Precise user and item embedding learning is the key to building a successful recommender system. Traditionally, Collaborative Filtering (CF) provides way learn embeddings from user-item interaction history. However, performance limited due sparseness of behavior data. With emergence online social networks, systems have been proposed utilize each user's local neighbors' preferences alleviate data sparsity for better modeling. We argue that, platform, her potential influenced by trusted users,...

Abstract Ferrous sulfides with the high theoretic capacity are promising anode for sodium ion batteries. However, fading and inferior rate capability still hinder their practical application. In this work, Na‐doped Fe 7 S 8 microrods cationic vacancies weakened Fe─S bond constructed through a facile scalable sulfurized route. The experimental results combined theoretical analysis thoroughly reveal generation of strength induced by doping, which modulates energy band structure , provides more...

Tree boosting, which combines weak learners (typically decision trees) to generate a strong learner, is highly effective and widely used machine learning method. However, the development of high performance tree boosting model time-consuming process that requires numerous trial-and-error experiments. To tackle this issue, we have developed visual diagnosis tool, BOOSTVis, help experts quickly analyze diagnose training boosting. In particular, designed temporal confusion matrix visualization,...

As a key application of artificial intelligence, recommender systems are among the most pervasive computer aided to help users find potential items interests. Recently, researchers paid considerable attention fairness issues for intelligence applications. Most these approaches assumed independence instances, and designed sophisticated models eliminate sensitive information facilitate fairness. However, differ greatly from as naturally form user-item bipartite graph, collaboratively...

News recommendation is important for online news services. Existing models are usually learned from users' click behaviors. Usually the behaviors of users with same sensitive attributes (e.g., genders) have similar patterns and can easily capture these patterns. It may lead to some biases related user in results, e.g., always recommending sports male users, which unfair since not receive diverse information. In this paper, we propose a fairness-aware approach decomposed adversarial learning...

Knowledge graphs (KGs) have been widely used to improve recommendation accuracy. The multi-hop paths on KGs also enable reasoning, which is considered a crystal type of explainability. In this paper, we propose reinforcement learning framework for multi-level reasoning over KGs, leverages both ontology-view and instance-view model user interests. This ensures convergence more satisfying solution by effectively transferring high-level knowledge lower levels. Based the framework, path...

The detection of fake news often requires sophisticated reasoning skills, such as logically combining information by considering word-level subtle clues. In this paper, we move towards fine-grained for better reflecting the logical processes human thinking and enabling modeling particular, propose a framework following human’s information-processing model, introduce mutual-reinforcement-based method incorporating knowledge about which evidence is more important, design prior-aware bi-channel...

The wide spread of fake news has caused serious societal issues. We propose a subgraph reasoning paradigm for detection, which provides crystal type explainability by revealing subgraphs the propagation network are most important verification, and concurrently improves generalization discrimination power graph-based detection models removing task-irrelevant information. In particular, we reinforced generation method, perform fine-grained modeling on generated developing Hierarchical...

Collaborative Filtering (CF) is one of the most successful approaches for recommender systems. With emergence online social networks, recommendation has become a popular research direction. Most these models utilized each user's local neighbors' preferences to alleviate data sparsity issue in CF. However, they only considered neighbors user and neglected process that users' are influenced as information diffuses network. Recently, Graph Convolutional Networks~(GCN) have shown promising...

News recommendation systems play a key role in online news reading service. Knowledge graphs (KG), which contain comprehensive structural knowledge, are well known for their potential to enhance both accuracy and explainability. While existing works intensively study using KG improve accuracy, reasoning has not been fully explored. A few such as KPRN [18], [22] ADAC [25] have discussed knowledge some other domains music or movie, but methods practical the news. How make scalable generic KGs,...

Existing review-based recommendation models mainly learn long- term user and item representations from a set of reviews. Due to the ignorance rich side information reviews, these suffer two drawbacks: 1) they fail capture short-term changes preferences features reflected in reviews 2) cannot accurately model high-order user-item collaborative signals To overcome limitations, we propose multi-view approach named Set-Sequence-Graph (SSG), augment existing single-view (i.e., view set) methods...

The vertical profiles of aerosol optical properties are vital to clarify their transboundary transport, climate forcing and environmental health influences. Based on synergistic measurements multiple advanced detection techniques, this study investigated structure characteristics during two dust haze events in Lanzhou northwest China. Dust particles originated from remote deserts traveled eastward at different altitudes reached 10 April 2020. trans-regional aloft (~4.0 km) were entrained...

Abstract Based on CE318‐T sun‐sky‐lunar photometric measurements, this study examined the diurnal and nocturnal variations of dust aerosol characteristics from June to October 2023 at Ayvaj southwest Tajikistan for first time. The results indicated that ROLO‐RCF (RObotic Lunar Observatory model with correction factor)‐modified method was capable reliably calculating optical depths (AODs) Ångström exponent (AE 440–870 nm ) measured photometer, which validated synchronous observations a lidar....

Large Language Models (LLMs) gain substantial reasoning and decision-making capabilities from thought structures. However, existing methods such as Tree of Thought Retrieval Augmented Thoughts often fall short in complex tasks due to the limitations insufficient local retrieval factual knowledge inadequate global selection strategies. These make it challenging for these balance accuracy comprehensive logical optimization effectively. To address limitations, we introduce (RATT), a novel...

Chronic insomnia is a disease which brings intense mental pain and disturbing complications to patients worldwide. The oral microbiome exhibits mechanistic influence on human health. Therefore, it crucial understand the microbial diversity in insomnia. Tongue diagnosis has been considered critical basic procedure therapeutic decision-making Traditional Chinese Medicine (TCM). Hence, significant elucidate various differences chronic with different tongue features. In this paper, we used 16S...

News recommendation aims to help online news platform users find their preferred articles. Existing methods usually learn models from historical user behaviors on news. However, these are biased providers. Models trained data may capture and even amplify the biases providers, unfair for some minority In this paper, we propose a provider fairness-aware framework (named ProFairRec), which can fair different providers data. The core idea of ProFairRec is provider-fair representations achieve...

Background Cardioembolic Stroke (CS) and Atrial Fibrillation (AF) are prevalent diseases that significantly impact the quality of life impose considerable financial burdens on society. Despite increasing evidence a significant association between two diseases, their complex interactions remain inadequately understood. We conducted bioinformatics analysis employed machine learning techniques to investigate potential shared biomarkers CS AF. Methods retrieved AF datasets from Gene Expression...

Oral bioavailability (OBA)-related pharmacokinetic properties, such as aqueous solubility, lipophilicity, and intestinal membrane permeability, play a significant role in drug discovery. However, their measurement is usually costly time-consuming. Therefore, prediction models based on diverse approaches have been established recent decades. Computational of molecular properties has become an important step discovery, aiming to identify potential drug-like candidates reduce costs. limitations...

Although several effective learning-from-crowd methods have been developed to infer correct labels from noisy crowdsourced labels, a method for post-processed expert validation is still needed. This paper introduces semi-supervised learning algorithm that capable of selecting the most informative instances and maximizing influence labels. Specifically, we complete uncertainty assessment facilitate selection instances. The are then propagated similar via regularized Bayesian inference....

Abstract A Cimel Sun‐sky‐lunar photometer (CE318‐T) is designed to perform daytime and nighttime photometric measurements calculate diurnal cycle of aerosol optical depth (AOD). Nevertheless, the determination nocturnal AOD from CE318‐T requires a precise knowledge extraterrestrial lunar irradiance, which significantly changes with moon's phase angle (MPA) libration in single night. This study evaluated 1‐year AODs at Lanzhou by using three different methods, were validated collocated DIAL...