arXiv subjects

Meng Yu

Publications and source records attributed to Meng Yu.

88 records · Page 5Linked to original sources

arXiv2019

Equilibrium in Production Chains with Multiple Upstream Partners

In this paper, we extend and improve the production chain model introduced by Kikuchi et al. (2018). Utilizing the theory of monotone concave operators, we prove the existence, uniqueness, and global stability of equilibrium price, hence improving their results on production networks with multiple upstream partners. We propose an algorithm for computing the equilibrium price function that is more than ten times faster than successive evaluations of the operator. The model is then generalized to a stochastic setting that offers richer implications for the distribution of firms in a production network.

econ.TH↗

arXiv2019

A comprehensive study of speech separation: spectrogram vs waveform separation

Speech separation has been studied widely for single-channel close-talk microphone recordings over the past few years; developed solutions are mostly in frequency-domain. Recently, a raw audio waveform separation network (TasNet) is introduced for single-channel data, with achieving high Si-SNR (scale-invariant source-to-noise ratio) and SDR (source-to-distortion ratio) comparing against the state-of-the-art solution in frequency-domain. In this study, we incorporate effective components of the TasNet into a frequency-domain separation method. We compare both for alternative scenarios. We introduce a solution for directly optimizing the separation criterion in frequency-domain networks. In addition to speech separation objective and subjective measurements, we evaluate the separation performance on a speech recognition task as well. We study the speech separation problem for far-field data (more similar to naturalistic audio streams) and develop multi-channel solutions for both frequency and time-domain separators with utilizing spectral, spatial and speaker location information. For our experiments, we simulated multi-channel spatialized reverberate WSJ0-2mix dataset. Our experimental results show that spectrogram separation can achieve competitive performance with better network design. Multi-channel framework as well is shown to improve the single-channel performance relatively up to +35.5% and +46% in terms of WER and SDR, respectively.

cs.SD↗

arXiv2019

End-to-End Multi-Channel Speech Separation

The end-to-end approach for single-channel speech separation has been studied recently and shown promising results. This paper extended the previous approach and proposed a new end-to-end model for multi-channel speech separation. The primary contributions of this work include 1) an integrated waveform-in waveform-out separation system in a single neural network architecture. 2) We reformulate the traditional short time Fourier transform (STFT) and inter-channel phase difference (IPD) as a function of time-domain convolution with a special kernel. 3) We further relaxed those fixed kernels to be learnable, so that the entire architecture becomes purely data-driven and can be trained from end-to-end. We demonstrate on the WSJ0 far-field speech separation task that, with the benefit of learnable spatial features, our proposed end-to-end multi-channel model significantly improved the performance of previous end-to-end single-channel method and traditional multi-channel methods.

cs.SD↗

arXiv2019

Tracking the footprints of the radio pulsar B1727$-$47: proper motion, host supernova remnant, and the glitches

The bright radio pulsar B1727$-$47 with a characteristic age of 80 kyr is among the first pulsars discovered 50 yr ago. Using its regular timing observations and interferometric positions at three epochs, we measured, for the first time, the pulsar proper motion of $151 \pm 19$ mas yr$^{-1}$. At the dispersion measure distance of $\gtrsim 2.7$ kpc, this would suggest a record transverse velocity of the pulsar $\gtrsim 1900$ km s$^{-1}$. However, a backward extrapolation of the pulsar track to its birth epoch points remarkably close to the center of the evolved nearby supernova remnant RCW 114, which suggests genuine association of the two objects. In this case, the pulsar is substantially closer ($\sim 0.6$ kpc) and younger ($\sim 50$ kyr), and its velocity ($\sim400$ km s$^{-1}$) is compatible with the observed pulsar velocity distribution. We also identified two new glitches of the pulsar. We discuss implications of our results on the pulsar and remnant properties.

astro-ph.HE↗

arXiv2019

Improved Speaker-Dependent Separation for CHiME-5 Challenge

This paper summarizes several follow-up contributions for improving our submitted NWPU speaker-dependent system for CHiME-5 challenge, which aims to solve the problem of multi-channel, highly-overlapped conversational speech recognition in a dinner party scenario with reverberations and non-stationary noises. We adopt a speaker-aware training method by using i-vector as the target speaker information for multi-talker speech separation. With only one unified separation model for all speakers, we achieve a 10\% absolute improvement in terms of word error rate (WER) over the previous baseline of 80.28\% on the development set by leveraging our newly proposed data processing techniques and beamforming approach. With our improved back-end acoustic model, we further reduce WER to 60.15\% which surpasses the result of our submitted CHiME-5 challenge system without applying any fusion techniques.

eess.AS↗

arXiv2019

FAST ultra-wideband observation of abnormal emission-shift events of PSR B0919+06

