Odd Erik Gundersen

Sayash Kapoor, Emily Cantrell, Kenny Peng et al. · princeton

Machine learning (ML) methods are proliferating in scientific research. However, the adoption of these methods has been accompanied by failures of validity, reproducibility, and generalizability. These failures can hinder scientific progress, lead to false consensus around invalid claims, and undermine the credibility of ML-based science. ML methods are often applied and fail in similar ways across disciplines. Motivated by this observation, our goal is to provide clear reporting standards for ML-based science. Drawing from an extensive review of past literature, we present the REFORMS checklist ($\textbf{Re}$porting Standards $\textbf{For}$ $\textbf{M}$achine Learning Based $\textbf{S}$cience). It consists of 32 questions and a paired set of guidelines. REFORMS was developed based on a consensus of 19 researchers across computer science, data science, mathematics, social sciences, and biomedical sciences. REFORMS can serve as a resource for researchers when designing and implementing a study, for referees when reviewing papers, and for journals when enforcing standards for transparency and reproducibility.

Thijs Snelleman, Peter Lundestad Lawrence, Holger H. Hoos et al.

Background. Reproducibility is essential to the scientific method, but reproduction is often a laborious task. Recent works have attempted to automate this process and relieve researchers of this workload. However, due to varying definitions of reproducibility, a clear problem statement is missing. Objectives. Create a generalisable problem statement, applicable to any empirical study. We hypothesise that we can represent any empirical study using a structure based on the scientific method and that this representation can be automatically extracted from any publication, and captures the essence of the study. Methods. We apply our definition of reproducibility as a problem statement for the automatisation of reproducibility by automatically extracting the hypotheses, experiments and interpretations of 20 studies and assess the quality based on assessments by the original authors of each study. Results. We create a dataset representing the reproducibility problem, consisting of the representation of 20 studies. The majority of author feedback is positive, for all parts of the representation. In a few cases, our method failed to capture all elements of the study. We also find room for improvement at capturing specific details, such as results of experiments. Conclusions. We conclude that our formulation of the problem is able to capture the concept of reproducibility in empirical AI studies across a wide range of subfields. Authors of original publications generally agree that the produced structure is representative of their work; we believe improvements can be achieved by applying our findings to create a more structured and fine-grained output in future work.

Odd Erik Gundersen, Kevin Coakley, Christine Kirkpatrick et al.

Background: Many published machine learning studies are irreproducible. Issues with methodology and not properly accounting for variation introduced by the algorithm themselves or their implementations are attributed as the main contributors to the irreproducibility.Problem: There exist no theoretical framework that relates experiment design choices to potential effects on the conclusions. Without such a framework, it is much harder for practitioners and researchers to evaluate experiment results and describe the limitations of experiments. The lack of such a framework also makes it harder for independent researchers to systematically attribute the causes of failed reproducibility experiments. Objective: The objective of this paper is to develop a framework that enable applied data science practitioners and researchers to understand which experiment design choices can lead to false findings and how and by this help in analyzing the conclusions of reproducibility experiments. Method: We have compiled an extensive list of factors reported in the literature that can lead to machine learning studies being irreproducible. These factors are organized and categorized in a reproducibility framework motivated by the stages of the scientific method. The factors are analyzed for how they can affect the conclusions drawn from experiments. A model comparison study is used as an example. Conclusion: We provide a framework that describes machine learning methodology from experimental design decisions to the conclusions inferred from them.

Waqas Ahmed, Sheeba Samuel, Kevin Coakley et al.

To enhance the reproducibility and reliability of deep learning models, we address a critical gap in current training methodologies: the lack of mechanisms that ensure consistent and robust performance across runs. Our empirical analysis reveals that even under controlled initialization and training conditions, the accuracy of the model can exhibit significant variability. To address this issue, we propose a Custom Loss Function (CLF) that reduces the sensitivity of training outcomes to stochastic factors such as weight initialization and data shuffling. By fine-tuning its parameters, CLF explicitly balances predictive accuracy with training stability, leading to more consistent and reliable model performance. Extensive experiments across diverse architectures for both image classification and time series forecasting demonstrate that our approach significantly improves training robustness without sacrificing predictive performance. These results establish CLF as an effective and efficient strategy for developing more stable, reliable and trustworthy neural networks.

