Search

In this study, we propose a machine learning method called Distributionally Robust Safe Sample Screening (DRSSS). DRSSS aims to identify unnecessary training samples, even when the distribution of the training samples changes in the future. To achieve this, we effectively combine the distributionally robust (DR) paradigm, which aims to enhance model robustness against variations in data distribution, with the safe sample screening (SSS), which identifies unnecessary training samples prior to model training. Since we need to consider an infinite number of scenarios regarding changes in the distribution, we applied SSS because it does not require model training after the change of the distribution. In this paper, we employed the covariate shift framework to represent the distribution of training samples and reformulated the DR covariate-shift problem as a weighted empirical risk minimization problem, where the weights are subject to uncertainty within a predetermined range. By extending the existing SSS technique to accommodate this weight uncertainty, the DRSSS method is capable of reliably identifying unnecessary samples under any future distribution within a specified range. We provide a theoretical guarantee for the DRSSS method and validate its performance through numerical experiments on both synthetic and real-world datasets.

This integrated assessment modeling research analyzes what Korea's 2050 carbon neutrality would require for the national energy system and the role of the power sector concerning the availability of critical mitigation technologies. Our scenario-based assessments show that Korea's current policy falls short of what the nation's carbon-neutrality ambition would require. Across all technology scenarios examined in this study, extensive and rapid energy system transition is imperative, requiring the large-scale deployment of renewables and carbon capture & storage (CCS) early on and negative emission technologies (NETs) by the mid-century. Importantly, rapid decarbonization of the power sector that goes with rapid electrification of end-uses seems to be a robust national decarbonization strategy. Furthermore, we contextualize our net-zero scenario results using policy costs, requirements for natural resources, and the expansion rate of zero-carbon technologies. We find that the availability of nuclear power lowers the required expansion rate of renewables and CCS, alleviating any stress on terrestrial and geological systems. By contrast, the limited availability of CCS without nuclear power necessarily demands a very high penetration of renewables and significantly high policy compliance costs, which would decrease the feasibility of achieving the carbon neutrality target.

While modern deep neural networks are performant perception modules, performance (accuracy) alone is insufficient, particularly for safety-critical robotic applications such as self-driving vehicles. Robot autonomy stacks also require these otherwise blackbox models to produce reliable and calibrated measures of confidence on their predictions. Existing approaches estimate uncertainty from these neural network perception stacks by modifying network architectures, inference procedure, or loss functions. However, in general, these methods lack calibration, meaning that the predictive uncertainties do not faithfully represent the true underlying uncertainties (process noise). Our key insight is that calibration is only achieved by imposing constraints across multiple examples, such as those in a mini-batch; as opposed to existing approaches which only impose constraints per-sample, often leading to overconfident (thus miscalibrated) uncertainty estimates. By enforcing the distribution of outputs of a neural network to resemble a target distribution by minimizing an $f$-divergence, we obtain significantly better-calibrated models compared to prior approaches. Our approach, $f$-Cal, outperforms existing uncertainty calibration approaches on robot perception tasks such as object detection and monocular depth estimation over multiple real-world benchmarks., Comment: For more details about $f$-Cal, visit https://f-cal.github.io