arXiv subjects

Jing Shao

Publications and source records attributed to Jing Shao.

At least 199 records · Page 11Linked to original sources

arXiv2018

Exploring Disentangled Feature Representation Beyond Face Identification

This paper proposes learning disentangled but complementary face features with minimal supervision by face identification. Specifically, we construct an identity Distilling and Dispelling Autoencoder (D2AE) framework that adversarially learns the identity-distilled features for identity verification and the identity-dispelled features to fool the verification system. Thanks to the design of two-stream cues, the learned disentangled features represent not only the identity or attribute but the complete input image. Comprehensive evaluations further demonstrate that the proposed features not only maintain state-of-the-art identity verification performance on LFW, but also acquire competitive discriminative power for face attribute recognition on CelebA and LFWA. Moreover, the proposed system is ready to semantically control the face generation/editing based on various identities and attributes in an unsupervised manner.

cs.CV↗

arXiv2017

HydraPlus-Net: Attentive Deep Features for Pedestrian Analysis

Pedestrian analysis plays a vital role in intelligent video surveillance and is a key component for security-centric computer vision systems. Despite that the convolutional neural networks are remarkable in learning discriminative features from images, the learning of comprehensive features of pedestrians for fine-grained tasks remains an open problem. In this study, we propose a new attention-based deep neural network, named as HydraPlus-Net (HP-net), that multi-directionally feeds the multi-level attention maps to different feature layers. The attentive deep features learned from the proposed HP-net bring unique advantages: (1) the model is capable of capturing multiple attentions from low-level to semantic-level, and (2) it explores the multi-scale selectiveness of attentive features to enrich the final feature representations for a pedestrian image. We demonstrate the effectiveness and generality of the proposed HP-net for pedestrian analysis on two tasks, i.e. pedestrian attribute recognition and person re-identification. Intensive experimental results have been provided to prove that the HP-net outperforms the state-of-the-art methods on various datasets.

cs.CV↗

arXiv2011

A Top Seesaw on a 5D Playground

We study quantum corrections to an extra dimensional Yukawa theory where a single flat extra spatial dimension is compactified on an interval. At a UV scale this theory can be made equivalent to a 5D theory with a bulk four-fermion operator by choosing appropriate boundary conditions for the running of the low-energy theory. Using the fermion bubble approximation we find vacuum solutions which break the chiral symmetries that arise from compactification. Of particular interest are the form of brane localized terms which arise from fermion loops. For example, quadratically divergent contributions to the scalar mass are absent at one loop due to a remnant of 5D Lorentz invariance that is only lost after introducing fermion bulk masses. The model is interpreted as an implementation of top condensation in extra dimensions with an automatic seesaw mechanism.

hep-ph↗

arXiv2011

The Baryon-Dark Matter Ratio Via Moduli Decay After Affleck-Dine Baryogenesis

Low-scale supersymmetry breaking in string motivated theories implies the presence of O(100) TeV scale moduli, which generically lead to a significant modification of the history of the universe prior to Big Bang Nucleosynthesis. Such an approach implies a non-thermal origin for dark matter resulting from scalar decay, where the lightest supersymmetric particle can account for the observed dark matter relic density. We study the further effect of the decay on the baryon asymmetry of the universe, and find that this can satisfactorily address the problem of the over-production of the baryon asymmetry by the Affleck-Dine mechanism in the MSSM. Remarkably, there is a natural connection between the baryon and dark matter abundances today, which leads to a solution of the `Cosmic Coincidence Problem'.

hep-ph↗

arXiv2011

Goldstone Fermion Dark Matter

We propose that the fermionic superpartner of a weak-scale Goldstone boson can be a natural WIMP candidate. The p-wave annihilation of this `Goldstone fermion' into pairs of Goldstone bosons automatically generates the correct relic abundance, whereas the XENON100 direct detection bounds are evaded due to suppressed couplings to the Standard Model. Further, it is able to avoid indirect detection constraints because the relevant s-wave annihilations are small. The interactions of the Goldstone supermultiplet can induce non-standard Higgs decays and novel collider phenomenology.

hep-ph↗

arXiv2011

Simplified Models for LHC New Physics Searches

This document proposes a collection of simplified models relevant to the design of new-physics searches at the LHC and the characterization of their results. Both ATLAS and CMS have already presented some results in terms of simplified models, and we encourage them to continue and expand this effort, which supplements both signature-based results and benchmark model interpretations. A simplified model is defined by an effective Lagrangian describing the interactions of a small number of new particles. Simplified models can equally well be described by a small number of masses and cross-sections. These parameters are directly related to collider physics observables, making simplified models a particularly effective framework for evaluating searches and a useful starting point for characterizing positive signals of new physics. This document serves as an official summary of the results from the "Topologies for Early LHC Searches" workshop, held at SLAC in September of 2010, the purpose of which was to develop a set of representative models that can be used to cover all relevant phase space in experimental searches. Particular emphasis is placed on searches relevant for the first ~50-500 pb-1 of data and those motivated by supersymmetric models. This note largely summarizes material posted at http://lhcnewphysics.org/, which includes simplified model definitions, Monte Carlo material, and supporting contacts within the theory community. We also comment on future developments that may be useful as more data is gathered and analyzed by the experiments.

