Yang Yu
YOU?
Author Swipe
View article: VELVET-Med: Vision and Efficient Language Pre-training for Volumetric Imaging Tasks in Medicine
VELVET-Med: Vision and Efficient Language Pre-training for Volumetric Imaging Tasks in Medicine Open
Vision-and-language models (VLMs) have been increasingly explored in the medical domain, particularly following the success of CLIP in general domain. However, unlike the relatively straightforward pairing of 2D images and text, curating l…
View article: Self-Preservation in the Digital Age: An Exploration of the Positive Utilization of News Avoidance
Self-Preservation in the Digital Age: An Exploration of the Positive Utilization of News Avoidance Open
In recent years, the rapid development of the Internet has made it more convenient for the public to access all kinds of information. However, the number of people avoiding news is gradually increasing. The phenomenon of news avoidance is …
View article: Characterization and association patterns of <scp>Chinese</scp> food–medicine homologous species based on big data analytics
Characterization and association patterns of <span>Chinese</span> food–medicine homologous species based on big data analytics Open
BACKGROUND This study employs big data analytics to explore the characteristics and association patterns of 102 Chinese food–medicine homologous (CFMH) species recognized by the National Health Commission of China, focusing on their medici…
View article: Fault Diagnosis of Wind Turbine Blades Based on One-Dimensional Convolutional Neural Network-Bidirectional Long Short-Term Memory-Adaptive Boosting and Multi-Source Data Fusion
Fault Diagnosis of Wind Turbine Blades Based on One-Dimensional Convolutional Neural Network-Bidirectional Long Short-Term Memory-Adaptive Boosting and Multi-Source Data Fusion Open
To prevent wind turbine blade accidents and improve fault detection accuracy, a hybrid deep learning model based on 1D CNN-BiLSTM-AdaBoost for wind turbine-blade fault classification is proposed. Fault data are first preprocessed by segmen…
View article: Behavior-Regularized Diffusion Policy Optimization for Offline Reinforcement Learning
Behavior-Regularized Diffusion Policy Optimization for Offline Reinforcement Learning Open
Behavior regularization, which constrains the policy to stay close to some behavior policy, is widely used in offline reinforcement learning (RL) to manage the risk of hazardous exploitation of unseen actions. Nevertheless, existing litera…
View article: Near-Optimal Online Learning for Multi-Agent Submodular Coordination: Tight Approximation and Communication Efficiency
Near-Optimal Online Learning for Multi-Agent Submodular Coordination: Tight Approximation and Communication Efficiency Open
Coordinating multiple agents to collaboratively maximize submodular functions in unpredictable environments is a critical task with numerous applications in machine learning, robot planning and control. The existing approaches, such as the…
View article: Development of a reduced combustion mechanism for 2-butyltetrahydrofuran and its blends with diesel
Development of a reduced combustion mechanism for 2-butyltetrahydrofuran and its blends with diesel Open
The 2-butyltetrahydrofuran (2-BTHF) is being investigated as a potential alternative fuel for Diesel engines due to its self-ignition properties. This paper presents a reduced 2-BTHF chemical kinetic mechanism for the first time, accompani…
View article: The Association Between Work Passion and Research Ability in Male Nurses: A Multicenter Cross‐Sectional Study in China
The Association Between Work Passion and Research Ability in Male Nurses: A Multicenter Cross‐Sectional Study in China Open
Aim: To investigate the relationship between work passion and research ability in male nurses from China. Background: Male nurses are vital to global nursing, yet face severe shortages, occupational stress, and research constraints. While …
View article: A mixture transmuted generalized extreme value distribution: Definition and properties
A mixture transmuted generalized extreme value distribution: Definition and properties Open
Extreme events are often described using generalized extreme value models, which are crucial for quantifying their impact. In prior studies, researchers have utilized the quadratic rank transmutation map to construct a comprehensive family…
View article: Cigarette Quality Problem Prediction Based on WOA‐BiLSTM‐SPA
Cigarette Quality Problem Prediction Based on WOA‐BiLSTM‐SPA Open
In modern cigarette production, proactive quality prediction is essential for ensuring product consistency and manufacturing efficiency, yet traditional post‐occurrence defect detection methods often lead to delays, inefficiencies, and hig…
View article: Influence of Elevated Potassium Fertilization on Structural and Functional Properties of Sweet Potato Root Tuber Starch
Influence of Elevated Potassium Fertilization on Structural and Functional Properties of Sweet Potato Root Tuber Starch Open
Nine sweet potato varieties with different flesh colors were cultivated under uniform environmental conditions with potassium (K) fertilizer treatments at levels of 0, 22.5, and 45 kg/ha. The structural and functional properties of the sta…
View article: Provably and Practically Efficient Adversarial Imitation Learning with General Function Approximation
Provably and Practically Efficient Adversarial Imitation Learning with General Function Approximation Open
As a prominent category of imitation learning methods, adversarial imitation learning (AIL) has garnered significant practical success powered by neural network approximation. However, existing theoretical studies on AIL are primarily limi…
View article: Graph Neural Patching for Cold-Start Recommendations
Graph Neural Patching for Cold-Start Recommendations Open
