arXiv subjects

Zhengyu Su

Publications and source records attributed to Zhengyu Su.

7 recordsLinked to original sources

arXiv2026

OneShot: Index-in-Ranking with Neural Scoring for Large-Scale Retrieval

In modern recommendation systems, retrieval serves as a primary stage responsible for filtering billions of candidate items down to thousands prior to refined ranking. To make this massive search effective and efficient, the system relies on ranking accuracy and indexing efficiency. However, these two objectives are traditionally misaligned: while the former optimizes for the alignment between ranking predictions and user behavior, the latter optimizes for a structural grouping of item representations which enables fast search among billions of candidates. Thus, despite extensive efforts to scale up interaction modeling for retrieval, they remain fundamentally limited by the structural misalignment between the ranking objectives and the proximity-learned index. In this work, we address this long-standing dichotomy by proposing a new holistic retrieval framework, OneShot. It is an end-to-end, in-model index learning framework that natively aligns index learning with ranking objectives. Using this joint learning as a structural foundation, OneShot pushes the boundaries of retrieval expressiveness by scaling interaction modeling with neural scoring beyond the persistent dot-product bottleneck. OneShot is fully deployed in Instagram's industrial short-video recommendation system, driving significant wins in user daily sessions, engagement, and time-spent. Additionally, OneShot achieves a $20\%$ recall gain at the operational ranking volume and a 10x efficiency improvement at an equivalent recall level.

cs.IR↗

arXiv2026

SlimPer: Make Personalization Model Slim and Smart

Transformer-style architectures are increasingly adopted for industrial recommendation systems, yet they inherit a design premise misaligned with the task: generative models rely on per-token autoregressive prediction, which justifies maintaining large intermediate tensors that scale with sequence length. In contrast, recommendation systems produce a single set of relevance scores for each pair without token-level supervision. Leveraging this observation, we propose SlimPer, which reformulates personalized ranking as iterative refinement of a compact, unified knowledge base. At each layer, the model selectively queries raw multi-modal user-side tokens, computes explicit relevance matching scores, and refines the knowledge base, all in O(N) per-layer cost with a fixed-size intermediate representation. As a result, model depth is decoupled from user history length, enabling deeper relevance understanding without proportional growth in compute or memory; request-only optimization further trims memory by sharing a single copy of user-side tokens across all candidate items. SlimPer unifies sparse, dense, and sequence features within a single backbone and provides inherent interpretability through its attention mechanism. Deployed on Instagram Reels and Feed, SlimPer yields measurable improvements in user engagement while streamlining the overall system and enabling effective modeling of 10k+ fine-grained user history events.

cs.IR↗

arXiv2026

The Bell-Bloom-type optically-pumped FID Rubidium atomic magnetometer with a multi-passing probe beam and two counter-propagating pump beams

The Bell-Bloom-type optically pumped atomic magnetometers are well suited for weak geomagnetic field detection. However, conventional single-beam pumping introduces an atomic spin polarization gradient, which limits the measurement accuracy and sensitivity. To address this issue, this paper proposes and experimentally demonstrates a Bell-Bloom-type rubidium FID magnetometer scheme integrating orthogonally polarized counter-propagating pumping and multi-pass probe detection. This design homogenizes the atomic spin polarization distribution and suppresses light shifts and power broadening effects induced by the pump beam. Meanwhile, the five-pass probe configuration significantly enhances the signal amplitude. Experimental results reveal that, compared with the traditional single-beam pumping and single-pass detection scheme, the proposed magnetometer achieves a remarkable improvement in magnetic field measurement accuracy, and the magnetic field sensitivity is improved from 18.9 pT/\sqrt{Hz} to 3.1 pT/\sqrt{Hz}. This work provides an effective technical approach and reference for optimizing the performance of atomic magnetometers and extending their applications in integrated arrays.

physics.atom-ph↗

arXiv2026

Optimization and vectorization of a Mz-type optically-pumped Rubidium magnetometer

