BZPEERPapers, people, evidence and research paths
Menu

Discover

SearchFind papers, people, topics, institutions and opportunities from one place.GraphSee how papers, authors, topics and evidence connect.

Research tools

HomeContinue from the papers, people and topics that matter now.CollectionsTurn saved papers into reading lists, evidence maps and shareable context.ResearchReconstruct the path that led you from search to paper to question.CollaborationMove from a paper, topic or expert request into a focused thread.PublishCreate full-text research articles with metadata, formulas and versions.

Network

ExpertsFind specialists whose work and reviews explain why they can help.CommunitiesJoin research groups organized around papers, topics and shared evidence.

Opportunities

ListingsBrowse roles, calls and collaborations connected to real research context.

Account

Log inReturn to your saved papers, graph views and active threads.Create accountStart saving papers, building a research trail and joining teams.
Log inCreate account

Source author record

Peihao Huang

Peihao Huang appears in the imported research catalog. Authorship, coauthor and topic links are available while profile ownership is still unclaimed.

ResearcherUnclaimed source record
quant-phcond-mat.mes-hallInformation RetrievalArtificial Intelligencecond-mat.mtrl-scicond-mat.supr-con

Catalog footprint

What is connected

14works
6topics
4close collaborators

Actions

Connect this record

Log in to claim
Open graphBrowse works

Research graph

See the researcher in context

Open full explorer

Inspect adjacent papers, topics, institutions and collaborators without losing the researcher page.

Building this map preview

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

Published work

14 published item(s)

preprint2022arXiv

AutoFAS: Automatic Feature and Architecture Selection for Pre-Ranking System

Industrial search and recommendation systems mostly follow the classic multi-stage information retrieval paradigm: matching, pre-ranking, ranking, and re-ranking stages. To account for system efficiency, simple vector-product based models are commonly deployed in the pre-ranking stage. Recent works consider distilling the high knowledge of large ranking models to small pre-ranking models for better effectiveness. However, two major challenges in pre-ranking system still exist: (i) without explicitly modeling the performance gain versus computation cost, the predefined latency constraint in the pre-ranking stage inevitably leads to suboptimal solutions; (ii) transferring the ranking teacher's knowledge to a pre-ranking student with a predetermined handcrafted architecture still suffers from the loss of model performance. In this work, a novel framework AutoFAS is proposed which jointly optimizes the efficiency and effectiveness of the pre-ranking model: (i) AutoFAS for the first time simultaneously selects the most valuable features and network architectures using Neural Architecture Search (NAS) technique; (ii) equipped with ranking model guided reward during NAS procedure, AutoFAS can select the best pre-ranking architecture for a given ranking teacher without any computation overhead. Experimental results in our real world search system show AutoFAS consistently outperforms the previous state-of-the-art (SOTA) approaches at a lower computing cost. Notably, our model has been adopted in the pre-ranking module in the search system of Meituan, bringing significant improvements.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Contrastive Information Transfer for Pre-Ranking Systems

Real-word search and recommender systems usually adopt a multi-stage ranking architecture, including matching, pre-ranking, ranking, and re-ranking. Previous works mainly focus on the ranking stage while very few focus on the pre-ranking stage. In this paper, we focus on the information transfer from ranking to pre-ranking stage. We propose a new Contrastive Information Transfer (CIT) framework to transfer useful information from ranking model to pre-ranking model. We train the pre-ranking model to distinguish the positive pair of representation from a set of positive and negative pairs with a contrastive objective. As a consequence, the pre-ranking model can make full use of rich information in ranking model's representations. The CIT framework also has the advantage of alleviating selection bias and improving the performance of recall metrics, which is crucial for pre-ranking models. We conduct extensive experiments including offline datasets and online A/B testing. Experimental results show that CIT achieves superior results than competitive models. In addition, a strict online A/B testing at one of the world's largest E-commercial platforms shows that the proposed model achieves 0.63\% improvements on CTR and 1.64\% improvements on VBR. The proposed model now has been deployed online and serves the main traffic of this system, contributing a remarkable business growth.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Sampling Is All You Need on Modeling Long-Term User Behaviors for CTR Prediction

