Xintong Zhao

ORCID: 0000-0001-8401-356X
Publications
Citations
Views
---
Saved
---
About
Contact & Profiles
Research Areas
  • Machine Learning in Materials Science
  • Data Quality and Management
  • Semantic Web and Ontologies
  • Industrial Vision Systems and Defect Detection
  • Advancements in Photolithography Techniques
  • Topic Modeling
  • Biomedical Text Mining and Ontologies
  • Integrated Circuits and Semiconductor Failure Analysis
  • Research Data Management Practices
  • Metal-Organic Frameworks: Synthesis and Applications
  • Scientific Computing and Data Management
  • Advanced Graph Neural Networks
  • Image Processing Techniques and Applications
  • Geochemistry and Geologic Mapping
  • Computational Drug Discovery Methods
  • Covalent Organic Framework Applications
  • Advanced Text Analysis Techniques
  • Advanced Neural Network Applications
  • Natural Language Processing Techniques
  • X-ray Diffraction in Crystallography
  • Electron and X-Ray Spectroscopy Techniques
  • VLSI and FPGA Design Techniques
  • Image and Signal Denoising Methods

Shanghai Huali Microelectronics (China)
2023-2024

Drexel University
2021-2024

Sichuan University of Science and Engineering
2023

Abstract Purpose This paper reports on a scientometric analysis bolstered by human-in-the-loop, domain experts, to examine the field of metal-organic frameworks (MOFs) research. Scientometric analyses reveal intellectual landscape field. The study engaged MOF scientists in design and review our research workflow. materials are an essential component next-generation renewable energy storage biomedical technologies. approach demonstrates how engaging via human-in-the-loop processes, can help...

10.2478/jdis-2024-0019 article EN Journal of Data and Information Science 2024-06-01

Scientific literature presents a wellspring of cutting-edge knowledge for materials science, including valuable data (e.g., numerical from experiment results, material properties and structure). These are critical accelerating discovery by data-driven machine learning (ML) methods. The challenge is, it is impossible humans to manually extract retain this due the extensive growing volume publications.To end, we explore fine-tuned BERT model extracting knowledge. Our preliminary results show...

10.1109/bigdata52589.2021.9671697 article EN 2021 IEEE International Conference on Big Data (Big Data) 2021-12-15

Building a knowledge graph is time-consuming and costly process which often applies complex natural language processing (NLP) methods for extracting triples from text corpora. Pre-trained large Language Models (PLM) have emerged as crucial type of approach that provides readily available range AI applications. However, it unclear whether feasible to construct domain-specific graphs PLMs. Motivated by the capacity accelerate data-driven materials discovery, we explored set state-of-the-art...

10.1109/bigdata55660.2022.10020568 article EN 2021 IEEE International Conference on Big Data (Big Data) 2022-12-17

In this position paper, we describe research on knowledge graph-empowered materials science prediction and discovery. The consists of several key components including ontology mapping, data annotation, information extraction from unstructured scholarly articles. We argue that although big generated by simulations experiments have motivated accelerated the data-driven science, distribution heterogeneity science-related hinders major advancements in field. Knowledge graphs, as semantic hubs,...

10.1109/bigdata52589.2021.9671503 article EN 2021 IEEE International Conference on Big Data (Big Data) 2021-12-15

To address the problems of low productivity and sizeable dispensing positioning errors in manual semi-automatic processes small- medium-sized electronic enterprises, this study proposes a fully automatic method based on visual with RJDNEL-type PCBs as research object. The system is constructed through construction mechanical structure, selection optical equipment, debugging control system. imaging technology. Firstly, area extracted image preprocessing; then, edge detected by Sobel operator,...

10.3390/app13169206 article EN cc-by Applied Sciences 2023-08-13

ABSTRACT This paper reports on a demonstration of YAMZ (Yet Another Metadata Zoo) as mechanism for building community consensus around metadata terms. The is motivated by the complexity standards environment and need more user-friendly approaches researchers to achieve vocabulary consensus. reviews series standardization challenges, explores crowdsourcing factors that offer possible solutions, introduces system. A presented with members Toberer materials science laboratory at Colorado School...

10.1162/dint_a_00211 article EN Data Intelligence 2023-01-01

Traditional Scanning Electron Microscopy (SEM) image noise reduction techniques, such as frame averaging or utilizing higher resolution SEM images, may result in potential electron beam damage and could also limit the speed of screening. In this paper, we propose a deep-learning-based denoising method using Transformer-based architecture that addresses these challenges. This effectively reduces while preserving details, providing comparable measurements line width roughness are only...

10.1109/iwaps60466.2023.10366153 article EN 2023-10-26

In semiconductor manufacturing, automatic defect classification is of paramount importance. Even the slightest defects can compromise chip performance or lead to complete failure, subsequently impacting yield rates. Currently, still heavily relies on manual processes, often leading a significant number misclassifications. this paper, we propose method based MAE (Masked Autoencoder) for in manufacturing. The core concept involves applying high-proportion random mask images, creating...

10.1109/iwaps60466.2023.10366134 article EN 2023-10-26

Knowledge Organization Systems (KOS) as networks of knowledge have the potential to inform AI operations. This paper explores natural language processing and machine learning in context KOS Helping Interdisciplinary Vocabulary Engineering (HIVE) technology. The presents three use cases: HIVE Historical Networks, for Materials Science (HIVE-4-MAT), Using Enhance Explore Medical Ontologies. background section reviews foundations, while cases provide a frame reference discussing current...

