Machine Learning

The Role of Hyperparameters in Machine Learning Models and How to Tune Them

Hyperparameters critically influence how well machine learning models perform on unseen, out-of-sample data. Systematically comparing the performance of different hyperparameter settings will often go a long way in building confidence about a model's performance. However, analyzing 64 machine learning related manuscripts published in three leading political science journals (APSR, PA, and PSRM) between 2016 and 2021, we find that only 13 publications (20.31%) report the hyperparameters and also how they tuned them in either the paper or the appendix. We illustrate the dangers of cursory attention to model and tuning transparency in comparing machine learning models' capability to predict electoral violence from tweets. The tuning of hyperparameters and their documentation should become a standard component of robustness checks for machine learning models.

Private Post-GAN Boosting

Differentially private GANs have proven to be a promising approach for generating realistic synthetic data without compromising the privacy of individuals. However, due to the privacy-protective noise introduced in the training, the convergence of GANs becomes even more elusive, which often leads to poor utility in the output generator at the end of training. We propose Private post-GAN boosting (Private PGB), a differentially private method that combines samples produced by the sequence of generators obtained during GAN training to create a high-quality synthetic dataset. Our method leverages the Private Multiplicative Weights method (Hardt and Rothblum, 2010) and the discriminator rejection sampling technique (Azadi et al., 2019) for reweighting generated samples, to obtain high quality synthetic data even in cases where GAN training does not converge. We evaluate Private PGB on a Gaussian mixture dataset and two US Census datasets, and demonstrate that Private PGB improves upon the standard private GAN approach across a collection of quality measures. Finally, we provide a non-private variant of PGB that improves the data quality of standard GAN training.

Really Useful Synthetic Data -- A Framework to Evaluate the Quality of Differentially Private Synthetic Data

Recent advances in generating synthetic data that allow to add principled ways of protecting privacy -- such as Differential Privacy -- are a crucial step in sharing statistical information in a privacy preserving way. But while the focus has been on privacy guarantees, the resulting private synthetic data is only useful if it still carries statistical information from the original data. To further optimise the inherent trade-off between data privacy and data quality, it is necessary to think closely about the latter. What is it that data analysts want? Acknowledging that data quality is a subjective concept, we develop a framework to evaluate the quality of differentially private synthetic data from an applied researcher's perspective. Data quality can be measured along two dimensions. First, quality of synthetic data can be evaluated against training data or against an underlying population. Second, the quality of synthetic data depends on general similarity of distributions or specific tasks such as inference or prediction. It is clear that accommodating all goals at once is a formidable challenge. We invite the academic community to jointly advance the privacy-quality frontier.

How Cross-Validation Can Go Wrong and What to Do About it.

The introduction of new "machine learning" methods and terminology to political science complicates the interpretation of results. Even more so, when one term – like cross-validation – can mean very different things. We find different meanings of cross-validation in applied political science work. In the context of predictive modeling, cross-validation can be used to obtain an estimate of true error or as a procedure for model tuning. Using a single cross-validation procedure to obtain an estimate of the true error and for model tuning at the same time leads to serious misreporting of performance measures. We demonstrate the severe consequences of this problem with a series of experiments. We also observe this problematic usage of cross-validation in applied research. We look at Muchlinski et al. (2016) on the prediction of civil war onsets to illustrate how the problematic cross-validation can affect applied work. Applying cross-validation correctly, we are unable to reproduce their findings. We encourage researchers in predictive modeling to be especially mindful when applying cross-validation.

Poster Presentation - Deep Multiple Imputation - Using Mixture Density Networks to Impute Missing Values.