The cold start problem in recommender systems remains a critical challenge. Current solutions often train hybrid models on auxiliary data for both cold and warm users/items, potentially degrading the experience for the latter. This drawbac…
View article: Hindsight Preference Learning for Offline Preference-based Reinforcement Learning
Hindsight Preference Learning for Offline Preference-based Reinforcement Learning Open
Offline preference-based reinforcement learning (RL), which focuses on optimizing policies using human preferences between pairs of trajectory segments selected from an offline dataset, has emerged as a practical avenue for RL applications…
View article: CIS: Composable Instruction Set for Data Streaming Applications
CIS: Composable Instruction Set for Data Streaming Applications Open
The enhanced efficiency of hardware accelerators, including Single Instruction Multiple Data (SIMD) architectures and Coarse-Grained Reconfigurable Architectures (CGRAs), is driving significant advancements in Artificial Intelligence and M…
View article: BWArea Model: Learning World Model, Inverse Dynamics, and Policy for Controllable Language Generation
BWArea Model: Learning World Model, Inverse Dynamics, and Policy for Controllable Language Generation Open
Large language models (LLMs) have catalyzed a paradigm shift in natural language processing, yet their limited controllability poses a significant challenge for downstream applications. We aim to address this by drawing inspiration from th…
View article: Generalizable Task Representation Learning for Offline Meta-Reinforcement Learning with Data Limitations
Generalizable Task Representation Learning for Offline Meta-Reinforcement Learning with Data Limitations Open
Generalization and sample efficiency have been long-standing issues concerning reinforcement learning, and thus the field of Offline Meta-Reinforcement Learning (OMRL) has gained increasing attention due to its potential of solving a wide …
View article: Focus-Then-Decide: Segmentation-Assisted Reinforcement Learning
Focus-Then-Decide: Segmentation-Assisted Reinforcement Learning Open
Visual Reinforcement Learning (RL) is a promising approach to achieve human-like intelligence. However, it currently faces challenges in learning efficiently within noisy environments. In contrast, humans can quickly identify task-relevant…
View article: Episodic Return Decomposition by Difference of Implicitly Assigned Sub-Trajectory Reward
Episodic Return Decomposition by Difference of Implicitly Assigned Sub-Trajectory Reward Open
Real-world decision-making problems are usually accompanied by delayed rewards, which affects the sample efficiency of Reinforcement Learning, especially in the extremely delayed case where the only feedback is the episodic reward obtained…
View article: Communication-robust multi-agent learning by adaptable auxiliary multi-agent adversary generation
Communication-robust multi-agent learning by adaptable auxiliary multi-agent adversary generation Open
Communication can promote coordination in cooperative Multi-Agent Reinforcement Learning (MARL). Nowadays, existing works mainly focus on improving the communication efficiency of agents, neglecting that real-world communication is much mo…
View article: Model-Based Reinforcement Learning with Multi-Step Plan Value Estimation
Model-Based Reinforcement Learning with Multi-Step Plan Value Estimation Open
A promising way to improve the sample efficiency of reinforcement learning is model-based methods, in which many explorations and evaluations can happen in the learned models to save real-world samples. However, when the learned model has …
View article: Powder Preparation by Vacuum Atomization and 3D Printing Forming Performance of 316 Stainless Steel
Powder Preparation by Vacuum Atomization and 3D Printing Forming Performance of 316 Stainless Steel Open
This work describes the process of preparing 316 L stainless steel powder by vacuum atomization, then a 3D print sample of 316 L stainless steel is prepared by selective laser melting (SLM) experiments. Over 77% of the powder particle size…
View article: Model-Based Offline Weighted Policy Optimization (Student Abstract)
Model-Based Offline Weighted Policy Optimization (Student Abstract) Open
A promising direction for applying reinforcement learning to the real world is learning from offline datasets. Offline reinforcement learning aims to learn policies from pre-collected datasets without online interaction with the environmen…
View article: Deep Anomaly Detection and Search via Reinforcement Learning (Student Abstract)
Deep Anomaly Detection and Search via Reinforcement Learning (Student Abstract) Open
Semi-supervised anomaly detection is a data mining task which aims at learning features from partially-labeled datasets. We propose Deep Anomaly Detection and Search (DADS) with reinforcement learning. During the training process, the agen…
View article: Anti-drifting Feature Selection via Deep Reinforcement Learning (Student Abstract)
Anti-drifting Feature Selection via Deep Reinforcement Learning (Student Abstract) Open
Feature selection (FS) is a crucial procedure in machine learning pipelines for its significant benefits in removing data redundancy and mitigating model overfitting. Since concept drift is a widespread phenomenon in streaming data and cou…
View article: Learning Generalizable Batch Active Learning Strategies via Deep Q-networks (Student Abstract)
Learning Generalizable Batch Active Learning Strategies via Deep Q-networks (Student Abstract) Open
To handle a large amount of unlabeled data, batch active learning (BAL) queries humans for the labels of a batch of the most valuable data points at every round. Most current BAL strategies are based on human-designed heuristics, such as u…