Universal transformer Hawkes process with adaptive recursive iteration

نویسندگان

چکیده

Asynchronous events sequences are widely distributed in the natural world and human activities, such as earthquakes records, users’ activities social media, so on. How to distill information from these seemingly disorganized data is a persistent topic that researchers focus One of most useful models point process model, on basis, obtain many noticeable results. Moreover, recent years, foundation neural networks, especially recurrent networks (RNN) proposed compare with traditional models, their performance greatly improved. Enlighten by transformer which can learn sequence efficiently without convolutional structure, Hawkes comes out, achieves state-of-the-art performance. However, there some research proving re-introduction recursive calculations further improve transformer’s Thus, we come out new kind universal (UTHP), contains both mechanism self-attention mechanism, local perception ability also introduce network (CNN) position-wise-feed-forward part. We conduct experiments several datasets validate effectiveness UTHP explore changes after introduction mechanism. These multiple demonstrate our model has certain improvement compared previous models.

برای دانلود باید عضویت طلایی داشته باشید

برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

منابع مشابه

Bi-iteration recursive instrumental variable subspace tracking and adaptive filtering

In this paper, we propose a class of fast sequential bi-iteration singular value (Bi-SVD) subspace tracking algorithms for adaptive eigendecomposition of the cross covariance matrix in the recursive instrumental variable (RIV) method of system identification. These algorithms can be used for RIV subspace processing of signals in unknown correlated Gaussian noise. Realizations with O(Nr) and O(N...

متن کامل

Modelling Dyadic Interaction with Hawkes Process

We apply Hawkes process to the analysis of dyadic interaction. Hawkes process is applicable to excitatory interactions, wherein the actions of each individual increase the probability of further actions in the near future. We consider the representation of Hawkes process both as a conditional intensity function and as a cluster Poisson process. The former treats the probability of an action in ...

متن کامل

Markov-modulated Hawkes process with stepwise decay

This paper proposes a new model—the Markov-modulated Hawkes process with stepwise decay (MMHPSD)—to investigate the variation in seismicity rate during a series of earthquake sequences including multiple main shocks. The MMHPSD is a self-exciting process which switches among different states, in each of which the process has distinguishable background seismicity and decay rates. Parameter estim...

متن کامل

Process Algebra with Iteration

We introduce iteration in process algebra by means of (the binary version of) Kleene's star operator: x y is the process that chooses between x and y, and upon termination of x has this choice again. It is proved that adding respectively interleaving, communication and abstraction operators increases expressivity up to the regular processes. However, if the distinction between (successful) term...

متن کامل

ذخیره در منابع من


  با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید

ژورنال

عنوان ژورنال: Engineering Applications of Artificial Intelligence

سال: 2021

ISSN: ['1873-6769', '0952-1976']

DOI: https://doi.org/10.1016/j.engappai.2021.104416