Kevin Coakley, Christine R. Kirkpatrick, Odd Erik Gundersen

Reproducing published deep learning papers to validate their conclusions can be difficult due to sources of irreproducibility. We investigate the impact that implementation factors have on the results and how they affect reproducibility of deep learning studies. Three deep learning experiments were ran five times each on 13 different hardware environments and four different software environments. The analysis of the 780 combined results showed that there was a greater than 6% accuracy range on the same deterministic examples introduced from hardware or software environment variations alone. To account for these implementation factors, researchers should run their experiments multiple times in different hardware and software environments to verify their conclusions are not affected.

Aleksandra Jekic, Afroditi Natsaridou, Signe Riemer-Sørensen et al.

Approximating solutions to partial differential equations (PDEs) is fundamental for the modeling of dynamical systems in science and engineering. Physics-informed neural networks (PINNs) are a recent machine learning-based approach, for which many properties and limitations remain unknown. PINNs are widely accepted as inferior to traditional methods for solving PDEs, such as the finite element method, both with regard to computation time and accuracy. However, PINNs are commonly claimed to show promise in solving inverse problems and handling noisy or incomplete data. We compare the performance of PINNs in solving inverse problems with that of a traditional approach using the finite element method combined with a numerical optimizer. The models are tested on a series of increasingly difficult fluid mechanics problems, with and without noise. We find that while PINNs may require less human effort and specialized knowledge, they are outperformed by the traditional approach. However, the difference appears to decrease with higher dimensions and more data. We identify common failures during training to be addressed if the performance of PINNs on noisy inverse problems is to become more competitive.

Odd Erik Gundersen, Odd Cappelen, Martin Mølnå et al.

A reproducibility crisis has been reported in science, but the extent to which it affects AI research is not yet fully understood. Therefore, we performed a systematic replication study including 30 highly cited AI studies relying on original materials when available. In the end, eight articles were rejected because they required access to data or hardware that was practically impossible to acquire as part of the project. Six articles were successfully reproduced, while five were partially reproduced. In total, 50% of the articles included was reproduced to some extent. The availability of code and data correlate strongly with reproducibility, as 86% of articles that shared code and data were fully or partly reproduced, while this was true for 33% of articles that shared only data. The quality of the data documentation correlates with successful replication. Poorly documented or miss-specified data will probably result in unsuccessful replication. Surprisingly, the quality of the code documentation does not correlate with successful replication. Whether the code is poorly documented, partially missing, or not versioned is not important for successful replication, as long as the code is shared. This study emphasizes the effectiveness of open science and the importance of properly documenting data work.

Håkon Hanisch Kjærnli, Lluis Mas-Ribas, Hans Jakob Håland et al.

When deploying time series forecasting models based on machine learning to real world settings, one often encounter situations where the data distribution drifts. Such drifts expose the forecasting models to out-of-distribution (OOD) data, and machine learning models lack robustness in these settings. Robustness can be improved by using deep generative models or genetic algorithms to augment time series datasets, but these approaches lack interpretability and are computationally expensive. In this work, we develop an interpretable and simple framework for generating time series. Our method combines time-series decompositions with analytic functions, and is able to generate time series with characteristics matching both in- and out-of-distribution data. This approach allows users to generate new time series in an interpretable fashion, which can be used to augment the dataset and improve forecasting robustness. We demonstrate our framework through EXPRTS, a visual analytics tool designed for univariate time series forecasting models and datasets. Different visualizations of the data distribution, forecasting errors and single time series instances enable users to explore time series datasets, apply transformations, and evaluate forecasting model robustness across diverse scenarios. We show how our framework can generate meaningful OOD time series that improve model robustness, and we validate EXPRTS effectiveness and usability through three use-cases and a user study.

Odd Erik Gundersen

Reproducibility is a confused terminology. In this paper, I take a fundamental view on reproducibility rooted in the scientific method. The scientific method is analysed and characterised in order to develop the terminology required to define reproducibility. Further, the literature on reproducibility and replication is surveyed, and experiments are modeled as tasks and problem solving methods. Machine learning is used to exemplify the described approach. Based on the analysis, reproducibility is defined and three different types of reproducibility as well as four degrees of reproducibility are specified.