Assessing Israel's Motives In Annexing the Jordan Valley

This article analyzes Israel's motives in annexing the Jordan Valley—a plan that, if approved, will eliminate any possibility of establishing a Palestinian state, even on a small part of historic Palestine. This promises to be one of the most critical strategic turning points in the contemporary Israeli‐Palestinian conflict. The analysis reveals that the Israeli annexation decision, even if postponed, has become a reality imposed by Israel on the international community, insinuated into formal and official government announcements and declarations. In addition, the article highlights the danger of imposing Israeli sovereignty over the Jordan Valley since it carries with it a threat to regional stability. As part of this examination, the study traces the development of the positions of successive Israeli governments toward the issue of annexing this region, from 1967 through the dissolution of the Netanyahu government in 2021.

Does personalization of officeholders undermine the legitimacy of the office? On perceptions of objectivity in legal decisionmaking

Abstract Public legitimation of legal decisionmaking can be promoted through various strategies. We examine strategies of legitimation that are premised on personalizing the public image of legal agents. A personalized public administration emphasizes individual decisionmakers and seeks legitimacy through familiarity with the character, identity, and virtues of individual agents, whereas a non‐personalized public administration projects an ethos of technocratic decisionmaking, seeking legitimacy through institutional objectivity and impartiality. We conducted an experiment to examine the efficacy of personalization strategies in the context of a politically charged legal affair: the criminal cases involving the prime minister of Israel, Benjamin Netanyahu. We focus on people's perceived objectivity of the office of the Israeli attorney general (AG), given exposure (vs. no exposure) to different types of personal information about the AG, and while manipulating the salience of contrasting decisions concerning Netanyahu (indicting him on several counts of corruption versus exculpating him in others). We find that exposure to personal information about the AG decreased the perceived objectivity of his office, compared to no exposure to personal information, regardless of the type of information, decision salience, and respondents' political leanings. Our findings, therefore, support the legitimating potential of the non‐personalization of decisionmakers, and show that it pertains to people positioned as both “losers” and “winners” with regard the political impact of the decision. The study further reflects the capacity of nonabstract real‐world, real‐time, analyses to shed light on the drivers of public trust in legal decisionmaking in politically polarized contexts—an issue of pertinence in many contemporary democracies.

The Ultimate Revenge of Benjamin Netanyahu

Media malaise or mobilization during repeat elections? Evidence from Israel’s three consecutive rounds of elections (2019–2020)

In 2019–2020, Israel went through three consecutive elections in less than a year on grounds of alleged corruption by Prime Minister Netanyahu, and his lack of ability to form a coalition. This study aims to contribute to analyses of the media mobilization/malaise effect by examining the impact of such a prolonged period of campaigning on citizens’ political behavior. Thus, we conducted six online surveys using a longitudinal sample of Israeli society before and after each election. The analysis found that, despite participants’ testimonies that they were increasingly “tired of dealing with elections,” there was a significant increase in participants’ reported certainty in their vote, news consumption, participation in online political discussions, and level of political efficacy between the elections. Next, a multivariate analysis aiming to explain variations in voters’ political efficacy found that political trust, participants’ reported certainty about their vote, and political interest all explained high levels of political efficacy. The analysis provides one of the strongest reinforcements to date regarding the validity of political mobilization theory, demonstrating its relevance under challenging conditions. We discuss further implications and generalizability of our findings.

Learning to Extrapolate: A Transductive Approach

PathRTM: Real-time prediction of KI-67 and tumor-infiltrated lymphocytes

Unsupervised Iterative U-Net with an Internal Guidance Layer for Vertebrae Contrast Enhancement in Chest X-Ray Images

Diagnosis, Feedback, Adaptation: A Human-in-the-Loop Framework for Test-Time Policy Adaptation

Learning to Extrapolate: A Transductive Approach

Machine learning systems, especially with overparameterized deep neuralnetworks, can generalize to novel test instances drawn from the samedistribution as the training data. However, they fare poorly when evaluated onout-of-support test points. In this work, we tackle the problem of developingmachine learning systems that retain the power of overparameterized functionapproximators while enabling extrapolation to out-of-support test points whenpossible. This is accomplished by noting that under certain conditions, a"transductive" reparameterization can convert an out-of-support extrapolationproblem into a problem of within-support combinatorial generalization. Wepropose a simple strategy based on bilinear embeddings to enable this type ofcombinatorial generalization, thereby addressing the out-of-supportextrapolation problem under certain conditions. We instantiate a simple,practical algorithm applicable to various supervised learning and imitationlearning tasks.

PathRTM: Real-time prediction of KI-67 and tumor-infiltrated lymphocytes

In this paper, we introduce PathRTM, a novel deep neural network detectorbased on RTMDet, for automated KI-67 proliferation and tumor-infiltratedlymphocyte estimation. KI-67 proliferation and tumor-infiltrated lymphocyteestimation play a crucial role in cancer diagnosis and treatment. PathRTM is anextension of the PathoNet work, which uses single pixel keypoints for withineach cell. We demonstrate that PathRTM, with higher-level supervision in theform of bounding box labels generated automatically from the keypoints usingNuClick, can significantly improve KI-67 proliferation and tumorinfiltratedlymphocyte estimation. Experiments on our custom dataset show that PathRTMachieves state-of-the-art performance in KI-67 immunopositive, immunonegative,and lymphocyte detection, with an average precision (AP) of 41.3%. Our resultssuggest that PathRTM is a promising approach for accurate KI-67 proliferationand tumor-infiltrated lymphocyte estimation, offering annotation efficiency,accurate predictive capabilities, and improved runtime. The method also enablesestimation of cell sizes of interest, which was previously unavailable, throughthe bounding box predictions.