Search

Embedding techniques have become essential components of large databases in the deep learning era. By encoding discrete entities, such as words, items, or graph nodes, into continuous vector spaces, embeddings facilitate more efficient storage, retrieval, and processing in large databases. Especially in the domain of recommender systems, millions of categorical features are encoded as unique embedding vectors, which facilitates the modeling of similarities and interactions among features. However, numerous embedding vectors can result in significant storage overhead. In this paper, we aim to compress the embedding table through quantization techniques. Given that features vary in importance levels, we seek to identify an appropriate precision for each feature to balance model accuracy and memory usage. To this end, we propose a novel embedding compression method, termed Mixed-Precision Embeddings (MPE). Specifically, to reduce the size of the search space, we first group features by frequency and then search precision for each feature group. MPE further learns the probability distribution over precision levels for each feature group, which can be used to identify the most suitable precision with a specially designed sampling strategy. Extensive experiments on three public datasets demonstrate that MPE significantly outperforms existing embedding compression methods. Remarkably, MPE achieves about 200x compression on the Criteo dataset without comprising the prediction accuracy., Comment: under submision

Background-oriented Schlieren tomography (BOST) is a prevalent method for visualizing intricate turbulent flows, valued for its ease of implementation and capacity to capture three-dimensional distributions of a multitude of flow parameters. However, the voxel-based meshing scheme leads to significant challenges, such as inadequate spatial resolution, substantial discretization errors, poor noise immunity, and excessive computational costs. This work presents an innovative reconstruction approach termed neural refractive index field (NeRIF) which implicitly represents the flow field with a neural network, which is trained with tailored strategies. Both numerical simulations and experimental demonstrations on turbulent Bunsen flames suggest that our approach can significantly improve the reconstruction accuracy and spatial resolution while concurrently reducing computational expenses. Although showcased in the context of background-oriented schlieren tomography here, the key idea embedded in the NeRIF can be readily adapted to various other tomographic modalities including tomographic absorption spectroscopy and tomographic particle imaging velocimetry, broadening its potential impact across different domains of flow visualization and analysis., Comment: 10 pages, 5 figures

We propose the implementation of a rapid adiabatic passage (RAP) scheme to generate entanglement in Rydberg atom-array systems. This method transforms a product state in a multi-qubit system into an entangled state with high fidelity and robustness. By employing global and continuous driving laser fields, we demonstrate the generation of two-qubit Bell state and three-qubit W state, via sequential RAP pulses within the Rydberg blockade regime. As an illustrative example, applying this technique to alkali atoms, we predict fidelities exceeding 0.9995 for two-qubit Bell and three-qubit W state, along with excellent robustness. Furthermore, our scheme can be extended to generate entanglement between weakly coupled atoms and to create four-qubit Greenberger- Horne-Zeilinger states through spatial correlations. Our approach holds the potential for extension to larger atomic arrays, offering a straightforward and efficient method to generate high-fidelity entangled states in neutral atom systems., Comment: 9 pages, 5 figures

High-fidelity and robust quantum gates are essential for quantum information processing, where neutral Rydberg atoms trapped in optical tweezer arrays serving as a versatile platform for the implementation. We propose a rapid adiabatic passage (RAP) scheme for achieving a high-fidelity controlled-Z (CZ) gate on a neutral atom Rydberg platform. Utilizing only global laser dressing, our scheme involves echoing two identical RAP pulses within the Rydberg blockade regime to realize a CZ gate and can be readily extended to a C$^k$Z gate with additional qubits. We predict a CZ gate with fidelity over 0.9995 using akali-atom parameters, and a CCZ gate with fidelity exceeding 0.999. Moreover, the direct utilization of echoing RAP pulses enables the implementation of a four-bit CCCZ gate at fidelity over 0.996 without further optimization. The proposed scheme, remarkably robust to variations in driving fields and realistic decoherence effects, holds promise for future quantum information processing applications., Comment: (5+3) pages, 3 figures

