Yajun Peng

Yajun Peng

Auto-generated profile, based on publicly available information

Yajun Peng

Department of Geosciences Princeton University Princeton New Jersey USA

Research Focus & Metrics

Academic Career Length

15Years (2012-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
24 April 2015 in Journal of Geophysical Research: Solid Earth
Verified
Journal Article
Open access
PDF file
YPYajun Peng
Yajun Peng
ARAllan M. Rubin
Allan M. Rubin
MBMichael G. Bostock
Michael G. Bostock
JAJohn G. Armbruster
John G. Armbruster
We develop a cross‐station method to detect and locate tremor and low‐frequency earthquakes (LFEs), based on the original work of Armbruster et al. (2014) that compares waveforms from the same time window at stations separated by roughly 10 km. To improve the signal‐to‐noise ratio, we first rotate the horizontal components into the empirical shear wave particle motion direction. The large‐scale “rapid tremor reversals” beneath southern Vancouver Island are best recorded by stations that exhibit pronounced shear wave splitting, which obscures this optimal direction. We correct for splitting using the stacked templates of 11 low‐frequency earthquake families obtained in this region by Bostock et al. (2012). We find that the style of rapid tremor migrations (RTMs) evolves as the main front passes over a region. Very close to the main front, numerous small‐scale migrations occur with recurrence intervals far shorter than tidal periods. These usually propagate along the main front even when that is not parallel to dip. Several larger RTMs propagating along the main front have prominent elongation orthogonal to the propagation direction, inconsistent with the interpretation that their large propagation speed is an “apparent” velocity caused by the slow main front intersecting a preexisting linear structure on the plate interface. Farther behind the main front, RTMs gradually progress to being tidally modulated and have generally slower propagation speeds. Many reversal‐like RTMs are observed, some of which evolve from fronts initially propagating along the main front. These reversals are sometimes coherent across regions of low tremor density.

28 January 2016 in Geochemistry, Geophysics, Geosystems
Verified
Journal Article
Open access
PDF file
YPYajun Peng
Yajun Peng
ARAllan M. Rubin
Allan M. Rubin
Episodic tremor and slip (ETS) in subduction zones is generally interpreted as the manifestation of shear slip near the base of earthquake‐generating portion of the plate interface. Here we devise a new method of cross‐correlating stacked Array of Arrays seismic data that provides greatly improved tremor locations, a proxy for the underlying slow slip, beneath the Olympic Peninsula. This increased resolution allows us to image many features of tremor that were not visible previously. We resolve the spatial transition between the rupture zones of the inter‐ETS and major ETS episodes in 2010, suggesting stress redistribution by the former. Most tremor migrations propagated along the slowly advancing main tremor front during both the inter‐ETS and the major ETS episodes, even though the main front of the former deviated strongly from its usual (along‐dip) orientation. We find a distinct contrast between along‐dip rupture extent of large‐scale rapid tremor reversals (RTRs) to the south and that to the north in our study region that anticorrelates with the locations of inter‐ETS events. These RTRs originate from the main front, similar to smaller‐scale RTRs previously observed at high‐resolution, and many start by propagating along the main front. This could be consistent with RTRs being triggered by a cascading failure of brittle asperities. After initiation, the RTRs repeatedly occupy the same source region, and the early repetitions appear not to be tidally driven. Their stress drop may come from continuing fault weakening processes within the tremor zone, or loading by aseismic slip in surrounding regions.

26 January 2017 in Geophysical Research Letters
Verified
Journal Article
Open access
PDF file
YPYajun Peng
Yajun Peng
ARAllan M. Rubin
Allan M. Rubin
Slow slip events exhibit significant complexity in slip evolution and variations in recurrence intervals. Behavior that varies systematically with recurrence interval is likely to reflect different extents of fault healing between these events. Here we use high‐resolution tremor catalogs beneath Guerrero, Mexico, to investigate the mechanics of slow slip. We observe complex tremor propagation styles, including rapid tremor migrations propagating either along the main tremor front or backward, reminiscent of those in northern Cascadia. We also find many migrations that originate well behind the front and repeatedly occupy the same source region during a tremor episode, similar to those previously reported from Shikoku, Japan. These migrations could be driven by slow slip in the surrounding regions, with recurrence intervals possibly modulated by tides. The propagation speed of these migrations decreases systematically with time since the previous migration over the same source area. Tremor amplitudes seem consistent with changes in the propagation speeds being controlled primarily by changes in the slip speeds. One interpretation is that the high propagation speeds and inferred high slip speeds during the migrations with short recurrence intervals are caused by incomplete healing within the host rock adjacent to the shear zone, which could lead to high permeability and reduced dilatant strengthening of the fault gouge. Similar processes may operate in other slow slip source regions such as Cascadia.

