Oxygen vacancy concentration in metal oxides is crucial to gas sensing performance. This study introduces a novel strategy of oxygen vacancies modulation on CoFe2O4, by two-step controllable incorporation of TiO2@MXene and carbon. In the first step, MXene simultaneously acts as an incorporation agent and a titanium source to grow TiO2 and regulate the oxygen vacancy concentration in CoFe2O4. The fabricated n-n heterojunctions of TiO2/CoFe2O4 further induced a gradient distribution of oxygen vacancies through energy band bending. And a second step, combining glucose as a carbon source, further increases the oxygen vacancy concentration in CoFe2O4 by chemical reduction during a hydrothermal process, leading to the formation of a CoFe2O4-TiO2@MXene-C nanostructured composite. Through controlling the glucose content, the ratio of Co3+ to Co2+ within the composite can be sequentially adjusted, which allows for the further regulation of oxygen vacancy concentration on the composite surface. The effectiveness of this two-step incorporation is demonstrated through enhanced acetone sensing performance, providing valuable insights into the fabrication of high-performance metal oxide-based gas sensors via controlled oxygen vacancy modulation.

AI applications powered by deep learning models are increasingly run natively at edge. A deployed model not only encounters continuously evolving input distributions (domains) but also faces adversarial attacks from third-party. This necessitates adapting the model to shifting domains to maintain high natural accuracy, while avoiding degrading the model's robust accuracy. However, existing domain adaptation and adversarial attack preventation techniques often have conflicting optimization objectives and they rely on time-consuming training process. This paper presents RobustDA, an on-device lightweight approach that co-optimizes natural and robust accuracies in model retraining. It uses a set of low-rank adapters to retain all learned domains' knowledge with small overheads. In each model retraining, RobustDA constructs an adapter to separate domain-related and robust-related model parameters to avoid their conflicts in updating. Based on the retained knowledge, it quickly generates adversarial examples with high-quality pseudo-labels and uses them to accelerate the retraining process. We demonstrate that, comparing against 14 state-of-the-art DA techniques under 7 prevalent adversarial attacks on edge devices, the proposed co-optimization approach improves natural and robust accuracies by 6.34% and 11.41% simultaneously. Under the same accuracy, RobustDA also speeds up the retraining process by 4.09x.