10.1080/01639374.2021.1995918 article EN Cataloging & Classification Quarterly 2021-11-10

Purpose The output of academic literature has increased significantly due to digital technology, presenting researchers with a challenge across every discipline, including materials science, as it is impossible manually read and extract knowledge from millions published literature. purpose this study address by exploring extraction in applied scholarship. An overriding goal help inform readers about the status science. Design/methodology/approach authors conducted two-part analysis,...

10.1108/el-11-2020-0320 article EN The Electronic Library 2021-08-09

Abstract Scientific literature is one of the most significant resources for sharing knowledge. Researchers turn to scientific as a first step in designing an experiment. Given extensive and growing volume literature, common approach reading manually extracting knowledge too time consuming, creating bottleneck research cycle. This challenge spans nearly every domain. For materials science, experimental data distributed across millions publications are extremely helpful predicting properties...

10.1002/pra2.497 article EN publisher-specific-oa Proceedings of the Association for Information Science and Technology 2021-10-01

HIVE4MAT is a linked data interactive application for navigating ontologies of value to materials science. HIVE enables automatic indexing textual resources with standardized terminology. This article presents the motivation underlying HIVE4MAT, explains system architecture, reports on two evaluations, and discusses future plans.

10.48550/arxiv.2411.00676 preprint EN arXiv (Cornell University) 2024-11-01

In semiconductor manufacturing, the detection of defects efficiently and accurately plays an important role in improving production quality process optimization. However, most current defect inspection methods need to collect reference images on wafer. Based machine learning (ML) model, this paper first using layout generate corresponding Scanning Electron Microscopy (SEM) image as for defect, then by comparing similarity with generated achieve accurate identification localization defect....

10.1117/12.3052994 article EN 2024-12-10

In the integrated circuits field, rapid and accurate detection of defects anomalies is a critical factor in improving lithography process yields. Research on large-scale chip layout pattern feature extraction clustering algorithms plays crucial role enhancing manufacturing yield processes. This paper proposes graph matching-based method, leveraging high redundancy relatively simple circuit structure patterns. Our method innovatively employs graph-based representation to capture keypoint...

10.1117/12.3052980 article EN 2024-12-10

Photolithography is a pivotal stage in integrated circuit chip manufacturing, exerting direct influence on both the performance and yield of chips. Its efficacy hinges heavily meticulous control parameters such as focus exposure dose. Traditionally, production speed limited by multiply rounds lengthy production-adjust process. Speeding up this process manufacturing has become pressing problem. To tackle challenge, we introduce novel framework that integrates conditional adversarial network...

10.1117/12.3052912 article EN 2024-12-10

We present a comprehensive benchmark dataset for Knowledge Graph Question Answering in Materials Science (KGQA4MAT), with focus on metal-organic frameworks (MOFs). A knowledge graph (MOF-KG) has been constructed by integrating structured databases and extracted from the literature. To enhance MOF-KG accessibility domain experts, we aim to develop natural language interface querying graph. have developed comprised of 161 complex questions involving comparison, aggregation, complicated...

10.48550/arxiv.2309.11361 preprint EN cc-by arXiv (Cornell University) 2023-01-01

Abstract Processes and practices—and in general, informational doings their diverse constellations—are pertinent elements of the information landscape. This panel presents research on documentation description processes practices field addressing: 1) how different conceptualisations influence they emerge as describable entities; 2) what approaches to document describe exist have been proposed science technology research; 3) aspects capture, make visible invisible; 4) novel insights from...

10.1002/pra2.509 article EN Proceedings of the Association for Information Science and Technology 2021-10-01

Growth in computational materials science and initiatives such as the Materials Genome Initiative (MGI) European Modelling Council (EMMC) has motivated development application of ontologies. A key factor been increased adoption FAIR principles, making research data findable, accessible, interoperable, reusable (Wilkinson et al. 2016). This paper characterizes semantic interoperability among a subset ontologies MatPortal repository. Background context covers interoperability, ontological...

10.48550/arxiv.2310.00078 preprint EN cc-by-sa arXiv (Cornell University) 2023-01-01

Research methods and procedures are core aspects of the research process. Metadata focused on these components is critical to supporting FAIR principles, particularly reproducibility. The reported in this paper presents a methodological framework for metadata documentation reproducibility producing Metal Organic Frameworks (MOFs). MOF case study involved natural language processing extract key synthesis experiment information from corpus literature. Following, classification activity was...

10.48550/arxiv.2310.12417 preprint EN cc-by arXiv (Cornell University) 2023-01-01

Automated image analysis and classification system that based on machine learning have been developed applied to the PWQ/FEM flow enhance process stability accuracy. Moving these task from artificial automation has added benefit of avoiding person-to-person inconsistencies in classification. In this paper, we approach a data consists an automated learning-enabled window relies CDSEM images taken Focus/Exposure Matrix wafer. Combined systematic preprocessing aggregation multi approaches,...

10.1109/iwaps60466.2023.10366089 article EN 2023-10-26

In semiconductor manufacturing, a forward etching process model that can accurately predict the pattern transformation from post-lithography (ADI) to post-etch (AEI) is greatly desired. However, current etch bias based, it unable offer rich information as SEM image does for engineers do wafer disposition. We propose an AEI generation using Pix2PixHD. It generate images corresponding ADI images. Experimental results demonstrate our generated achieve SSIM of 96.7 <sup...

10.1109/iwaps60466.2023.10366092 article EN 2023-10-26
Coming Soon ...