Network analysis is becoming one of the most active research areas in statistics. Significant advances have been made recently on developing theories, methodologies and algorithms for analyzing networks. However, there has little fundamental study optimal estimation. In this paper, we establish rate convergence graphon For stochastic block model with $k$ clusters, show that under mean squared error $n^{-1}\log k+k^2/n^2$. The minimax upper bound improves existing results literature through a...

Entity alignment associates entities in different knowledge graphs if they are semantically same, and has been successfully used the graph construction connection. Most of recent solutions for entity based on embedding, which maps a low-dimension space where connected with guidance prior aligned pairs. The study this paper focuses two important issues that limit accuracy current solutions: 1) labeled data priorly pairs difficult expensive to acquire, whereas abundant unlabeled not used; 2)...

Recently vision transformer models have become prominent for a range of tasks. These models, however, usually suffer from intensive computational costs and heavy memory requirements, making them impractical deployment on edge platforms. Recent studies proposed to prune transformers in an unexplainable manner, which overlook the relationship between internal units model target class, thereby leading inferior performance. To alleviate this problem, we propose novel explainable pruning...

Recently vision transformer models have become prominent for a multitude of tasks. These models, however, are usually opaque with weak feature interpretability, making their predictions inaccessible to the users. While there has been surge interest in development post-hoc solutions that explain model decisions, these methods can not be broadly applied different architectures, as rules interpretability change accordingly based on heterogeneity data and structures. Moreover, is no method...

Cross-lingual entity alignment identifies pairs that share the same meanings but locate in different language knowledge graphs (KGs). The study this paper is to address two limitations widely exist current solutions: 1) loss functions defined at level serve well purpose of aligning labeled entities fail match whole picture and unlabeled KGs; 2) translation from one domain other has been considered (e.g., X Y by M1 or M2). However, important duality between KGs (X M2) ignored. We propose a...

Adenoid hypertrophy is one of the most common upper respiratory tract disorders during childhood, leading to a range symptoms such as nasal congestion, mouth breathing and obstructive sleep apnea. Current diagnostic methods, including computerized tomography scans endoscopy, are invasive or involve ionizing radiation, rendering them unsuitable for long-term assessments. To address these clinical challenges, this paper proposes novel deep learning approach non-invasive detection adenoid using...

Score-based diffusion models have emerged as powerful tools in generative modeling, yet their theoretical foundations remain underexplored. In this work, we focus on the Wasserstein convergence analysis of score-based models. Specifically, investigate impact various discretization schemes, including Euler discretization, exponential integrators, and midpoint randomization methods. Our provides a quantitative comparison these discrete approximations, emphasizing influence behavior....

Histone deacetylation is one of the well characterized post-translational modifications related to transcriptional repression in eukaryotes. The process histone achieved by deacetylases (HDACs). Over last decade, substantial advances our understanding mechanism fruit ripening have been achieved, but role HDACs this has not elucidated. In study, an RNA interference (RNAi) expression vector targeting SlHDA1 was constructed and transformed into tomato plants. Shorter time decreased storability...

Cross-lingual entity alignment aims at associating semantically similar entities in knowledge graphs with different languages. It has been an essential research problem for integration and graph connection, studied supervised or semi-supervised machine learning methods the assumption of clean labeled data. However, labels from human annotations often include errors, which can largely affect results. We thus aim to formulate explore robust problem, is non-trivial, due deficiency noisy labels....

Recent self-supervised learning methods are able to learn high-quality image representations and closing the gap with supervised approaches. However, these unable acquire new knowledge incrementally – they are, in fact, mostly used only as a pre-training phase over IID data. In this work we investigate continual regimes without any replay mechanism. We show that naive functional regularization, also known feature distillation, leads lower plasticity limits performance. Instead, propose...

The abundance of high-dimensional data in the modern sciences has generated tremendous interest penalized estimators such as lasso, scaled square-root elastic net, and many others. In this paper, we establish a general oracle inequality for prediction linear regression with methods. Since proof relies only on convexity continuity arguments, result holds irrespective design matrix applies to wide range estimators. Overall, bound demonstrates that generic can provide consistent any matrix....

Using numerical simulation methods in additive manufacturing (AM) processes can effectively reduce the cost of process exploration. This study presents a multi-scale based on powder bed model for grain growth laser fusion (L-PBF) Inconel 718 material. The finite volume method (FVM) and cellular automaton (CA) were used to reproduce process. results consistent with experimental terms melt pool morphology, structure, average size. competitive epitaxial grains discussed. effects power, scanning...

Path planning is a key problem in the autonomous navigation of mobile robots and research hotspot field robotics. Harris Hawk Optimization (HHO) faces challenges such as low solution accuracy slow convergence speed, it easy falls into local optimization path applications. For this reason, paper proposes Multi-strategy Improved (MIHHO) algorithm. First, double adaptive weight strategy used to enhance search capability algorithm significantly improve speed planning; second, Dimension...

In recent years rank aggregation has received significant attention from the machine learning community. The goal of such a problem is to combine (partially revealed) preferences over objects large population into single, relatively consistent ordering those objects. However, in many cases, we might not want single ranking and instead opt for individual rankings. We study version known as collaborative ranking. this assume that users provide us with pairwise (for example purchasing one item...

Yu Lu, Jiali Zeng, Jiajun Zhang, Shuangzhi Wu, Mu Li. Proceedings of the 59th Annual Meeting Association for Computational Linguistics and 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers). 2021.

Biclustering structures in data matrices were first formalized a seminal paper by John Hartigan (1972) where one seeks to cluster cases and variables simultaneously. Such are also prevalent block modeling of networks. In this paper, we develop unified theory for the estimation completion with biclustering structures, is partially observed noise contaminated matrix certain structure. particular, show that constrained least squares estimator achieves minimax rate-optimal performance several...

Recent years have witnessed increasing attention on the application of graph alignment to on-Web tasks, such as knowledge integration and social network linking. Despite achieving remarkable performance, prevailing models still suffer from noisy supervision, yet how mitigate impact noise in labeled data is under-explored. The negative sampling based discrimination model has been a feasible solution detect filter them out. However, due its sensitivity distribution, would lead an inaccurate...