BZPEERReading, trust and collaboration for research
Menu

Discover

GraphFollow connections around a paper, topic or saved view.

Workspaces

HomePick up where your research flow left off.InboxHandle requests, replies and notifications from one place.CollectionsCurate reading lists, evidence sets and shareable dossiers.ResearchSee your private provenance graph, trail and imports.CollaborationMove from discovery into focused discussion rooms.PublishDraft, preview and publish full-text research articles.

Network

ExpertsFind specialists worth following or asking for help.CommunitiesJoin research circles organized around shared work.PartnersSee external organizations active around the same topics.

Opportunities

ListingsBrowse roles, calls and collaborations with clearer fit signals.

Account

Log inReturn to your reading and collaboration workspace.Create accountStart saving work, following people and joining teams.
Log inCreate account

Researcher profile

Fengshan Liu

Fengshan Liu contributes to research discovery and scholarly infrastructure.

ResearcherAffiliation not importedOpen to collaborate
physics.med-phastro-ph.COastro-ph.GAastro-ph.IMComputer VisionMachine Learningphysics.ao-phphysics.app-phphysics.data-an

Trust snapshot

Quick read

Trust 21 - EmergingVerification L1Unclaimed author
7works
0followers
9topics
4close collaborators

Actions

Decide how to stay connected

Follow researcher0

Identity and collaboration

How to connect with this researcher

Claiming links this public author record to a researcher profile and unlocks direct collaboration workflows.

Log in to claim

Direct collaboration

Open a focused conversation when the fit is right

Claim this author entity first to unlock direct invitations.

Research graph

See the researcher in context

Open full explorer

Inspect adjacent work, topics, institutions and collaborators without jumping out to a separate graph page.

Building this graph slice

BZPEER is loading the nearby papers, people, topics and institutions for this page.

Published work

7 published item(s)

preprint2025arXiv

Introduction to the Chinese Space Station Survey Telescope (CSST)

The Chinese Space Station Survey Telescope (CSST) is an upcoming Stage-IV sky survey telescope, distinguished by its large field of view (FoV), high image quality, and multi-band observation capabilities. It can simultaneously conduct precise measurements of the Universe by performing multi-color photometric imaging and slitless spectroscopic surveys. The CSST is equipped with five scientific instruments, i.e. Multi-band Imaging and Slitless Spectroscopy Survey Camera (SC), Multi-Channel Imager (MCI), Integral Field Spectrograph (IFS), Cool Planet Imaging Coronagraph (CPI-C), and THz Spectrometer (TS). Using these instruments, CSST is expected to make significant contributions and discoveries across various astronomical fields, including cosmology, galaxies and active galactic nuclei (AGN), the Milky Way and nearby galaxies, stars, exoplanets, Solar System objects, astrometry, and transients and variable sources. This review aims to provide a comprehensive overview of the CSST instruments, observational capabilities, data products, and scientific potential.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Comparison of measurement systems for assessing number- and mass-based particle filtration efficiency

The particle filtration efficiency (PFE) of a respirator or face mask is one of its key properties. While the physics of particle filtration results in the PFE being size-dependent, measurement standards are specified using a single, integrated PFE, for simplicity. This integrated PFE is commonly defined with respect to either the number (NBFE) or mass (MBFE) distribution of particles as a function of size. This relationship is non-trivial; it is influenced by both the shape of the particle distribution and the fact that multiple practical definitions of particle size are used. This manuscript discusses the relationship between NBFE and MBFE in detail, providing a guide to practitioners. Our discussion begins with a theoretical discussion of the underlying principles. We then present experimental results for a database of size-resolved PFE (SPFE) measurements for over 900 candidate respirators and filter media, including filter media with systematically varied properties and commercial samples that span the 20%-99.8% MBFE. The observed relationships between NBFE and MBFE are discussed in terms of the most-penetrating particle size (MPPS) and charge state of the media. For the NaCl particles used here, we observed that MBFEs were greater than NBFEs for charged materials and vice versa for uncharged materials. This relationship is observed because a shift from NBFE to MBFE weights the distribution towards larger sizes, while charged materials shift the MPPS to smaller sizes. Results are validated by comparing the output of a pair of TSI 8130As, a common instrument used in gauging standards compliance, to that of MBFEs computed from a system capable of measuring SPFE.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Size-dependent mass absorption cross-section of soot particles from various sources

The mass absorption cross-section (MAC) of combustion-generated soot is used in pollution and emissions measurements to quantify the mass concentration of soot and in atmospheric modelling to predict the radiative effects of soot on climate. Previous work has suggested that the MAC of soot particles may change with their size, due to (1) internal scattering among monomers in the soot aggregate, (2) the correlation of soot primary-particle diameter with aggregate size, (3) quantum confinement effects, or (4) a size-dependent degree of soot graphitization. Here, we report a size-dependent MAC for ex-situ soot sampled from two commercially available diffusion-flame soot generators, one aviation turbine engine, and one diesel generator. We also incorporate literature data. We show that the MAC increases with aggregate size until a plateau is reached at single particle masses between 4 and 30 fg (approximately 300-650 nm soot mobility diameter). The smallest particles may have MACs 50% to 80% smaller than the largest, depending on the source, while the largest particles have MACs within the range reported by previous measurements on polydisperse samples. Moreover, we show that models of hypotheses (1), (2), and (3) do not describe our measurement results, leaving hypothesis (4) as the only remaining candidate.

