Hpc-gpt: Integrating large language model for high-performance computing

X Ding, L Chen, M Emani, C Liao, PH Lin… - Proceedings of the SC' …, 2023 - dl.acm.org
Large Language Models (LLMs), including the LLaMA model, have exhibited their efficacy
across various general-domain natural language processing (NLP) tasks. However, their …

[HTML][HTML] Caching, transcoding, delivery and learning for advanced video streaming services

M Choi, T Xiang, H Lim, Y Kim, M Ahn, S Oh, H Kim - ICT Express, 2024 - Elsevier
Online video services are responsible for a large amount of global data traffic, which is why
wireless video caching has become a major focus in order to deal with the increasing …

Efficient distributed continual learning for steering experiments in real-time

T Bouvier, B Nicolae, A Costan, T Bicer, I Foster… - Future Generation …, 2025 - Elsevier
Deep learning has emerged as a powerful method for extracting valuable information from
large volumes of data. However, when new training data arrives continuously (ie, is not fully …

Lobster: Load balance-aware I/O for distributed DNN training

J Liu, B Nicolae, D Li - Proceedings of the 51st International Conference …, 2022 - dl.acm.org
The resource-hungry and time-consuming process of training Deep Neural Networks
(DNNs) can be accelerated by optimizing and/or scaling computations on accelerators such …

Accelerated laminographic image reconstruction using GPUs

B Ma, V Nikitin, D Li, T Bicer - Electronic Imaging, 2024 - library.imaging.org
Laminography is a specialized 3D imaging technique optimized for examining flat,
elongated structures. Laminographic reconstruction is the process of generating 3D volume …

Efficient Data-Parallel Continual Learning with Asynchronous Distributed Rehearsal Buffers

T Bouvier, B Nicolae, H Chaugier, A Costan… - arXiv preprint arXiv …, 2024 - arxiv.org
Deep learning has emerged as a powerful method for extracting valuable information from
large volumes of data. However, when new training data arrives continuously (ie, is not fully …

Diaspora: Resilience-Enabling Services for Real-Time Distributed Workflows

B Nicolae, JM Wozniak, T Bicer… - 2024 IEEE 20th …, 2024 - ieeexplore.ieee.org
The need for real-time processing to enable automated decision making and experimental
steering has driven a shift from high-performance computing workflows on a centralized …

Optimizing the Training of Co-Located Deep Learning Models Using Cache-Aware Staggering

K Assogba, B Nicolae… - 2023 IEEE 30th …, 2023 - ieeexplore.ieee.org
Despite significant advances, training deep learning models remains a time-consuming and
resource-intensive task. One of the key challenges in this context is the ingestion of the …

APAM: Adaptive Pre-Training and Adaptive Meta Learning in Language Model for Noisy Labels and Long-Tailed Learning

B Dong, Y Xu, S Chi, Z Shi, Z Du - … International Conference on …, 2023 - ieeexplore.ieee.org
Practical natural language processing (NLP) tasks often exhibit long-tailed distributions
accompanied by noisy labels, posing significant challenges to the generalization and …

Data-driven Performance Optimization for Data-intensive Applications

J Liu - 2024 - search.proquest.com
Data-intensive applications have attracted considerable attention from researchers in
information sciences and enterprises, as these applications have made evolutionary …