Engineering topics

Feng, Yunhe

Publications and source records attributed to Feng, Yunhe.

4 records

DOE OSTI2024

Improving Text Classification with Large Language Model-Based Data Augmentation

Large Language Models (LLMs) such as ChatGPT possess advanced capabilities in understanding and generating text. These capabilities enable ChatGPT to create text based on specific instructions, which can serve as augmented data for text classification tasks. Previous studies have approached data augmentation (DA) by either rewriting the existing dataset with ChatGPT or generating entirely new data from scratch. However, it is unclear which method is better without comparing their effectiveness. This study investigates the application of both methods to two datasets: a general-topic dataset (Reuters news data) and a domain-specific dataset (Mitigation dataset). Our findings indicate that: 1. ChatGPT generated new data consistently enhanced model’s classification results for both datasets. 2. Generating new data generally outperforms rewriting existing data, though crafting the prompts carefully is crucial to extract the most valuable information from ChatGPT, particularly for domain-specific data. 3. The augmentation data size affects the effectiveness of DA; however, we observed a plateau after incorporating 10 samples. 4. Combining the rewritten sample with new generated sample can potentially further improve the model’s performance.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2021

Optimizing Error-Bounded Lossy Compression for Scientific Data on GPUs

Error-bounded lossy compression is a critical technique for significantly reducing scientific data volumes. With ever-emerging heterogeneous high-performance computing (HPC) architecture, GPU-accelerated error-bounded compressors (such as CUSZ and cuZFP) have been developed. However, they suffer from either low performance or low compression ratios. To this end, we propose CUSZ+ to target both high compression ratios and throughputs. We identify that data sparsity and data smoothness are key factors for high compression throughputs. Our key contributions in this work are fourfold: (1) We propose an efficient compression workflow to adaptively perform run-length encoding and/or variable-length encoding. (2) We derive Lorenzo reconstruction in decompression as multidimensional partial-sum computation and propose a fine-grained Lorenzo reconstruction algorithm for GPU architectures. (3) We carefully optimize each of CUSZ kernels by leveraging state-of-the-art CUDA parallel primitives. (4) We evaluate CUSZ+ using seven real-world HPC application datasets on V100 and A100 GPUs. Experiments show CUSZ+ improves the compression throughputs and ratios by up to 18.4x and 5.3x, respectively, over CUSZ on the tested datasets.

Tian, Jiannan↗

DOE OSTI2019

Real-Time Automated Health Information Technology Hazard Detection

The objective of this project is to develop an automated surveillance tool that will help the US Department of Veterans Affairs (VA) to detect potential hazards in health information technology (HIT) systems. To achieve this objective, Oak Ridge National Laboratory (ORNL) researchers have developed computational methods and a surveillance tool that will detect safety concerns and identify opportunities to improve HIT and safety within the VA system using Corporate Data Warehouse (CDW) data. This work is limited to the safe operation and safe use of HIT. It does not address the use of HIT to make health care safer. This project leveraged ORNL expertise in computational sciences and its computing resources accordingly.This report documents the FY 2019 accomplishments on this project. There were three major tasks. The first task developed data extraction approaches, which helped provide a better overall understanding of the important data elements in CDW. The outputs of this task served as inputs into the second task, which is the hazard detection and characterization task. This second task involved development of three methods that can be categorized into probabilistic model, event-based hazards detection, and system-based reliability performance evaluation. The third task dealt with the design, development, and demonstration of the automated surveillance tool. The tool currently resides in ORNL’s Private Health Information enclave at Oak Ridge, Tennessee, and operates on the VA’s CDW data. The developed capabilities and tool have been tested on CDW data from different clinical domains including Consult, Radiology, Laboratory, and Medication.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2019

DeepWelding: a Deep Learning Enhanced Approach to GTAW Using Multi-source Sensing Images

Deep learning has great potential to reshape manufacturing industries. In this paper, we present DeepWelding, a novel framework that applies deep learning techniques to improve gas tungsten arc welding (GTAW) process monitoring and penetration detection using multi-source sensing images. The framework is capable of analyzing multiple types of optical sensing images synchronously and consists of three deep learning enhanced consecutive phases: image preprocessing, image selection, and weld penetration classification. Specifically, we adopted generative adversarial networks (pix2pix) for image denoising and classic convolutional neural networks (AlexNet) for image selection. Both pix2pix and AlexNet delivered satisfactory performance. However, five individual neural networks with heterogeneous architectures demonstrated inconsistent generalization capabilities in the classification phase when holding out multi-source images generated with specific experimental settings. Therefore, two ensemble methods combining multiple neural networks were designed to improve the model performance on unseen data collected from different experimental settings. In conclusion, we also found that the quality of model prediction was heavily influenced by the data stream collection environment. We think these findings are beneficial for the broad intelligent welding community.

42 ENGINEERING↗