Rich user behavior data has been proven to be of great value for Click-Through Rate (CTR) prediction applications, especially in industrial recommender, search, or advertising systems. However, it's non-trivial for real-world systems to make full use of long-term user behaviors due to the strict requirements of online serving time. Most previous works adopt the retrieval-based strategy, where a small number of user behaviors are retrieved first for subsequent attention. However, the retrieval-based methods are sub-optimal and would cause more or less information losses, and it's difficult to balance the effectiveness and efficiency of the retrieval algorithm. In this paper, we propose SDIM (Sampling-based Deep Interest Modeling), a simple yet effective sampling-based end-to-end approach for modeling long-term user behaviors. We sample from multiple hash functions to generate hash signatures of the candidate item and each item in the user behavior sequence, and obtain the user interest by directly gathering behavior items associated with the candidate item with the same hash signature. We show theoretically and experimentally that the proposed method performs on par with standard attention-based models on modeling long-term user behaviors, while being sizable times faster. We also introduce the deployment of SDIM in our system. Specifically, we decouple the behavior sequence hashing, which is the most time-consuming part, from the CTR model by designing a separate module named BSE (behavior Sequence Encoding). BSE is latency-free for the CTR server, enabling us to model extremely long user behaviors. Both offline and online experiments are conducted to demonstrate the effectiveness of SDIM. SDIM now has been deployed online in the search system of Meituan APP.

0 citations0 reviewsNo code linkedOpen access
preprint2021arXiv

Impact of $\mathcal{T}$-symmetry on spin decoherence and control in a synthetic spin-orbit field

The electrical control of a spin qubit in a quantum dot relies on spin-orbit coupling (SOC), which could be either intrinsic to the underlying crystal lattice or heterostructure, or extrinsic via, for example, a micro-magnet. Here we show that a key difference between the intrinsic SOC and the synthetic SOC introduced by a micro-magnet is their symmetry under time reversal. Specifically, the time-reversal symmetry ($\mathcal{T}$-symmetry) of the intrinsic SOC leads to not only the traditional van Vleck cancellation known for spin relaxation, but also vanishing spin dephasing to the lowest order of SOC, which we term as "longitudinal spin-orbit field cancellation". On the other hand, the synthetic SOC from a micro-magnet breaks the $\mathcal{T}$-symmetry, therefore eliminates both the "van Vleck cancellation" and the "longitudinal spin-orbit field cancellation". In other words, the effective field $\vecΩ$ experienced by the spin qubit does not depend on the quantization magnetic field anymore, and a longitudinal component is allowed for $\vecΩ$ to the first order of SOC. Consequently, spin relaxation and dephasing are qualitatively modified compared with the case of the intrinsic SOC. Furthermore, the fidelity of electric-dipole spin resonance based on $\vecΩ$ could be optimized, with potential applications in spin-based quantum computing.

0 citations0 reviewsNo code linkedOpen access
preprint2021arXiv

Pressure-induced Superconductivity in dual-topological semimetal Pt2HgSe3

Recently monolayer jacutingaite (Pt2HgSe3), a naturally occurring exfoliable mineral, discovered in Brazil in 2008, has been theoretically predicted as a candidate quantum spin Hall system with a 0.5 eV band gap, while the bulk form is one of only a few known dual-topological insulators which may host different surface states protected by symmetries. In this work, we systematically investigate both structure and electronic evolution of bulk Pt2HgSe3 under high pressure up to 96 GPa. The nontrivial topology persists up to the structural phase transition observed in the high-pressure regime. Interestingly, we found that this phase transition is accompanied by the appearance of superconductivity at around 55 GPa and the critical transition temperature Tc increases with applied pressure. Our results demonstrate that Pt2HgSe3 with nontrivial topology of electronic states displays new ground states upon compression and raises potentials in application to the next-generation spintronic devices.

