structured pseudocode or algorithm blocks. |
| Open Source Code | No | The paper does not provide any links to open-source code for the described methodology or state that the code is publicly available. |
| Open Datasets | No | We used a dataset from a hospital in Japan. All the patients in the dataset underwent ICU treatment at some point during their hospital stay. This dataset was constructed as part of the Quality Indicator/Improvement Project (Lee et al. 2011). |
| Dataset Splits | Yes | We randomly sampled 60% of the patients to create the training dataset and used the remaining 40% for evaluation. ... The hyperparameters were tuned by three-fold cross validation in the training dataset. |
| Hardware Specification | No | The paper does not specify any hardware details such as CPU, GPU models, or memory used for running the experiments. |
| Software Dependencies | No | The paper mentions applying the LBFGS optimizer and using logistic regression but does not provide specific version numbers for these or any other software dependencies. |
| Experiment Setup | Yes | For our proposed method, λ1 was tuned among {0, 10 3, 10 1}, and λ2 was set to 10 5. The number of latent tasks K was tuned among {22, 23, 24}. |