Papers where the original authors have engaged with our plain-language explanation.
For every paper on this page, at least one of the original authors has seen our plain-language explanation and engaged with it — either confirming it reads accurately or requesting corrections that we then applied. An endorsement does not mean the authors formally approve every sentence, but it does mean the explanation has passed the eyes of the people who wrote the paper.
1351 papers reviewed by authors · 731–740 / 1351
This paper introduces Thalia, a global, multi-modal dataset comprising 38 spatiotemporal datacubes with expert annotations and atmospheric variables to overcome data scarcity and advance deep learning-based automated monitoring of volcanic deformation using InSAR.
The Pixel Watch 2 achieves significantly improved heart rate accuracy during motion-heavy activities by leveraging a 10-channel multipath PPG sensor and an on-device deep learning model trained on 10,000 hours of diverse data, resulting in substantially tighter error margins compared to previous devices.
The paper introduces MARLNet, a PPO-based reinforcement learning agent that integrates motion priors and action smoothness penalties to refine object detection bounding boxes, achieving consistent performance gains on VOC and VisDrone datasets while addressing reward interference and representational limitations.
This study analyzes retail markets in Uganda to demonstrate that the affordability of intoxication is driven by access to low-cost, high-strength ethanol through informal channels and plastic packaging, particularly in rural and slum areas, suggesting that alcohol policies should target ethanol content and informal market regulation rather than just beverage prices.
This paper reports the first statistically significant detection of a population of "Gamma-Loud Protostars," providing evidence that protostellar jets accelerate hadrons to produce gamma rays via pion decay, thereby establishing a new class of Galactic gamma-ray sources and revealing a critical link between particle acceleration and the mechanical power of star formation.
This paper introduces RACL, a Reasoning-Agent Control Layer that sits atop existing metaheuristics to observe, reason about, and dynamically adjust an optimizer's internal search behavior through bounded hypotheses and interventions, demonstrating significant cost improvements in vehicle routing tasks without modifying business constraints or incurring material computational overhead.
The paper introduces PCE, a model-agnostic algorithm that generates diverse, plausible, and Pareto-optimal counterfactual explanations by combining an isolation forest for data distribution adherence and SHAP values for computational efficiency, thereby outperforming existing methods in solution quality and speed.
This paper investigates functional redundancy in one-layer tanh recurrent neural networks by using ordered real Schur coordinates to identify task-specific, approximate symmetries where structured ablations of nonnormal couplings can be performed without disrupting input-output behavior, revealing that such invariances vary across tasks and solutions rather than representing universal weight-space symmetries.
Using JWST NIRSpec IFU observations from the GA-NIFS survey, this study reveals that moderate-luminosity AGN at redshifts exhibit a remarkably high incidence (>75%) of powerful, spatially resolved ionized gas outflows with velocities up to 2000 km/s, suggesting that AGN feedback was more intense and potentially more impactful on host galaxy evolution in the early Universe than at later cosmic times.
PSCT-Net is a geometry-aware deep learning framework that reconstructs 3D pediatric skull CTs from sparse bi-planar X-rays by integrating differentiable back-projection, attention-guided refinement, and a Bidirectional Mamba module to overcome depth ambiguity and improve bone boundary accuracy, while introducing a specialized pediatric dataset to address the lack of child-specific training data.