hep-ph↗

arXiv2011

T-Quarks at the Large Hadron Collider: 2010-12

We study the potential of the current Large Hadron Collider (LHC) 7 TeV run to search for heavy, colored vector-like fermions, which are assumed to carry a conserved Z2 quantum number forcing them to be pair-produced. Each fermion is assumed to decay directly into a Standard Model quark and an invisible stable particle. T-odd quarks (T-quarks) and the lightest T-odd particle (LTP) of the Littlest Higgs model with T-parity provide an example of this setup. We estimate the bounds based on the published CMS search for events with jets and missing transverse energy in the 35 pb-1 data set collected in the 2010 run. We find that T-quark masses below about 450 GeV are ruled out for the LTP mass about 100 GeV. This bound is somewhat stronger than the published Tevatron constraint. We also estimate the reach with higher integrated luminosities expected in the 2011-12 run. If no deviation from the SM is observed, we expect that a bound on the T-quark mass of about 650 GeV, for the LTP mass of 300 GeV and below, can be achieved with 1 fb-1 of data. We comment on the possibility of using initial-state radiation jets to constrain the region with nearly-degenerate T-quark and LTP.

hep-ph↗

arXiv2011

SUSY Digs up a Buried Higgs

The Higgs boson may dominantly decay to 4 light jets through a light pseudo-scalar intermediary: h -> 2 eta -> 4j, making reconstruction at the LHC particularly challenging. We explore the phenomenology of such "Buried Higgs" scenarios in which the primary discovery channel of the Higgs is in cascade decays of superpartners. QCD backgrounds that would otherwise overwhelm the Higgs decay are suppressed by the requirement of high p_T jets and large missing transverse momentum that are the typical signatures of TeV scale supersymmetry. Utilizing jet substructure techniques, we find that for buried Higgses in the 100-120 GeV range, a 5-sigma discovery can be expected with roughly 10-25 inverse fb of data at E_CM = 14 TeV. For lighter Higgs bosons, the signal is contaminated by hadronically decaying W bosons, and discovery remains an unsolved challenge.

hep-ph↗

arXiv2011

CP-violating Phases in M-theory and Implications for EDMs

We demonstrate that in effective theories arising from a class of N=1 fluxless compactifications of M-theory on a G2 manifold with low energy supersymmetry, CP-violating phases do not appear in the soft-breaking Lagrangian except via the Yukawas appearing in the trilinear parameters. Such a mechanism may be present in other string compactifications as well; we describe properties sufficient for this to occur. CP violation is generated via the Yukawas since the soft trilinear matrices are generically not proportional to the Yukawa matrices. Within the framework considered, the estimated theoretical upper bounds for electric dipole moments (EDM) of the electron, neutron and mercury are all within the current experimental limits and could be probed in the near future.

hep-ph↗

arXiv2010

Mass Measurement in Boosted Decay Chains with Missing Energy

We explore a novel method of mass reconstruction in events with missing transverse momentum at hadron colliders. In events with sizeable boost factors in the final steps of dual multi-stage decay chains, the missing energy particles may each be approximately collinear with a visible standard model particle, spanning a narrow "MET-cone." We exploit this collinear approximation, when applicable, to reconstruct the masses of exotica.

hep-ph↗

arXiv2010

F-theory and the LHC: Stau Search

F-theory GUT models favor a relatively narrow range of soft supersymmetry breaking parameters in the MSSM Lagrangian. This leads to the specific predictions that a 10-100 MeV mass gravitino is the LSP, and the NLSP is quasi-stable, with a lifetime between a second to an hour. In a wide range of parameter space, the NLSP turns out to be a stau, though a bino-like lightest neutralino is also possible. Focusing on F-theory GUTs with a stau NLSP, we study the discovery potential at the LHC for such scenarios. Models with a quasi-stable stau predict a striking signature of a heavy charged particle passing through the detector. As a function of the parameters of minimal F-theory GUTs, we study how many of such events to expect, and additional signatures correlated with the presence of quasi-stable staus. We also study the prospects for staus to become stopped in or near the detector, as well as potential ways to distinguish such models from minimal gauge mediation models with similar spectra.

hep-ph↗

arXiv2009

Enhanced Tau Lepton Signatures at LHC in Constrained Supersymmetric Seesaw

We discuss the possible enhancement of the tau lepton events at LHC when the left-handed stau doublet becomes light (which can be even lighter than the right-handed stau). This is illustrated in the constrained supersymmetric seesaw model where the slepton doublet mass is suppressed by the effects of a large neutrino Yukawa coupling. We study a few representative parameter sets in the sneutrino coannihilation regions where the tau sneutrino is NLSP and the stau coannihilation regions where the stau is NLSP both of which yield the thermal neutralino LSP abundance determined by WMAP.

