Pair-then-Aggregate (PtA) introduces a programming paradigm and an automated parallel execution engine for large-scale secure multi-party (MPC) computations, drawing inspiration from the widely-used yet not explicitly defined Table-Generation-and-Look-up (TGL) pattern in privacy-preserving algorithm design. PtA offers an easy-to-use API and a versatile execution engine that harnesses various levels of parallelism and adapts to different MPC deployments, algorithms, and input sizes. Evaluations on a real-world MPC platform demonstrate significant enhancements in scalability, adaptability, and ease of programming. PtA can process one billion input elements with 3-23 lines of C++ code in 5-74 seconds. It outperforms state-of-the-art implementations in 91.4 % of 35 test cases, achieving up to a$12.4 \times$speedup with much less coding effort.11Our code is provided in https://github.com/Fannxy/Pair-then-Aggregate

Artificial intelligence (AI) has had a profound impact on the medical industry. As the ultimate consumers, patients’ acceptance of medical AI plays an important role in realizing its widespread adoption in healthcare. However, previous studies have indicated that patients’ acceptance of medical AI is still inconsistent. This study considered the application of medical AI in different scenarios and divided them into medical AI for diagnosis and treatment (MAI-DT) and medical AI for non-diagnosis and non-treatment (MAI-NDT). Drawing on the theories of uniqueness and dehumanization, this study explored the impacts of patients’ sense of uniqueness (SOU) on their acceptance of medical AI under the two scenarios and the underlying mechanisms. Analysis of the data from 238 samples collected during the COVID-19 period in China showed that participants with a higher SOU had a lower acceptance of MAI-DT but a higher acceptance of MAI-NDT than participants with a lower SOU. In addition, perceived AI knowledge and perceived AI empathy mediated the relationship between the participants’ SOU and their acceptance of AI under the two scenarios. The theoretical and practical implications are discussed.

We study the maximum weight perfect $f$-factor problem on any general simple graph $G=(V,E,w)$ with positive integral edge weights $w$, and $n=|V|$, $m=|E|$. When we have a function $f:V\rightarrow \mathbb{N}_+$ on vertices, a perfect $f$-factor is a generalized matching so that every vertex $u$ is matched to $f(u)$ different edges. The previous best algorithms on this problem have running time $O(m f(V))$ [Gabow 2018] or $\tilde{O}(W(f(V))^{2.373}))$ [Gabow and Sankowski 2013], where $W$ is the maximum edge weight, and $f(V)=\sum_{u\in V}f(u)$. In this paper, we present a scaling algorithm for this problem with running time $\tilde{O}(mn^{2/3}\log W)$. Previously this bound is only known for bipartite graphs [Gabow and Tarjan 1989]. The running time of our algorithm is independent of $f(V)$, and consequently it first breaks the $Ω(mn)$ barrier for large $f(V)$ even for the unweighted $f$-factor problem in general graphs.

Given an undirected simple graph G = (V, E) that undergoes edge insertions and deletions, we wish to efficiently maintain an edge coloring with only a few colors. The previous best dynamic algorithm by [3] could deterministically maintain a valid edge coloring using 2Δ – 1 colors with O(log Δ) update time, where Δ stands for the current maximum vertex degree of graph G. In this paper, we first propose a new static (1 + ∊)Δ edge coloring algorithm that runs in near-linear time. Based on this static algorithm, we show that there is a randomized dynamic algorithm for this problem that only uses (1 + ∊)Δ colors with O(log8 n/∊4) amortized update time when Δ ≥ Ω(log2 n/∊2), where ∊ > 0 is an arbitrarily small constant.