Federated learning is a promising distributed training paradigm that effectively safeguards data privacy. However, it may involve significant communication costs, which hinders training efficiency. In this paper, we aim to enhance communication efficiency from a new perspective. Specifically, we request the distributed clients to find optimal model updates relative to global model parameters within predefined random noise. For this purpose, we propose Federated Masked Random Noise (FedMRN), a novel framework that enables clients to learn a 1-bit mask for each model parameter and apply masked random noise (i.e., the Hadamard product of random noise and masks) to represent model updates. To make FedMRN feasible, we propose an advanced mask training strategy, called progressive stochastic masking (PSM). After local training, each client only need to transmit local masks and a random seed to the server. Additionally, we provide theoretical guarantees for the convergence of FedMRN under both strongly convex and non-convex assumptions. Extensive experiments are conducted on four popular datasets. The results show that FedMRN exhibits superior convergence speed and test accuracy compared to relevant baselines, while attaining a similar level of accuracy as FedAvg., Comment: Accepted by MM 2024

Federated learning is a promising distributed machine learning paradigm that can effectively exploit large-scale data without exposing users' privacy. However, it may incur significant communication overhead, thereby potentially impairing the training efficiency. To address this challenge, numerous studies suggest binarizing the model updates. Nonetheless, traditional methods usually binarize model updates in a post-training manner, resulting in significant approximation errors and consequent degradation in model accuracy. To this end, we propose Federated Binarization-Aware Training (FedBAT), a novel framework that directly learns binary model updates during the local training process, thus inherently reducing the approximation errors. FedBAT incorporates an innovative binarization operator, along with meticulously designed derivatives to facilitate efficient learning. In addition, we establish theoretical guarantees regarding the convergence of FedBAT. Extensive experiments are conducted on four popular datasets. The results show that FedBAT significantly accelerates the convergence and exceeds the accuracy of baselines by up to 9\%, even surpassing that of FedAvg in some cases., Comment: Accepted by ICML 2024

As an emerging non-volatile memory technology, magnetic random access memory (MRAM) has key features and advantages including non-volatility, high speed, endurance, low power consumption and radiation tolerance. Conventional MRAM utilizes magnetic tunnel junctions (MTJs), which consist of two ferromagnetic layers separated by an insulating tunnel barrier. The orientation of the magnetic layers represents the binary data (0 or 1), and electrical resistance changes depending on the relative orientation of these magnetic layers. Despite these advancements, the quest for a swifter, more stable magneto-resistive random-access memory paradigm persists. In this vein, we present a groundbreaking development: room-temperature antiferromagnetic tunnel junctions devoid of any net magnetic moment. Over 200% tunneling altermagnetic resistance (TAR) ratio was measured at RuO2 (110)/MgO/RuO2 (110)/W structure, which is achieved by changing the antiferromagnetic Neel vector of RuO2 with an ultralow current density 2 MA*cm-2., Comment: Modification required

Collinear antiferromagnetic (AFM) materials have unique promise of no stray fields, display ultrafast dynamics, and being robust against perturbation filed which motivates the extensive research of antiferromagnetic spintronics. However, the manipulation and detection of antiferromagnetic order remain formidable challenges. Here, we report the electrical detection of colinear antiferromagnetism in all-epitaxial RuO2/MgO/RuO2 three-terminal tunnel junctions (TJ) using spin-flop tunnel anisotropy magnetoresistance (TAMR). We measured a TAMR ratio of around 60% at room temperature, which arises between the parallel and perpendicular configurations of the adjacent collinear AFM state. Furthermore, we carried out angular dependent measurements using this AFM-TJ and showed that the magnitude of anisotropic longitudinal magnetoresistance in the AFM-TJ can be controlled by the direction of magnetic field. We also theoretically found that the colinear antiferromagnetic MTJ may produce a substantially large TAMR ratio as a result of the time-reversal, strong spin orbit coupling (SOC) characteristic of antiferromagnetic RuO2. Our work not only propels antiferromagnetic materials to the forefront of spintronic device innovation but also unveils a novel paradigm for electrically governed antiferromagnetic spintronics, auguring transformative advancements in high-speed, low-energy information devices.