0 citations0 reviewsNo code linkedOpen access
preprint2015arXiv

Phonon mediated spin relaxation in a moving quantum dot: Doppler shift, Cherenkov radiation, and spin relaxation boom

We study relaxation of a moving spin qubit caused by phonon noise. As we vary the speed of the qubit, we observe several interesting features in spin relaxation and the associated phonon emission, induced by Doppler effect. In particular, in the supersonic regime, the phonons emitted by the relaxing qubit is concentrated along certain directions, similar to the shock waves produced in classical Cherenkov effect. As the speed of the moving qubit increases from the subsonic regime to the supersonic regime, the qubit experiences a peak in the spin relaxation rate near the speed of sound, which we term a spin relaxation boom in analogy to the classical sonic boom. We also find that the moving spin qubit may have a lower relaxation rate than a static qubit, which hints at the possibility of coherence-preserving transportation for a spin qubit. While the physics we have studied here has strong classical analogies, we do find that quantum confinement for the spin qubit plays an important role in all the phenomena we observe. Specifically, it produces a correction on the Cherenkov angle, and removes the divergence in relaxation rate at the sonic barrier. It is our hope that our results would encourage further research into approaches for transferring and preserving quantum information in spin qubit architectures.

0 citations0 reviewsNo code linkedOpen access
preprint2014arXiv

Decoherence of a Driven Qubit

We study decoherence of a field-driven qubit in the presence of environmental noises. For a general qubit, we find that driving, whether on-resonance or off-resonance, alters the qubit decoherence rates (including dissipation and pure dephasing), allowing both blue and red sideband contributions from the reservoir. Depending on the noise spectral density, driving field detuning and driving field phase shift, the qubit decoherence rates could be either accelerated or reduced. We apply our general theory to the system of an electron spin qubit that is confined in a quantum dot and driven by an in-plane electric field. We analyze how spin relaxation induced by the electrical noise due to electron-phonon interaction varies as a function of driving frequency, driving magnitude, driving field phase shift and spin-orbit coupling strengths.

0 citations0 reviewsNo code linkedOpen access
preprint2014arXiv

Spin Relaxation due to Charge Noise

We study decoherence of an electron spin qubit in a quantum dot due to charge noise. We find that at the lowest order, the pure dephasing channel is suppressed for both $1/f$ charge noise and Johnson noise, so that charge noise leads to a pure relaxation channel of decoherence. Because of the weaker magnetic field dependence, the spin relaxation rate due to charge noise could dominate over phonon noise at low magnetic fields in a gate-defined GaAs or Si quantum dot or a InAs self-assembled quantum dot. Furthermore, in a large InAs self-assembled quantum dot, the spin relaxation rate due to phonon noise could be suppressed in high magnetic field, and the spin relaxation due to charge noise could dominate in both low and high magnetic field. Numerically, in a 1 Tesla field, the spin relaxation time due to typical charge noise is about $100$ s in Si, $0.1$ s in GaAs for a gate-defined quantum dot with a $1$ meV confinement, and $10$ $μ$s in InAs self-assembled quantum dot with a $4$ meV confinement.

0 citations0 reviewsNo code linkedOpen access
preprint2014arXiv

Spin relaxation in a Si quantum dot due to spin-valley mixing

We study the relaxation of an electron spin qubit in a Si quantum dot due to electrical noise. In particular, we clarify how the presence of conduction-band valleys influences spin relaxation. In single-valley semiconductor quantum dots, spin relaxation is through the mixing of spin and envelope orbital states via spin-orbit interaction. In Si, the relaxation could also be through the mixing of spin and valley states. We find that the additional spin relaxation channel, via spin-valley mixing and electrical noise, is indeed important for an electron spin in a Si quantum dot. By considering both spin-valley and intravalley spin-orbit mixings and Johnson noise in a Si device, we find that the spin relaxation rate peaks at the hot spot, where the Zeeman splitting matches the valley splitting. Furthermore, because of a weaker field dependence, the spin relaxation rate due to Johnson noise could dominate over phonon noise at low magnetic fields, which fits well with recent experiments.

0 citations0 reviewsNo code linkedOpen access
preprint2013arXiv

Effect of bath temperature on the quantum decoherence

The dynamics of a qubit in two different environments are investigated theoretically. The first environment is a two level system coupled to a bosonic bath. And the second one is a damped harmonic oscillator. Based on a unitary transformation, we find that the decoherence of the qubit can be reduced with increasing temperature $T$ in the first case, which agree with the results in [Phys. Rev. Lett. 100, 120401], whereas, it can not be reduced with $T$ in the second case. In both cases, the qubit dynamics are changed substantially as the coupling increases or finite detuning appears.

