Background and motivation
Test-time adaptation (TTA) updates a deployed model using only unlabeled target data, but standard entropy-based methods often rely on deterministic confidence and can reinforce overconfident errors under distribution shift.
Proposed method: MIRA overview
We propose MIRA, Mutual Information guided calibration for Reliable Test-Time Adaptation, a source-free online adaptation method that uses lightweight last-layer Gaussian perturbations to probe prediction reliability.
For each target batch, MIRA reuses a single feature extraction pass and samples perturbed classifiers around the source classifier, forming an efficient local predictive distribution.
Scale selection via mutual information and agreement
Rather than fitting a Bayesian posterior, MIRA treats the perturbation scale as a local probing radius and selects a batch-specific scale using mutual information (MI) and prediction agreement.
Reliability signal and adaptation behavior
The resulting perturbation-sensitivity signal identifies predictions that are confident and locally stable under classifier perturbations, while suppressing brittle predictions with high stochastic disagreement.
MIRA then performs adaptation with an MI-confidence weighted entropy loss, which down-weights perturbation-sensitive or low-confidence samples before they drive entropy minimization.
This weighted loss reduces the influence of brittle predictions that can otherwise cause harmful adaptation under distribution shift.
Practical requirements and plug-and-play property
The method requires no source data, retraining, model ensembles, or offline posterior fitting, and can be applied as a plug-and-play modification to entropy-minimization TTA.
Experiments and takeaway
Experiments on distribution-shift benchmarks show improvements in accuracy, calibration, and negative log-likelihood, suggesting that mutual-information-guided perturbation probing and weighted adaptation yield more robust online TTA than deterministic confidence.