Zeroquant: Efficient and affordable post-training quantization for large-scale transformers

Z Yao, R Yazdani Aminabadi… - Advances in …, 2022 - proceedings.neurips.cc
How to efficiently serve ever-larger trained natural language models in practice has become
exceptionally challenging even for powerful cloud servers due to their prohibitive …

Unsupervised EHR‐based phenoty** via matrix and tensor decompositions

F Becker, AK Smilde, E Acar - Wiley Interdisciplinary Reviews …, 2023 - Wiley Online Library
Computational phenoty** allows for unsupervised discovery of subgroups of patients as
well as corresponding co‐occurring medical conditions from electronic health records …

Zeroquant (4+ 2): Redefining llms quantization with a new fp6-centric strategy for diverse generative tasks

X Wu, H ** electronic health records
A Afshar, I Perros, H Park, C Defilippi, X Yan… - Proceedings of the …, 2020 - dl.acm.org
Phenoty** electronic health records (EHR) focuses on defining meaningful patient groups
(eg, heart failure group and diabetes group) and identifying the temporal evolution of …

Static and streaming tucker decomposition for dense tensors

JG Jang, U Kang - ACM Transactions on Knowledge Discovery from …, 2023 - dl.acm.org
Given a dense tensor, how can we efficiently discover hidden relations and patterns in static
and online streaming settings? Tucker decomposition is a fundamental tool to analyze …

Fast and accurate domain adaptation for irregular tensor decomposition

J Kim, KH Park, JG Jang, U Kang - … of the 30th ACM SIGKDD Conference …, 2024 - dl.acm.org
Given an irregular tensor from a newly emerging domain, how can we quickly and
accurately capture its patterns utilizing existing irregular tensors in multiple domains? The …

Deepspeed data efficiency: Improving deep learning model quality and training efficiency via efficient data sampling and routing

C Li, Z Yao, X Wu, M Zhang, C Holmes, C Li… - Proceedings of the AAAI …, 2024 - ojs.aaai.org
Recent advances on deep learning models come at the price of formidable training cost. The
increasing model size is one of the root causes, but another less-emphasized fact is that …

Random-ltd: Random and layerwise token drop** brings efficient training for large-scale transformers

Z Yao, X Wu, C Li, C Holmes, M Zhang, C Li… - ar** Using Electronic Health Records
E Konyar, MR Gahrooei - IEEE Journal of Biomedical and …, 2025 - ieeexplore.ieee.org
Computational phenoty** uses data mining methods to extract clusters of clinical
descriptors, known as phenotypes, from electronic health records (EHR). Tensor …