13 May 2024
Verified
Conference Paper
Open access
BCBo Chang
Bo Chang
CMChangping Meng
Changping Meng
HMHe Ma
He Ma
SCShuo Chang
Shuo Chang
YGYang Gu
Yang Gu
MCMinmin Chen
Minmin Chen
Recommender systems are essential for finding personalized content for users on online platforms. These systems are often trained on historical user interaction data, which collects user feedback on system recommendations. This creates a feedback loop leading to popularity bias; popular content is over-represented in the data, better learned, and thus recommended even more. Less popular content struggles to reach its potential audiences. Popularity bias limits the diversity of content that users are exposed to, and makes it harder for new creators to gain traction. Existing methods to alleviate popularity bias tend to trade off the performance of popular items. In this work, we propose a new method for alleviating popularity bias in recommender systems, called the cluster anchor regularization, which partitions the large item corpus into hierarchical clusters, and then leverages the cluster information of each item to facilitate transfer learning from head items to tail items. Our results demonstrate the effectiveness of the proposed method with offline analyses and live experiments on a large-scale industrial recommendation platform, where it significantly increases tail recommendation without hurting the overall user experience.

18 April 2012 in Geophysical Journal International
Verified
Journal Article
Open access
PDF file
YPYajun Peng
Yajun Peng
SZShiyong Zhou
Shiyong Zhou
Jiancang ZhuangJiancang Zhuang
Jiancang Zhuang
JSJia Shi
Jia Shi
Large earthquakes could perturb the stress field in regions even thousands of kilometers away, leading to abrupt changes in background seismicity. We have developed a probability-based approach, based on the epidemic-type aftershock sequence model and the stochastic declustering method, to invert the smoothed temporal variation of background seismicity rate and to extract useful physical signals from complex seismicity patterns. An iterative algorithm is constructed to estimate the background seismicity simultaneously by using a combination of maximum likelihood estimate and weighted variable kernel estimate. We verify this approach through simulations and confirm that it can sensitively recover the onset of dynamic triggering. The algorithm is applied to an earthquake catalog in Yunnan Province, China, and successfully identifies a rapid increment of background seismicity rate following the occurrence of the 2004 Sumatra Mw 9.2 earthquake, whereas no remote triggering effect is detected following the occurrence of the 2005 Sumatra Mw 8.7 earthquake. This phenomenon agrees with GPS observations. It is found that the elevated seismic activity within 15 d after the Sumatra earthquake is mostly composed by shallow events, and direct triggering relationship is well established. Our approach also provides a way to detect the existence of other underlying physical processes based on catalog information.

2 October 2018 in Geophysical Research Letters
Verified
Journal Article
Open access
PDF file
YPYajun Peng
Yajun Peng
ARAllan M. Rubin
Allan M. Rubin
In this study, we analyze high‐resolution tremor catalogs from northern Cascadia, Guerrero, and northern Kii Peninsula. We find that tremor often occurs in short bursts that repeatedly occupy the same source area within a slow slip event. We hypothesize that these bursts are driven by loading from slow slip in areas surrounding the tremor zone. Adopting a rate‐and‐state friction law with a velocity‐weakening to velocity‐strengthening transition, we develop a finite fault model in which the size of the slow slip zone is larger than that of the tremor zone. Tidal forcing is added. Many asperities are randomly distributed within the tremor zone in order to generate burst‐like slip evolution while maintaining reasonable propagation speeds of the main slow slip front. We successfully reproduce the increasing recurrence intervals of the bursts as the main front moves across the tremor zone, as well as tidally modulated secondary fronts well behind the main front.