Optically pumped magnetometers (OPMs) have demonstrated significant potential in weak magnetic field detection due to their high sensitivity. In this study, we developed an Mz-type optically pumped rubidium magnetometer using a paraffin-coated anti-relaxation vapor cell. The system optimization and performance characterization were conducted inside a magnetic shield. Specifically, the pump light intensity and radio-frequency (RF) magnetic field were jointly optimized by using the linewidth-amplitude ratio as the core metric. Based on the frequency-domain noise spectrum, the sensitivity in open-loop mode was measured to be approximately 30.8 pT/Hz^{1/2}. Furthermore, a closed-loop feedback locking technique was applied, reducing the measured noise floor under the tested conditions and improving the sensitivity to 22.9 pT/Hz^{1/2}, with a measured -3 dB bandwidth of 123 Hz. The dynamic characteristics were evaluated via magnetic-field step response, showing that the system could track magnetic-field changes stably under closed-loop operation. Finally, by using tri-axial modulation and frequency-domain demodulation, we overcame the scalar measurement limitation of traditional Mz magnetometers. This work realizes vector magnetic field detection and provides a technical basis for applications such as geomagnetic navigation and magnetic anomaly detection.

physics.atom-ph↗

arXiv2025

Anomalous Gate-tunable Capacitance in Van der Waals Heterostructures

The ferroelectricity emerging in non-polar graphene/hexagonal boron nitride (hBN) heterostructures has drawn considerable attention because of its fascinating properties and promising high-frequency electrical polarization switching. Yet, the underlying mechanism is still under debate. Here in twisted double bilayer graphene (TDBLG) aligned with its neighboring hBN, we observed two types of hysteresis - delayed hysteresis in top gate induced by the anomalous screening, and advanced hysteresis in back gate caused by the anomalous gate-tunable capacitance. To investigate the role played by moiré potential in the anomalous hysteresis, we studied a moiréless graphene heterostructure as control experiment. Unexpectedly, we observed exactly the same phenomena in this control device. Our findings suggest that the anomalous ferroelectricity in graphene/hBN heterostructures may originate from the dielectric material hBN, calling for further structural investigations on hBN. The observation of gate-tunable capacitance provides more insights in the mysterious ferroelectricity in graphene/hBN heterostructures, and should enable new design of memory devices such as memcapacitor based on tunable capacitance.

cond-mat.mes-hall↗

arXiv2018

Spherical Parameterization Balancing Angle and Area Distortions

This work presents a novel framework for spherical mesh parameterization. An efficient angle-preserving spherical parameterization algorithm is introduced, which is based on dynamic Yamabe flow and the conformal welding method with solid theoretic foundation. An area-preserving spherical parameterization is also discussed, which is based on discrete optimal mass transport theory. Furthermore, a spherical parameterization algorithm, which is based on the polar decomposition method, balancing angle distortion and area distortion is presented. The algorithms are tested on 3D geometric data and the experiments demonstrate the efficiency and efficacy of the proposed methods.

cs.GR↗

arXiv2016

Capacitated Kinetic Clustering in Mobile Networks by Optimal Transportation Theory

We consider the problem of capacitated kinetic clustering in which $n$ mobile terminals and $k$ base stations with respective operating capacities are given. The task is to assign the mobile terminals to the base stations such that the total squared distance from each terminal to its assigned base station is minimized and the capacity constraints are satisfied. This paper focuses on the development of \emph{distributed} and computationally efficient algorithms that adapt to the motion of both terminals and base stations. Suggested by the optimal transportation theory, we exploit the structural property of the optimal solution, which can be represented by a power diagram on the base stations such that the total usage of nodes within each power cell equals the capacity of the corresponding base station. We show by using the kinetic data structure framework the first analytical upper bound on the number of changes in the optimal solution, i.e., its stability. On the algorithm side, using the power diagram formulation we show that the solution can be represented in size proportional to the number of base stations and can be solved by an iterative, local algorithm. In particular, this algorithm can naturally exploit the continuity of motion and has orders of magnitude faster than existing solutions using min-cost matching and linear programming, and thus is able to handle large scale data under mobility.

cs.NI↗