This paper proposes two rankings of statistical experiments based on the linear convex order, providing simpler and more tractable characterizations than Blackwell order, which relies on the convex order. We apply these rankings to compare statistical experiments in decision problems with a binary-action space and in decision problems that aggregate payoffs over a collection of binary-action decision problems. Furthermore, these rankings can be used to compare statistical experiments in moral hazard problems without requiring the first-order approach to be valid, thereby complementing the results in Holmström (1979) and Kim (1995). The full paper is available at: https://arxiv.org/abs/2505.14639

This paper studies learning from multiple informed agents where each agent has a small piece of information about the unknown state of the world in the form of a noisy signal and sends a message to the principal, who then makes a decision that is not constrained by predetermined rules. In contrast to the existing literature, we model the conflict of interest between the principal and the agents more generally and consider the case where the preferences of the principal and the agents are misaligned in some realized states. We show that if the conflict of interest between the principal and the agents is moderate, there is a discontinuity: when the number of agents is large enough, adding even a tiny probability of misaligned states leads to complete unraveling in which the agents ignore their signals, in contrast to the almost complete revealing that is predicted by the existing literature. Furthermore, we demonstrate that no matter how small the conflict of interest between the principal and the agents is, the information contained in each agent's message must vanish as the number of agents grows large. Finally, no matter how many agents there are, the total amount of information that is transmitted is limited, and the principal always fails to fully learn the unknown state.

Generalized nucleus segmentation techniques can contribute greatly to reducing the time to develop and validate visual biomarkers for new digital pathology datasets. We summarize the results of MoNuSeg 2018 Challenge whose objective was to develop generalizable nuclei segmentation techniques in digital pathology. The challenge was an official satellite event of the MICCAI 2018 conference in which 32 teams with more than 80 participants from geographically diverse institutes participated. Contestants were given a training set with 30 images from seven organs with annotations of 21,623 individual nuclei. A test dataset with 14 images taken from seven organs, including two organs that did not appear in the training set was released without annotations. Entries were evaluated based on average aggregated Jaccard index (AJI) on the test set to prioritize accurate instance segmentation as opposed to mere semantic segmentation. More than half the teams that completed the challenge outperformed a previous baseline. Among the trends observed that contributed to increased accuracy were the use of color normalization as well as heavy data augmentation. Additionally, fully convolutional networks inspired by variants of U-Net, FCN, and Mask-RCNN were popularly used, typically based on ResNet or VGG base architectures. Watershed segmentation on predicted semantic segmentation maps was a popular post-processing strategy. Several of the top techniques compared favorably to an individual human annotator and can be used with confidence for nuclear morphometrics.