0 citations0 reviewsNo code linkedOpen access
preprint2013arXiv

Spin Qubit Relaxation in a Moving Quantum Dot

Long-range quantum communication for spin qubits is an important open problem. Here we study decoherence of an electron spin qubit that is being transported in a moving quantum dot. We focus on spin decoherence due to spin-orbit interaction and a random electric potential. We find that at the lowest order, the motion induces longitudinal spin relaxation, with a rate linear in the dot velocity. Our calculated spin relaxation time ranges from sub $μ$s in GaAs to above ms in Si, making this relaxation a significant decoherence channel. Our results also give clear indications on how to reduce the decoherence effect of electron motion.

0 citations0 reviewsNo code linkedOpen access
preprint2010arXiv

Comparison of analytical and numerical methods and the effect of bath coupling on the quantum decoherence

The dynamics of a qubit in a structured environment is investigated theoretically. One point of view of the model is the spin-boson model with a Lorentz shaped spectral density. An alternative view is a qubit coupled to harmonic oscillator (HO), which in turn coupled to a Ohmic environment. Two different methods are applied and compared for this problem. One is a perturbation method based on a unitary transformation. Since the transformed hamiltonian is of rotating wave approximation (RWA) form, we call it the transformed rotating wave approximation (TRWA) method. And the other one is the numerically exact method of the quasi-adiabatic propagator path-integral (QUAPI) method. TRWA method can be applied from the first point of view. And the QUAPI method can applied from both points of views. We find that from the 1st point of view QUAPI only works well for large $Γ$. Since the memory time is too long for the practical evaluation of QUAPI when $Γ$ is small. We call this treatment as QUAPI1. And from the 2nd point of view, QUAPI works well for small $Γ$, since the non-adiabatic effect become more important as $Γ$ increases, one need smaller time-step and more steps to obtain accurate result which also quickly runs out the computational resources. This treatment is called QUAPI2. We find that the TRWA method works well for the whole parameter range of $Γ$ and show good agreement with QUAPI1 and QUAPI2. On the other hand, we find that the decoherence of the qubit can be reduced with increasing coupling between HO and bath. This result may be relevant to the design of quantum computer.

0 citations0 reviewsNo code linkedOpen access
preprint2010arXiv

The dynamics of a qubit in a spin-boson environment: a comparison between analytical and numerical method

The dynamics of a qubit under the decoherence of a two level fluctuator (TLF) in addition to its coupling to a bosonic bath is investigated theoretically. Two different methods are applied and compared for this problem. One is a perturbation method based on a unitary transformation. With the merit of our unitary transformation, non-adiabatic effect can be taken into account efficiently. And the other one is the numerically exact method, namely the quasi-adiabatic path-integral (QUAPI) propagator technique. We find that the analytical method works well for a wide parameter range and show good agreement with QUAPI. On the other hand, The enhancement and the reduction of quantum decoherence of the qubit is checked with varying bath temperature $T$ and TLF-bath coupling.

0 citations0 reviewsNo code linkedOpen access
preprint2009arXiv

Non-Markovian dynamics of a biased qubit coupled to a structured bath

A new analytical approach, beyond rotating wave approximation, based on unitary transformations and the non-Markovian master equation for the density operator, is applied to treat the biased spin boson model with a Lorentzian structured bath for arbitrary detunings at zero temperature. Compared to zero bias, we find that the dynamics demonstrates two more damping oscillation frequencies and one additional relaxation frequency for non-zero bias, where one of the damping oscillation frequencies is a new effect. Analytical expressions for the non-Markovian dynamics and the corresponding spectrum, the localized-delocalized transition point, the coherent-incoherent transition point, the analytical ground energy, the renormalized tunneling factor and the susceptibility are determined. The sum rule and the Shiba relation are checked in the coherent regime.

0 citations0 reviewsNo code linkedOpen access