A5100691427- Topic Modeling
- Natural Language Processing Techniques
- Robot Manipulation and Learning
- COVID-19 diagnosis using AI
- Radiomics and Machine Learning in Medical Imaging
- AI in cancer detection
- Polymer composites and self-healing
- Asphalt Pavement Performance Evaluation
- Explainable Artificial Intelligence (XAI)
- Text Readability and Simplification
- Robotics and Sensor-Based Localization
- Multimodal Machine Learning Applications
- Educational Technology and Assessment
- Medical Image Segmentation Techniques
- Grouting, Rheology, and Soil Mechanics
- Teleoperation and Haptic Systems
- Anomaly Detection Techniques and Applications
- Surgical Simulation and Training
- Advanced Neural Network Applications
- Online Learning and Analytics
- Adversarial Robustness in Machine Learning
- Polydiacetylene-based materials and applications
- Gaze Tracking and Assistive Technology
- Fluid Dynamics and Thin Films
- Neural Networks Stability and Synchronization
Northwestern University
2022-2025
Intel (United States)
2025
Institute of Rock and Soil Mechanics
2023-2024
University of Chinese Academy of Sciences
2023-2024
Sichuan University
2024
Yibin University
2024
Johns Hopkins University
2023-2024
Zhengzhou University
2024
Changchun Institute of Optics, Fine Mechanics and Physics
2024
Imperial College London
2024
Telekinesis, as commonly portrayed in science fiction literature and cinema, is a super power wherein users control manipulate objects absent physical interaction. In real world, enhancing human-robot interaction needs the synthesis of human intuitive processes with robotic arm. This paper introduces teleoperation system achieving essence telekinetic operations, combining profound capabilities augmented reality (AR) arm operations. Utilizing AR, proposed methodology offers operators visual...
The unprecedented performance of large language models (LLMs) necessitates improvements in evaluations. Rather than merely exploring the breadth LLM abilities, we believe meticulous and thoughtful designs are essential to thorough, unbiased, applicable Given importance world knowledge LLMs, construct a Knowledge-oriented Assessment benchmark (KoLA), which carefully design three crucial factors: (1) For ability modeling, mimic human cognition form four-level taxonomy knowledge-related...
Welding stands as a critical focus for the intelligent and digital transformation of machinery industry, with automated laser welding playing pivotal role in sector’s technological advancement. The management deformation such operations is fundamental, relying on advanced analysis prediction methods. endeavor to accurately analyze practical applications compounded by interplay numerous variables, pronounced coupling effect among these factors, reliance expert intuition. Thus, effective...
This study investigates the potential of eye-tracking technology and Segment Anything Model (SAM) to design a collaborative human-computer interaction system that automates medical image segmentation. We present \textbf{GazeSAM} enable radiologists collect segmentation masks by simply looking at region interest during diagnosis. The proposed tracks radiologists' eye movement utilizes eye-gaze data as input prompt for SAM, which automatically generates mask in real time. is first work...
The key to optimizing the performance of an anycast-based system (e.g., root DNS or a CDN) is choosing right set sites announce anycast prefix. One challenge here predicting catchments. A naïve approach advertise prefix from all subsets available and choose best-performing subset, but this does not scale well. We demonstrate that by conducting pairwise experiments between peering with tier-1 networks, we can predict catchments would result if any subset sites. prove our method effective in...
Motion errors in the trajectory of a six-joint industrial robotic arm’s end-effector can significantly impact machining precision. Complex milling operations lead to deviations from intended path due structural characteristics. These often exhibit periodic and position-dependent variations, underscoring need for meticulous control measures. To address this challenge, we propose novel motion decomposition-based error compensation technique arm. This approach involves breaking down robot’s...
The distributed control architecture is expected to dominate the gas turbine engine (GTE) systems in future, which sensors and actuators are connected controllers via a network. Hence, problem of network-enabled high-performance (DEC) becomes important modern systems. Due properties network, packet dropouts must be considered. This study introduces system based on networked cascade (NCCS). Typical turboshaft engine-distributed designed NCCS framework with <inline-formula...
In this paper, the water droplet erosion (WDE) performance of typical martensitic precipitation substrate 0Cr17Ni4Cu4Nb in steam turbine final stage, laser solid solution strengthened sample, cladding sample and brazed stellite alloy samples have been studied based on a high-speed rotating waterjet test system. The WDE resistance several materials from strong to weak is sequence: Brazed > substrate. Furthermore, mechanism failure mode revealed. It found that hard carbide starting point crack...
In recent years, mechanoluminescent (ML) materials have shown great potential in stress sensing, mechanical energy collection and conversion, so they attracted wide attention the field of stomatology. early stage this study, BaSi2O2N2:Eu2+ ML phosphors were synthesized by two-step high temperature solid state method, then mixed with Polydimethylsiloxane (PDMS) different proportions to obtain BaSi2O2N2:Eu2+/PDMS composites mass fractions (10%,20%,30%,40%,50%). Then its biosafety was evaluated...
Large-scale, big-variant, and high-quality data are crucial for developing robust successful deep-learning models medical applications since they potentially enable better generalization performance avoid overfitting. However, the scarcity of labeled always presents significant challenges. This paper proposes a novel approach to address this challenge by controllable diffusion image synthesis, called EMIT-Diff. We leverage recent probabilistic generate realistic diverse synthetic that...
Adaptive learning aims to stimulate and meet the needs of individual learners, which requires sophisticated system-level coordination diverse tasks, including modeling resources, estimating student states, making personalized recommendations. Existing deep methods have achieved great success over statistical models; however, they still lack generalization for tasks suffer from insufficient capacity since are composed highly-coupled task-specific architectures rely on small-scale,...
Textile defect detection is an important part of textile quality control. Due to the diversity fabric texture and lack images, methods based on deep learning which does not rely defective samples has been gradually applied. However, in previous methods, ability distinguish image features defects insufficient. In order solve this problem, paper proposed improved generative adversarial network, introduced a self-encoder with MLP layers into generator module. Fabric images will be reconstructed...
In this paper we discuss the error estimations for div least-squares finite element method on elliptic problems. Compared with previous work, present a complete analysis, which improves current \emph{state-of-the-art} results. The both scalar and flux variables are established by dual arguments, in most cases, only an $H^{1+\varepsilon}$ regularity is used. Numerical experiments strongly confirm our analysis.
When making mode decisions of inter frames in H.264 [1], the encoder must perform exhaustive calculations costs Rate Distortion (RD) on all modes and find optimal one. This procedure has high computational complexity for many redundant calculations. A Machine Learning based division strategy is proposed this paper, which can accurately quickly predict unnecessary improve overall encoding speed. The uses logistic regression model [2] includes three features: Quantization Parameter (QP),...
Automated volumetric segmentation of the pancreas on cross-sectional imaging is needed for diagnosis and follow-up pancreatic diseases. While CT-based more established, MRI-based methods are understudied, largely due to a lack publicly available datasets, benchmarking research efforts, domain-specific deep learning methods. In this retrospective study, we collected large dataset (767 scans from 499 participants) T1-weighted (T1W) T2-weighted (T2W) abdominal MRI series five centers between...
Abstract This study uses molecular dynamics simulations to investigate the wetting behavior of water droplets on a gold substrate with annular grooves. The research finds that droplet size and hydrophilicity similarly affect behavior. Enhanced changes velocity trend contact line movement, while impacts magnitude change. Contact angle fluctuations are observed analyzed theoretically based Young's equation. results provide insights into relationship between surface structure, contributing...