Early-Exit Neural Networks with Nested Prediction Sets
This addresses uncertainty quantification in EENNs for safety-critical domains, offering a novel solution to a specific bottleneck.
The paper tackles the problem of unreliable uncertainty estimates in early-exit neural networks (EENNs) for safety-critical applications, showing that standard methods like conformal prediction produce non-nested prediction sets across exits. It proposes using anytime-valid confidence sequences (AVCSs) to ensure nested sets, enabling fast and safe predictive modeling.
Early-exit neural networks (EENNs) enable adaptive and efficient inference by providing predictions at multiple stages during the forward pass. In safety-critical applications, these predictions are meaningful only when accompanied by reliable uncertainty estimates. A popular method for quantifying the uncertainty of predictive models is the use of prediction sets. However, we demonstrate that standard techniques such as conformal prediction and Bayesian credible sets are not suitable for EENNs. They tend to generate non-nested sets across exits, meaning that labels deemed improbable at one exit may reappear in the prediction set of a subsequent exit. To address this issue, we investigate anytime-valid confidence sequences (AVCSs), an extension of traditional confidence intervals tailored for data-streaming scenarios. These sequences are inherently nested and thus well-suited for an EENN's sequential predictions. We explore the theoretical and practical challenges of using AVCSs in EENNs and show that they indeed yield nested sets across exits. Thus our work presents a promising approach towards fast, yet still safe, predictive modeling