Zero-knowledge (ZK) proof systems have developed rapidly in recent years, with hash functions as one of their central building blocks. Since these often dominate the prover cost, circuit-friendly hash function design has become an active research area. Most hash proposals target prime fields, although recent protocols such as Binius and VOLE-based ZK operate natively over binary extension fields F2n. These binary field protocols reduce the cost of proving widely used binary and bitwise statements, thereby opening up new design opportunities. At the same time, the demand for ZK applications such as zkRollups is pushing towards performant hardware acceleration, a requirement that recent designs have largely neglected. Hence, a modern ZK hash function should also be efficient in hardware and fast in plain evaluation, to avoid new bottlenecks in non-circuit workloads.In this paper, we introduce RainHash2.0, a cryptographic permutation that addresses both gaps. RainHash2.0 is natively defined over binary extension fields, making it a natural match for F2n-based protocols such as Binius and VOLEitH, while being tailored for efficient hardware and competitive plain performance. To achieve this, we exploit new techniques from Binius to horizontally split the round function – arguably a novelty in itself that is particularly effective when finite fields of different sizes are used simultaneously. We implement RainHash2.0 in the Binius and VOLE-based ZK frameworks, comparing it against SHAKE, recent arithmetization-oriented designs, and its direct predecessor RainHash. Across proof size, prover- and verifier runtime, RainHash2.0 delivers significant improvements. In addition, we prototype RainHash2.0 on FPGA hardware and reach efficiency gains of up to 8.8x over related circuitfriendly hash functions. These results mark RainHash2.0 a practical choice for modern ZK applications.
PURPOSE:This study investigates the causal mechanisms underlying radiology report generation by analyzing how clinical information and prior imaging examinations contribute to annotation shifts. We systematically estimate why and how biases manifest, providing insights into the data generation process that influences radiology reporting. METHODS:This retrospective study analyzes 172,380 chest X-ray reports from 45,561 distinct patients in the MIMIC-IV CXR database. The study focuses on conditional effects for the diseases pneumonia, pleurisy, heart failure, rib fracture, and COPD. Propensity score matching is employed to balance the treatment and control groups, followed by logistic regression and neural network models to estimate causal effects. Statistical analysis involves calculating risk differences and 95% confidence intervals to determine significance (p ≤ 0.05). Sensitivity analysis is deployed to estimate the robustness of the effect estimates. RESULTS:The inclusion of clinical questions significantly influences the reporting of key observational findings. For instance, the probability of mentioning cardiomegaly increases by 15% (p ≤ 0.05) when a clinical question is posed conditioned on rib fracture. Similar effects are observed for support devices across multiple diseases. However, the impact of clinical information varies by disease. For instance, in the presence of clinical questions, the mention of pneumonia increases significantly for one disease, while for others there is no significant effect. CONCLUSION:This study demonstrates how annotation bias in radiology reports arises from clinical context and prior imaging access. Understanding these causal mechanisms is essential for mitigating biases in dataset curation, ensuring more reliable AI models, and improving the generalizability of multimodal medical imaging systems.
CyVerse, the largest publicly-funded open-source research cyberinfrastructure for life sciences, has played a crucial role in advancing data-driven research since the 2010s. As the technology landscape evolved with the emergence of cloud computing platforms, machine learning and artificial intelligence (AI) applications, CyVerse has enabled access by providing interfaces, Software as a Service (SaaS), and cloud-native Infrastructure as Code (IaC) to leverage new technologies. CyVerse services enable researchers to integrate institutional and private computational resources, custom software, perform analyses, and publish data in accordance with open science principles. Over the past 13 years, CyVerse has registered more than 124,000 verified accounts from 160 countries and was used for over 1,600 peer-reviewed publications. Since 2011, 45,000 students and researchers have been trained to use CyVerse. The platform has been replicated and deployed in three countries outside the US, with additional private deployments on commercial clouds for US government agencies and multinational corporations. In this manuscript, we present a strategic blueprint for creating and managing SaaS cyberinfrastructure and IaC as free and open-source software.
This study aims to improve the generalisation capabilities of machine learning models for modelling hourly air pollutant concentrations in scenarios where access to high-quality data is limited. A diverse set of techniques was implemented to tackle this challenge, encompassing the utilisation of the prophet, random forest, and three different deep learning architectures: long short-term memory networks, convolutional neural networks, and multilayer perceptrons. A hybrid model of random forest and prophet was also tested. The role of the hybrid model was to combine the forecasting strengths of the Prophet model with the predictive power of the Random Forest model to better capture complex temporal patterns in the data. After testing, the hybrid model demonstrated improved generalization capabilities, achieving statistically significant improvements in R2 for hourly concentrations of NO (improving by 26
Recent work suggests that music recommender systems are prone to disproportionally frequent recommendations of music from countries more prominently represented in the training data, notably the US. However, it remains unclear to what extent feedback loops in music recommendation influence the dynamics of such imbalance. In this work, we investigate the dynamics of representation of local (i.e., country-specific) and US-produced music in user profiles and recommendations. To this end, we conduct a feedback loop simulation study using the LFM-2b dataset. The results suggest that most of the investigated recommendation models decrease the proportion of music from local artists in their recommendations. Furthermore, we find that models preserving average proportions of US and local music do not necessarily provide country-calibrated recommendations. We also look into popularity calibration and, surprisingly, find that the most popularity-calibrated model in our study (ItemKNN) provides the least country-calibrated recommendations. In addition, users from less represented countries (e.g., Finland) are, in the long term, most affected by the under-representation of their local music in recommendations.
