This content will become publicly available on July 4, 2025
Inferring the Long-Term Causal Effects of Long-Term Treatments from Short-Term Experiments
- Award ID(s):
- 1846210
- NSF-PAR ID:
- 10520740
- Publisher / Repository:
- ICML
- Date Published:
- Journal Name:
- Proceedings of Machine Learning Research
- ISSN:
- 2640-3498
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
Climate change is characterized by global surface warming associated with the increase of greenhouse gas population since the start of the industrial era. Growing evidence shows that the upper atmosphere is experiencing appreciable cooling over the last several decades. The seminal modeling study by Roble and Dickinson (1989) suggested potential effects of increased greenhouse gases on the ionosphere and thermosphere cooling which appear consistent with some observations. However, several outstanding issues remain regarding the role of CO 2 , other important contributors, and impacts of the cooling trend in the ionosphere and thermosphere: for example, (1) what is the regional variability of the trends? (2) the very strong ionospheric cooling observed by multiple incoherent scatter radars that does not fit with the prevailing theory based on the argument of anthropogenic greenhouse gas increases, why? (3) what is the effect of secular changes in Earth’s main magnetic field? Is it visible now in the ionospheric data and can it explain some of the regional variability in the observed ionospheric trends? (4) what is the impact of long-term cooling in the thermosphere on operational systems? (5) what are the appropriate strategic plans to ensure the long-term monitoring of the critical space climate?more » « less
-
IceCube is a cubic kilometer neutrino detector located at the South Pole. It generates 1 TiB of raw data per day, which must be archived for possible retrieval years or decades later. Other low-level data products are also archived for easy retrieval in the event of a catastrophic data center failure. The Long Term Archive software is IceCube's answer to archiving this data across several computing sites.more » « less
-
A standard measure of the influence of a research paper is the number of times it is cited. However, papers may be cited for many reasons, and citation count offers limited information about the extent to which a paper affected the content of subsequent publications. We therefore propose a novel method to quantify linguistic influence in timestamped document collections. There are two main steps: first, identify lexical and semantic changes using contextual embeddings and word frequencies; second, aggregate information about these changes into per-document influence scores by estimating a high-dimensional Hawkes process with a low-rank parameter matrix. We show that this measure of linguistic influence is predictive of future citations: the estimate of linguistic influence from the two years after a paper’s publication is correlated with and predictive of its citation count in the following three years. This is demonstrated using an online evaluation with incremental temporal training/test splits, in comparison with a strong baseline that includes predictors for initial citation counts, topics, and lexical features.more » « less
