paper-with-me

홈 › Papers

Understanding and Improving Transformer From a Multi-Particle Dynamic System Point of View.

2020-02-26 · ICLR Workshop DeepDiffEq 2019 12 · Yiping Lu*, Zhuohan Li*, Di He, Zhiqing Sun, Bin Dong, Tao Qin, LiWei Wang, Tie-Yan Liu

The Transformer architecture is widely used in natural language processing. Despite its success, the design principle of the Transformer remains elusive. In this paper, we provide a novel perspective towards understanding the architecture: we show that the Transformer can be mathematically interpreted as a \emph{numerical Ordinary Differential Equation (ODE) solver for a convection-diffusion equation in a multi-particle dynamic system}. In particular, how words in a sentence are abstracted into contexts by passing through the layers of the Transformer can be interpreted as approximating multiple particles' movement in the space using the Lie-Trotter splitting scheme and the Euler's method. Inspired from such a relationship, we propose to replace the Lie-Trotter splitting scheme by the more accurate Strang-Marchuk splitting scheme and design a new network architecture called Macaron Net. Through extensive experiments, we show that the Macaron Net is superior to the Transformer on both supervised and unsupervised learning tasks.

📄 PDF Abstract BibTeX

Code (0)

등록된 구현이 없습니다.

Tasks

Sentence

Similar Papers 제목 키워드 기반

Understanding and Improving Transformer From a Multi-Particle Dynamic System Point of View

2019-06-06 · ICLR 2020 1 · Yiping Lu, Zhuohan Li, Di He, Zhiqing Sun 외

The Transformer architecture is widely used in natural language processing. Despite its success, the design principle of the Transformer remains elusive. In this paper, we provide a novel perspective towards understandin…

PositionSentence

Universal Approximation of Mean-Field Models via Transformers

2024-10-06 · Shiba Biswal, Karthik Elamvazhuthi, Rishi Sonthalia

This paper investigates the use of transformers to approximate the mean-field dynamics of interacting particle systems exhibiting collective behavior. Such systems are fundamental in modeling phenomena across physics, bi…

Transformer with Implicit Edges for Particle-based Physics Simulation

2022-07-22 · Yidi Shao, Chen Change Loy, Bo Dai

Particle-based systems provide a flexible and unified way to simulate physics systems with complex dynamics. Most existing data-driven simulators for particle-based systems adopt graph neural networks (GNNs) as their net…

Geometric Dynamics of Signal Propagation Predict Trainability of Transformers

2024-03-05 · Aditya Cowsik, Tamra Nebabu, Xiao-Liang Qi, Surya Ganguli

We investigate forward signal propagation and gradient back propagation in deep, randomly initialized transformers, yielding simple necessary and sufficient conditions on initialization hyperparameters that ensure traina…

Interpreting Transformers for Jet Tagging

2024-12-04 · Aaron Wang, Abhijith Gandrakota, Jennifer Ngadiuba, Vivekanand Sahu 외

Machine learning (ML) algorithms, particularly attention-based transformer models, have become indispensable for analyzing the vast data generated by particle physics experiments like ATLAS and CMS at the CERN LHC. Parti…

Jet Tagging