Randomness In Neural Network Training: Characterizing The Impact of Tooling

By Donglin Zhuang, Xingyao Zhang, Shuaiwen Leon Song, Sara Hooker. 2021

97 citations recorded by the source corpus.

The quest for determinism in machine learning has disproportionately focused on characterizing the impact of noise introduced by algorithmic design choices. In this work, we address a less well understood and studied question: how does our choice of tooling introduce randomness to deep neural network training. We conduct large scale experiments across different types of hardware, accelerators, state of art networks, and open-source datasets, to characterize how tooling choices contribute to the level of non-determinism in a system, the impact of said non-determinism, and the cost of eliminating different sources of noise. Our findings are surprising, and suggest that the impact of non-determinism in nuanced. While top-line metrics such as top-1 accuracy are not noticeably impacted, model performance on certain parts of the data distribution is far more sensitive to the introduction of randomness. Our results suggest that deterministic tooling is critical for AI safety. However, we also find that the cost of ensuring determinism varies dramatically between neural network architectures and hardware types, e.g., with overhead up to $746\%$, $241\%$, and $196\%$ on a spectrum of widely used GPU accelerator architectures, relative to non-deterministic training. The source code used in this paper is available at https://github.com/usyd-fsalab/NeuralNetworkRandomness.

Subfields: Artificial Intelligence, Evaluation & Benchmarks

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Evaluation & Benchmarks

Citations in the corpus

Evaluating the Social Impact of Generative AI Systems in Systems and Society (2023, background): “These can include choices about many stages of the optimization process [237, 129]; privacy constraints [24], widely used compression techniques [109, 15, 169] and the choice hardware [273] have all been found to amplify harm on underrepresented protected attributes [28].”
Predictability and Surprise in Large Generative Models (2022, background): “Generally, machine learning experiments are not precisely predictable — complex models trained on complex data typically yield noisy or variable results [79, 17].”
JailbreakBench: An Open Robustness Benchmark for Jailbreaking Large Language Models (2024, background): “For locally run models, the only source of randomness comes from GPU computations (Zhuang et al., 2022), and is usually negligible.”
FAIR-Ensemble: When Fairness Naturally Emerges From Deep Ensembling (2023, background): “Most relevant to our work is (Qian et al., 2021; Zhuang et al., 2022; Madhyastha & Jain, 2019; Summers & Dinneen, 2021) that evaluates how stochasticity in training impacts fairness in DNN Systems.”
Deep Reinforcement Learning at the Edge of the Statistical Precipice (2021, background): “[116] Donglin Zhuang, Xingyao Zhang, Shuaiwen Leon Song, and Sara Hooker.”