0 citations0 reviewsNo code linkedOpen access
preprint2021arXiv

IC Networks: Remodeling the Basic Unit for Convolutional Neural Networks

Convolutional neural network (CNN) is a class of artificial neural networks widely used in computer vision tasks. Most CNNs achieve excellent performance by stacking certain types of basic units. In addition to increasing the depth and width of the network, designing more effective basic units has become an important research topic. Inspired by the elastic collision model in physics, we present a general structure which can be integrated into the existing CNNs to improve their performance. We term it the "Inter-layer Collision" (IC) structure. Compared to the traditional convolution structure, the IC structure introduces nonlinearity and feature recalibration in the linear convolution operation, which can capture more fine-grained features. In addition, a new training method, namely weak logit distillation (WLD), is proposed to speed up the training of IC networks by extracting knowledge from pre-trained basic models. In the ImageNet experiment, we integrate the IC structure into ResNet-50 and reduce the top-1 error from 22.38% to 21.75%, which also catches up the top-1 error of ResNet-100 (21.75%) with nearly half of FLOPs.

0 citations0 reviewsNo code linkedOpen access
preprint2021arXiv

Systematic experimental comparison of particle filtration efficiency test methods for commercial respirators and face masks

Respirators, medical masks, and barrier face coverings all filter airborne particles using similar physical principles. However, they are tested for certification using a variety of standardized test methods, creating challenges for the comparison of differently certified products. We have performed systematic experiments to quantify and understand the differences between standardized test methods for N95 respirators (NIOSH TEB-APR-STP-0059 under US 42 CFR 84), medical face masks (ASTM F2299/F2100), and COVID-19-related barrier face coverings (ASTM F3502-21). Our experiments demonstrate the role of face velocity, particle properties (mean size, size variability, electric charge, density, and shape), measurement techniques, and environmental preconditioning. The measured filtration efficiency was most sensitive to changes in face velocity and particle charge. Relative to the NIOSH method, users of the ASTM F2299/F2100 method have commonly used non-neutralized (highly charged) aerosols as well as smaller face velocities, each of which may result in approximately 10% higher measured filtration efficiencies. In the NIOSH method, environmental conditioning at elevated humidity increased filtration efficiency in some commercial samples while decreasing it in others, indicating that measurement should be performed both with and without conditioning. More generally, our results provide an experimental basis for the comparison of respirators certified under various international methods, including FFP2, KN95, P2, Korea 1st Class, and DS2.

0 citations0 reviewsNo code linkedOpen access
preprint2020arXiv

IC-Network: Efficient Structure for Convolutional Neural Networks

Neural networks have been widely used, and most networks achieve excellent performance by stacking certain types of basic units. Compared to increasing the depth and width of the network, designing more effective basic units has become an important research topic. Inspired by the elastic collision model in physics, we present a universal structure that could be integrated into the existing network structures to speed up the training process and increase their generalization abilities. We term this structure the "Inter-layer Collision" (IC) structure. We built two kinds of basic computational units (IC layer and IC block) that compose the convolutional neural networks (CNNs) by combining the IC structure with the convolution operation. Compared to traditional convolutions, both of the proposed computational units have a stronger non-linear representation ability and can filter features useful for a given task. Using these computational units to build networks, we bring significant improvements in performance for existing state-of-the-art CNNs. On the imagenet experiment, we integrate the IC block into ResNet-50 and reduce the top-1 error from 22.85% to 21.49%, which also exceeds the top-1 error of ResNet-100 (21.75%).

0 citations0 reviewsNo code linkedOpen access
preprint2020arXiv

Quenching as a Contest between Galaxy Halos and their Central Black Holes

Existing models of galaxy formation have not yet explained striking correlations between structure and star-formation activity in galaxies, notably the sloped and moving boundaries that divide star-forming from quenched galaxies in key structural diagrams. This paper uses these and other relations to ``reverse-engineer'' the quenching process for central galaxies. The basic idea is that star-forming galaxies with larger radii (at a given stellar mass) have lower black-hole masses due to lower central densities. Galaxies cross into the green valley when the cumulative effective energy radiated by their black hole equals $\sim4\times$ their halo-gas binding energy. Since larger-radii galaxies have smaller black holes, one finds they must evolve to higher stellar masses in order to meet this halo-energy criterion, which explains the sloping boundaries. A possible cause of radii differences among star-forming galaxies is halo concentration. The evolutionary tracks of star-forming galaxies are nearly parallel to the green-valley boundaries, and it is mainly the sideways motions of these boundaries with cosmic time that cause galaxies to quench. BH-scaling laws for star-forming, quenched, and green-valley galaxies are different, and most BH mass growth takes place in the green valley. Implications include: the radii of star-forming galaxies are an important second parameter in shaping their black holes; black holes are connected to their halos but in different ways for star-forming, quenched, and green-valley galaxies; and the same BH-halo quenching mechanism has been in place since $z \sim 3$. We conclude with a discussion of black hole-galaxy co-evolution, the origin and interpretation of BH scaling laws.

0 citations0 reviewsNo code linkedOpen access