PSR B0919+06 is known for its abnormal emission phenomenon, where the pulse emission window occasionally shifts progressively in longitude and returns afterwards. The physical mechanism behind this phenomenon is still under investigation. In this paper, we present our ultra-wideband observation of this pulsar using the Five-hundred-meter Aperture Spherical radio Telescope (FAST), with simultaneous measurements in the frequency ranges 280-780 MHz and 1250-1550 MHz. We have identified three abnormal events, each of which becomes less apparent as the frequency decreases. At 1400 MHz, the averaged profile slightly shifted after the first and third abnormal events, implying a relationship between abnormal event and profile variation. We also found a linear trend in the left-edge position of the averaged profiles between the first and third event as well as after the third event, suggesting the existence of a slow-drifting mode between the two major events. The second event has a comparatively small shift in phase and is thus categorized as a `small flare state'. During the third event, a sequence of approximately nine pulses was seen to significantly weaken in all frequency bands, likely associated with the pseudo-nulling observed at 150 MHz. A three-component de-composition analysis of the normal averaged profiles shows that the trailing component is dominant at our observing frequencies, while the centre component has a comparatively steeper spectrum. We found the overall flux density in an abnormal event to slightly differ from that in an ordinary state, and the difference shows a frequency dependence. A comparison of the normal, abnormal and dimmed averaged profile indicates that the leading component is likely to be stable in all states.

astro-ph.HE↗

arXiv2019

The Radiation Structure of PSR B2016$+$28 Observed with FAST

With the largest dish Five-hundred-meter Aperture Spherical radio Telescope (FAST), both the mean and single pulses of PSR B2016$+$28, especially including the single-pulse structure, are investigated in detail in this study. The mean pulse profiles at different frequencies can be well fitted in a conal model, and the peak separation of intensity-dependent pulse profiles increases with intensity. The integrated pulses are obviously frequency dependent (pulse width decreases by $\sim\,20\%$ as frequency increases from 300 MHz to 750 MHz), but the structure of single pulses changes slightly (the corresponding correlation scale decreases by only $\sim\,1\%$). This disparity between mean and single pulses provides independent evidence for the existence of the RS-type vacuum inner gap, indicating a strong bond between particles on the pulsar surface. Diffused drifting sub-pulses are analyzed. The results show that the modulation period along pulse series ($P_3$) is positively correlated to the separation between two adjacent sub-pulses ($P_2$). This correlation may hint a rough surface on the pulsar, eventually resulting in the irregular drift of sparks. All the observational results may have significant implications in the dynamics of pulsar magnetosphere and are discussed extensively in this paper.

astro-ph.HE↗

arXiv2019

Study of Three Rotating Radio Transients with FAST

Rotating radio transients (RRATs) are peculiar astronomical objects whose emission mechanism remains under investigation. In this paper, we present observations of three RRATs, J1538+2345, J1854+0306 and J1913+1330, observed with the Five-hundred-meter Aperture Spherical radio Telescope (FAST). Specifically, we analyze the mean pulse profiles and temporal flux density evolutions of the RRATs. Owing to the high sensitivity of FAST, the derived burst rates of the three RRATs are higher than those in previous reports. RRAT J1854+0306 exhibited a time-dynamic mean pulse profile, whereas RRAT J1913+1330 showed distinct radiation and nulling segments on its pulse intensity trains. The mean pulse profile variation with frequency is also studied for RRAT J1538+2345 and RRAT J1913+1330, and the profiles at different frequencies could be well fitted with a cone-core model and a conal-beam model, respectively.

astro-ph.HE↗

arXiv2018

How shall we determine detection sensitivity in radio pulsar search?

Determination of detection sensitivity in a number of previous pulsar search programmes was done via the straightfoward use of the radiometer equation. In the same surveys, the Fourier domain method was used to search for pulsars. As detection sensitivity is partially a function of the searching method, the straightfoward use of the radiometer equation for detection sensitivity determination is not consistent with the Fourier searching method. In this proceeding, I clarify the problem and note the way for sensitivity determination which is consistent with the Fourier searching method. More details can be found in Yu (2018).

astro-ph.IM↗

arXiv2018

Harmonic Summing Improves Pulsar Detection Sensitivity: A Probability Analysis

Practical application of the harmonic summing technique in the power-spectrum analysis for searching pulsars has exhibited the technique's effectiveness. In this paper, theoretical verification of harmonic summing considering power's noise-signal probability distribution is given. With the top-hat and the modified von Mises pulse profile models, contours along which spectra total power is expected to exceed the 3\,$σ$ detection threshold with 0.999 confidence corresponding to $m=1, 2, 4, 8, 16$, or 32 harmonics summed are given with respect to the mean pulse amplitude and the pulse duty cycle. Optimized numbers of harmonics summed relative to the duty cycles are given. The routine presented builds a theoretical estimate of the minimum detectable mean flux density, i.e. sensitivity, under the power-spectrum searching method.

astro-ph.IM↗

arXiv2018

Deep Extractor Network for Target Speaker Recovery From Single Channel Speech Mixtures

