A new technical paper titled “Post-hoc Uncertainty Learning using a Dirichlet Meta-Model” was published (preprint) by researchers at MIT, University of Florida, and MIT-IBM Watson AI Lab (IBM Research).
The work demonstrates how to quantify the level of certainty in its predictions, while using less compute resources. “Uncertainty quantification is essential for both developers and users of machine-learning models. Developers can utilize uncertainty measurements to help develop more robust models, while for users, it can add another layer of trust and reliability when deploying models in the real world. Our work leads to a more flexible and practical solution for uncertainty quantification,” says Maohao Shen, an EE-CS graduate student and lead author of a paper in this MIT news article.
Find the technical paper here. December 2022.
Shen, Maohao, et al. “Post-hoc Uncertainty Learning using a Dirichlet Meta-Model.” arXiv preprint arXiv:2212.07359 (2022).
Related
AI Power Consumption Exploding
Exponential increase is not sustainable. But where is it all going?
Where And Why AI Makes Sense In Cars
True autonomy won’t happen quickly, but there are a lot of useful features that will be adopted in the interim.
Leave a Reply