hep-ph↗

arXiv2009

The Footprint of F-theory at the LHC

Recent work has shown that compactifications of F-theory provide a potentially attractive phenomenological scenario. The low energy characteristics of F-theory GUTs consist of a deformation away from a minimal gauge mediation scenario with a high messenger scale. The soft scalar masses of the theory are all shifted by a stringy effect which survives to low energies. This effect can range from 0 GeV up to ~ 500 GeV. In this paper we study potential collider signatures of F-theory GUTs, focussing in particular on ways to distinguish this class of models from other theories with an MSSM spectrum. To accomplish this, we have adapted the general footprint method developed recently for distinguishing broad classes of string vacua to the specific case of F-theory GUTs. We show that with only 5 fb^(-1) of simulated LHC data, it is possible to distinguish many mSUGRA models and low messenger scale gauge mediation models from F-theory GUTs. Moreover, we find that at 5 fb^(-1), the stringy deformation away from minimal gauge mediation produces observable consequences which can also be detected to a level of order ~ +/- 80 GeV. In this way, it is possible to distinguish between models with a large and small stringy deformation. At 50 fb^(-1), this improves to ~ +/- 10 GeV.

hep-ph↗

arXiv2008

Initial determination of the spins of the gluino and squarks at LHC

In principle particle spins can be measured from their production cross sections once their mass is approximately known. The method works in practice because spins are quantized and cross sections depend strongly on spins. It can be used to determine, for example, the spin of the top quark. Direct application of this method to supersymmetric theories will have to overcome the challenge of measuring mass at the LHC, which could require high statistics. In this article, we propose a method of measuring the spins of the colored superpatners by combining rate information for several channels and a set of kinematical variables, without directly measuring their masses. We argue that such a method could lead to an early determination of the spin of gluino and squarks. This method can be applied to the measurement of spin of other new physics particles and more general scenarios.

hep-ph↗

arXiv2008

Non-thermal Dark Matter and the Moduli Problem in String Frameworks

We address the cosmological moduli/gravitino problems and the issue of too little thermal but excessive non-thermal dark matter from the decays of moduli. The main examples we study are the G2-MSSM models arising from M theory compactifications, which allow for a precise calculation of moduli decay rates and widths. We find that the late decaying moduli satisfy both BBN constraints and avoid the gravitino problem. The non-thermal production of wino LSPs, which is a prediction of G2-MSSM models, gives a relic density of about the right order of magnitude.

hep-ph↗

arXiv2008

Explaining the Electroweak Scale and Stabilizing Moduli in M Theory

In a recent paper \cite{Acharya:2006ia} it was shown that in $M$ theory vacua without fluxes, all moduli are stabilized by the effective potential and a stable hierarchy is generated, consistent with standard gauge unification. This paper explains the results of \cite{Acharya:2006ia} in more detail and generalizes them, finding an essentially unique de Sitter (dS) vacuum under reasonable conditions. One of the main phenomenological consequences is a prediction which emerges from this entire class of vacua: namely gaugino masses are significantly suppressed relative to the gravitino mass. We also present evidence that, for those vacua in which the vacuum energy is small, the gravitino mass, which sets all the superpartner masses, is automatically in the TeV - 100 TeV range.

hep-th↗

arXiv2008

The $G_2$-MSSM - An $M$ Theory motivated model of Particle Physics

We continue our study of the low energy implications of $M$ theory vacua on $G_2$ manifolds, undertaken in \cite{Acharya:2007rc,Acharya:2006ia}, where it was shown that the moduli can be stabilized and a TeV scale generated, with the Planck scale as the only dimensionful input. A well-motivated phenomenological model - the $G_2$-MSSM, can be naturally defined within the above framework. In this paper, we study some of the important phenomenological features of the $G_2$-MSSM. In particular, the soft supersymmetry breaking parameters and the superpartner spectrum are computed. The $G_2$-MSSM generically gives rise to light gauginos and heavy scalars with wino LSPs when one tunes the cosmological constant. Electroweak symmetry breaking is present but fine-tuned. The $G_2$-MSSM is also naturally consistent with precision gauge coupling unification. The phenomenological consequences for cosmology and collider physics of the $G_2$-MSSM will be reported in more detail soon.

hep-ph↗

arXiv2007

Distributed delays stabilize neural feedback systems

We consider the effect of distributed delays in neural feedback systems. The avian optic tectum is reciprocally connected with the nucleus isthmi. Extracellular stimulation combined with intracellular recordings reveal a range of signal delays from 4 to 9 ms between isthmotectal elements. This observation together with prior mathematical analysis concerning the influence of a delay distribution on system dynamics raises the question whether a broad delay distribution can impact the dynamics of neural feedback loops. For a system of reciprocally connected model neurons, we found that distributed delays enhance system stability in the following sense. With increased distribution of delays, the system converges faster to a fixed point and converges slower toward a limit cycle. Further, the introduction of distributed delays leads to an increased range of the average delay value for which the system's equilibrium point is stable. The enhancement of stability with increasing delay distribution is caused by the introduction of smaller delays rather than the distribution per se.

physics.bio-ph↗