We find explicit and optimal upper bounds for the expected occupation density for an It\^o-process when its drift and diffusion coefficients are unknown under boundedness and ellipticity conditions on the coefficients. This is related to the optimal bound for the expected interval occupation found in Ankirchner and Wendt(2021). In contrast, our bound is for a single point and the resulting formula is less involved. Our findings allow us to find explicit upper bounds for mean path integrals., Comment: 20 pages

We investigate the orbit-to-charge conversion in YIG/Pt/nonmagnetic material (NM) trilayer heterostructures. With the additional Ru layer on the top of YIG/Pt stacks, the charge current signal increases nearly an order of magnitude in both longitudinal spin Seebeck effect (SSE) and spin pumping (SP) measurements. Through thickness dependence studies of the Ru metal layer and theoretical model, we quantitatively clarify different contributions of the increased SSE signal that mainly comes from the inverse orbital Hall effect (IOHE) of Ru, and partially comes from the orbital sink effect in the Ru layer. A similar enhancement of SSE(SP) signals is also observed when Ru is replaced by other materials (Ta, W, and Cu), implying the universality of the IOHE in transition metals. Our findings not only suggest a more efficient generation of the charge current via the orbital angular moment channel but also provides crucial insights into the interplay among charge, spin, and orbit.

This paper is concerned with finite dimensional models for the entire term structure for energy futures. As soon as a finite dimensional set of possible yield curves is chosen, one likes to estimate the dynamic behaviour of the yield curve evolution from data. The estimated model should be free of arbitrage which is known to result in some drift condition. If the yield curve evolution is modelled by a diffusion, then this leaves the diffusion coefficient open for estimation. From a practical perspective, this requires that the chosen set of possible yield curves is compatible with any obtained diffusion coefficient. In this paper, we show that this compatibility enforces an affine geometry of the set of possible yield curves., Comment: 17 pages

We find explicit upper bounds for the density of marginals of continuous diffusions where we assume that the diffusion coefficient is constant and the drift is solely assumed to be progressively measurable and locally bounded. In one dimension we extend our result to the case that the diffusion coefficient is a locally Lipschitz-continuous function of the state. Our approach is based on a comparison to a suitable doubly reflected Brownian motion whose density is known in a series representation., Comment: 13 pages

In transport, the topological Hall effect (THE) presents itself as non-monotonic features (or humps and dips) in the Hall signal and is widely interpreted as a sign of chiral spin textures, like magnetic skyrmions. However, when anomalous Hall effect (AHE) is also present, the co-existence of two AHEs could give rise to similar artifacts, making it difficult to distinguish between genuine THE with AHE and two-component AHE. Here we confirm genuine THE with AHE by means of transport and magneto-optical Kerr effect (MOKE) microscopy, in which magnetic skyrmions are directly observed, and find that genuine THE occurs in the transition region of the AHE. In sharp contrast, the artifact "THE", or two-component AHE occurs well beyond the saturation of the "AHE component" (under the false assumption of THE+AHE). Furthermore, we distinguish artifact "THE" from genuine THE by three methods: 1. Minor loops, 2. Temperature dependence, 3. Gate dependence. Minor loops of genuine THE with AHE are always within the full loop, while minor loops of the artifact "THE" may reveal a single loop that cannot fit into the "AHE component". Besides, the temperature or gate dependence of the artifact "THE" may also be accompanied by a polarity change of the "AHE component", as the non-monotonic features vanish, while the temperature dependence of genuine THE with AHE reveals no such change. Our work may help future researchers to exercise cautions and use these methods to examine carefully in order to ascertain genuine THE.

In this paper, we develop a novel weighted Laplacian method, which is partially inspired by the theory of graph Laplacian, to study recent popular graph problems, such as multilevel graph partitioning and balanced minimum cut problem, in a more convenient manner. Since the weighted Laplacian strategy inherits the virtues of spectral methods, graph algorithms designed using weighted Laplacian will necessarily possess more robust theoretical guarantees for algorithmic performances, comparing with those existing algorithms that are heuristically proposed. In order to illustrate its powerful utility both in theory and in practice, we also present two effective applications of our weighted Laplacian method to multilevel graph partitioning and balanced minimum cut problem, respectively. By means of variational methods and theory of partial differential equations (PDEs), we have established the equivalence relations among the weighted cut problem, balanced minimum cut problem and the initial clustering problem that arises in the middle stage of graph partitioning algorithms under a multilevel structure. These equivalence relations can indeed provide solid theoretical support for algorithms based on our proposed weighted Laplacian strategy. Moreover, from the perspective of the application to the balanced minimum cut problem, weighted Laplacian can make it possible for research of numerical solutions of PDEs to be a powerful tool for the algorithmic study of graph problems. Experimental results also indicate that the algorithm embedded with our strategy indeed outperforms other existing graph algorithms, especially in terms of accuracy, thus verifying the efficacy of the proposed weighted Laplacian.