Speaker-aware source separation methods are promising workarounds for major difficulties such as arbitrary source permutation and unknown number of sources. However, it remains challenging to achieve satisfying performance provided a very short available target speaker utterance (anchor). Here we present a novel "deep extractor network" which creates an extractor point for the target speaker in a canonical high dimensional embedding space, and pulls together the time-frequency bins corresponding to the target speaker. The proposed model is different from prior works in that the canonical embedding space encodes knowledges of both the anchor and the mixture during an end-to-end training phase: First, embeddings for the anchor and mixture speech are separately constructed in a primary embedding space, and then combined as an input to feed-forward layers to transform to a canonical embedding space which we discover more stable than the primary one. Experimental results show that given a very short utterance, the proposed model can efficiently recover high quality target speech from a mixture, which outperforms various baseline models, with 5.2% and 6.6% relative improvements in SDR and PESQ respectively compared with a baseline oracle deep attracor model. Meanwhile, we show it can be generalized well to more than one interfering speaker.

cs.SD↗

arXiv2018

Have we seen all glitches?

Neutron star glitches are observed via artificially scheduled pulsar pulse arrival-time observations. Detection probability density of glitch events for a given data set is essentially required knowledge for realizing glitch detectability with specified observing system and schedule. In Yu & Liu, the detection probability density was derived for the Yu et al. data set. In this proceeding, further discussions are presented.

astro-ph.HE↗

arXiv2017

TrustShadow: Secure Execution of Unmodified Applications with ARM TrustZone

The rapid evolution of Internet-of-Things (IoT) technologies has led to an emerging need to make it smarter. A variety of applications now run simultaneously on an ARM-based processor. For example, devices on the edge of the Internet are provided with higher horsepower to be entrusted with storing, processing and analyzing data collected from IoT devices. This significantly improves efficiency and reduces the amount of data that needs to be transported to the cloud for data processing, analysis and storage. However, commodity OSes are prone to compromise. Once they are exploited, attackers can access the data on these devices. Since the data stored and processed on the devices can be sensitive, left untackled, this is particularly disconcerting. In this paper, we propose a new system, TrustShadow that shields legacy applications from untrusted OSes. TrustShadow takes advantage of ARM TrustZone technology and partitions resources into the secure and normal worlds. In the secure world, TrustShadow constructs a trusted execution environment for security-critical applications. This trusted environment is maintained by a lightweight runtime system that coordinates the communication between applications and the ordinary OS running in the normal world. The runtime system does not provide system services itself. Rather, it forwards requests for system services to the ordinary OS, and verifies the correctness of the responses. To demonstrate the efficiency of this design, we prototyped TrustShadow on a real chip board with ARM TrustZone support, and evaluated its performance using both microbenchmarks and real-world applications. We showed TrustShadow introduces only negligible overhead to real-world applications.

cs.CR↗

arXiv2017

On the detection probability of neutron star glitches

Neutron stars are observed to undergo small, abrupt rotational speed-up. This phenomenon is known as glitch. In pulsar timing observations, detection of a neutron star glitch is constrained by the time of occurrence of the event relative to entire observing span and observing cadences, time of occurrence of preceding/subsequent glitches relative to observing cadences and the strength of timing noise. Using the Yu et al. (2013) data sets, in this paper, we analyse the observational selection in terms of detection probability. We define partial probabilities for the constraints and use Monte Carlo method with assuming glitches distribute uniformly to solve the complete probability formula for both group case involving 157 pulsars and individual cases for each of the seven pulsars with glitch numbers $\geqslant 5$. In the simulations, numerical Bayesian inference is used for glitch identification. With the derived detection probability density and observed results, we uncover glitch size probability distribution embedded in the data for both the group and individual cases. We find the most prominent correction occurred for PSR~J1341$-$6220, in which exponent of the power-law model varies from the observed $+0.7^{+1.4}_{-0.7}$ to $-0.4^{+1.0}_{-0.4}$. We suggest observers determine the detection probability for glitch theories, e.g. the self-organised criticality.

astro-ph.HE↗

arXiv2012

Glitches detected in southern radio pulsars

Parkes pulse arrival-time data for 165 radio pulsars spanning from 1990 to 2011 have been searched for period glitches. Forty-six events out of the detected 107 glitches were found to be new contributions to the entire glitch population which currently contains approximately 400 events.

astro-ph.HE↗

arXiv2009

Cooling and Heating Solid Quark Stars

We present here a phenomenological solid quark star pulsar model to interpret the observed thermal X-ray emission of isolated pulsars. The heat capacity for solid quark stars was found to be quite small, so that the residual internal stellar heat gained at the birth of the star could be dissipated in an extremely short timescale. However, the bombardment induced by backflowing plasma at the poles of solid quark stars would get the stars be reheated, so that long term soft X-ray emission can be sustained. Such a scenario could be used for those X-ray pulsars with significant magnetospheric activities, and their cooling processes would thus be established. Dim X-ray isolated neutron stars (XDINs) as well as compact central objects (CCOs) have been observed with dominant soft X-ray radiation combined with little magnetospheric manifestations. Such sources could be solid quark stars accreting in the propeller regime.

astro-ph.HE↗