More Like this

1. Cross-prediction-powered inference

   https://doi.org/10.1073/pnas.2322083121  

   Zrnic, Tijana; Candès, Emmanuel J (April 2024, Proceedings of the National Academy of Sciences)

   While reliable data-driven decision-making hinges on high-quality labeled data, the acquisition of quality labels often involves laborious human annotations or slow and expensive scientific measurements. Machine learning is becoming an appealing alternative as sophisticated predictive techniques are being used to quickly and cheaply produce large amounts of predicted labels; e.g., predicted protein structures are used to supplement experimentally derived structures, predictions of socioeconomic indicators from satellite imagery are used to supplement accurate survey data, and so on. Since predictions are imperfect and potentially biased, this practice brings into question the validity of downstream inferences. We introduce cross-prediction: a method for valid inference powered by machine learning. With a small labeled dataset and a large unlabeled dataset, cross-prediction imputes the missing labels via machine learning and applies a form of debiasing to remedy the prediction inaccuracies. The resulting inferences achieve the desired error probability and are more powerful than those that only leverage the labeled data. Closely related is the recent proposal of prediction-powered inference [A. N. Angelopoulos, S. Bates, C. Fannjiang, M. I. Jordan, T. Zrnic,Science382, 669–674 (2023)], which assumes that a good pretrained model is already available. We show that cross-prediction is consistently more powerful than an adaptation of prediction-powered inference in which a fraction of the labeled data is split off and used to train the model. Finally, we observe that cross-prediction gives more stable conclusions than its competitors; its CIs typically have significantly lower variability. 

   more » « less
2. Wearable self-powered biosensors

   https://doi.org/10.1016/j.coelec.2019.10.002  

   Reid, Russell C.; Mahbub, Ifana (February 2020, Current Opinion in Electrochemistry)
3. Artificial intelligence-powered electronic skin

   https://doi.org/10.1038/s42256-023-00760-z  

   Xu, Changhao; Solomon, Samuel A.; Gao, Wei (December 2023, Nature Machine Intelligence)
4. Chemically Powered Synthetic “Living” Systems

   https://doi.org/10.1016/j.chempr.2020.08.010  

   Gentile, Kayla; Somasundar, Ambika; Bhide, Ashlesha; Sen, Ayusman (September 2020, Chem)
5. Transpiration-powered desalination water bottle

   https://doi.org/10.1039/d1sm01470f  

   Cornish, Gracie A.; Eyegheleme, Ndidi L.; Hudson, Laurel S.; Troy, Kathleen J.; Vollen, Maia M.; Boreyko, Jonathan B. (February 2022, Soft Matter)

   Inspired by mangrove trees, we present a theoretical design and analysis of a portable desalinating water bottle powered by transpiration. The bottle includes an annular fin for absorbing solar heat, which is used to boost the evaporation rate of water from the interior synthetic leaf. This synthetic leaf comprises a nanoporous film deposited atop a supporting micromesh. Water evaporating from the leaf generates a highly negative Laplace pressure, which pulls the overlying source water across an upstream reverse osmosis membrane. Evaporated water is re-condensed in the bottom of the bottle for collection. The benefit of our hybrid approach to desalination is that reverse osmosis is spontaneously enabled by transpiration, while the thermal evaporation process is enhanced by heat localization and made more durable by pre-filtering the salt. We estimate that a 9.4 cm diameter bottle, with a 10 cm wide annular fin, could harvest about a liter of fresh water per day from ocean water. 

   more » « less