More Like this

1. Tragopogon dubius : Multiple introductions to North America and the formation of the New World tetraploids

   https://doi.org/10.1002/tax.12743  

   Soltis, Douglas E.; Mavrodiev, Evgeny V.; Gitzendanner, Matthew A.; Alexeev, Yuri E.; Godden, Grant T.; Soltis, Pamela S. (December 2022, TAXON)
2. America, Jump-Started: World War II R&D and the Takeoff of the US Innovation System

   https://doi.org/10.1257/aer.20221365  

   Gross, Daniel P; Sampat, Bhaven N (December 2023, American Economic Review)

   During World War II, the US government’s Office of Scientific Research and Development (OSRD) supported one of the largest public investments in applied R&D in US history. Using data on all OSRD-funded invention, we show this shock had a formative impact on the US innovation system, catalyzing technology clusters across the country, with accompanying increases in high-tech entrepreneur-ship and employment. These effects persist until at least the 1970s and appear to be driven by agglomerative forces and endogenous growth. In addition to creating technology clusters, wartime R&D permanently changed the trajectory of overall US innovation in the direction of OSRD-funded technologies. 

   more » « less
3. Learning to Manage the Multirisk World

   https://doi.org/10.1111/risa.13629  

   Wiener, Jonathan B. (November 2020, Risk Analysis)

   null (Ed.)
4. Learning to Superoptimize Real-World Programs

   Sypula, A. G.; Yin, Pengcheng; Lacomis, Jeremy; Le Goues, Claire; Schwarts, Edward J; Neubig, Graham (January 2022, Deep Learning for Code Workshop (ICLR 2022 Workshop))

   Program optimization is the process of modifying software to execute more efficiently. Superoptimizers attempt to find the optimal program by employing significantly more expensive search and constraint solving techniques. Generally, these methods do not scale well to programs in real development scenarios, and as a result superoptimization has largely been confined to small-scale, domain-specific, and/or synthetic program benchmarks. In this paper, we propose a framework to learn to superoptimize real-world programs by using neural sequence-to-sequence models. We created a dataset consisting of over 25K real-world x86-64 assembly functions mined from open-source projects and propose an approach, Self Imitation Learning for Optimization (SILO) that is easy to implement and outperforms a standard policy gradient learning approach on our dataset. Our method, SILO, superoptimizes 5.9% of our test set when compared with the gcc version 10.3 compiler’s aggressive optimization level -O3. We also report that SILO’s rate of superoptimization on our test set is over five times that of a standard policy gradient approach and a model pre-trained on compiler optimization demonstration. 

   more » « less
5. How to get to the end of the world

   https://doi.org/10.1145/3501291  

   Liu, Jen (January 2022, Interactions)