Applications

Cai Zhou, Zekai Wang, Menghua Wu et al.

This work addresses inefficiencies in LLM deployment for reasoning tasks, offering a method to reduce compute costs while maintaining performance, though it is incremental as it builds on conformal prediction and test-time training.

Yuanhao Chen, Peter Chin

For computational linguists and cognitive scientists, this demonstrates that LLMs learn formal-syntactic abstractions invisible to UD-based probes, revealing that UD provides a lower bound on syntactic encoding.

Hayden Helm, Carey Priebe

For researchers and practitioners using LLM-based surveys for market research, this work addresses the critical issue of prompt sensitivity invalidating standard statistical tests, offering a valid alternative.

Siqi Li, Chuan Hong, Ziye Tian et al.

For healthcare systems deploying models across settings with incomplete data, DRUM provides a robust transfer learning method without requiring imputation or target labels.

Peiyao Cai, Chengyu Cui, Felipe Maia Polo et al.

This work addresses the need for flexible scaling laws that account for heterogeneity across LLM families and benchmarks, offering a more nuanced tool for model development and evaluation.

Michael Hardy, Joshua Gilbert, Benjamin Domingue

This method addresses the challenge of ensuring assessment validity for AI researchers and educators by providing a scalable, model-agnostic tool to reduce reviewer effort in vetting large item sets.

Yichun Peng, Saike He, Peijie Zhang et al.

For researchers and policymakers in science of science, this provides a new, actionable principle for organizing teams to foster disruptive innovation, though it is an incremental extension of network-based team metrics.

Yao Zhao, Zhiyue Zhang, Yanxun Xu

This addresses the need for scalable and reproducible evidence synthesis in precision medicine, representing a novel integration rather than an incremental improvement.

Tian Bai, Ying Jin

This addresses the need for reliable uncertainty quantification in high-stakes applications like drug discovery and health prediction, offering a novel method for risk control.

Luxu Liang, Xiang Li

For practitioners needing robust detection of LLM-generated text, S2D offers a method that improves discriminative power over raw hidden representations, though it is an incremental improvement over existing representation-based detectors.

Sacha Braun, Michael I. Jordan, Francis Bach

For practitioners needing reliable prediction regions in multivariate regression, SLS offers a direct optimization alternative to two-step density-based methods, which are often difficult and computationally expensive.

Zachary R. Fox, Janet O. Agbaje, Dakotah Maguire et al.

For public health researchers and policymakers, this model enables timely and accurate air quality assessments, addressing the bottleneck of near real-time pollution exposure estimation.

Chao Li, Xing Su, Chao Fan et al.

For policymakers, this reveals that heat mitigation and adaptation policies have opposing effects on internal migration, which is crucial for understanding population dynamics under climate change.

Yi Yu, Junzhuo Ma, Chenghuang Shen et al.

This work provides a practical solution for deploying efficient technical service agents, though it appears incremental as it builds on existing adaptation methods.

Hyunwoo Kim, Munyoung Lee, Seung Hyub Jeon et al.

This addresses wafer-level spatial monitoring for plasma etching processes, which is incremental as it extends LLM reprogramming to spatial estimation.

Jose Benitez-Aurioles, Ricardo Silva, Brian McMillan et al.

This method corrects label bias due to differential testing frequency in clinical prediction models, improving fairness and accuracy for underdiagnosed groups.

Hao Wang, Licheng Pan, Qingsong Wen et al.

It offers a systematic overview for researchers in time-series forecasting, but is incremental as it synthesizes existing literature.

Ravish Gupta, Saket Kumar

This study addresses the problem of occupational displacement due to agentic AI for policymakers and regional planners, providing a novel framework to assess risks and opportunities in information-intensive sectors.

Keyi Wang, Jiaxiang Ji, Mahan Mansouri et al.

For electricity market operators and participants, the method offers an interpretable, low-complexity forecasting approach that outperforms existing methods on real-world data.

David Snyder, Apurva Badithela, Nikolai Matni et al.

This work addresses the need for reliable and efficient policy comparison in robotics, particularly for generalist manipulation policies, by providing a unified approach that handles various metrics beyond binary success, though it is incremental as it builds on existing sequential inference methods.