arXiv subjects

Wei Guo

Publications and source records attributed to Wei Guo.

At least 163 records · Page 9Linked to original sources

arXiv2025

Optimizing Sequential Recommendation Models with Scaling Laws and Approximate Entropy

Scaling Laws have emerged as a powerful framework for understanding how model performance evolves as they increase in size, providing valuable insights for optimizing computational resources. In the realm of Sequential Recommendation (SR), which is pivotal for predicting users' sequential preferences, these laws offer a lens through which to address the challenges posed by the scalability of SR models. However, the presence of structural and collaborative issues in recommender systems prevents the direct application of the Scaling Law (SL) in these systems. In response, we introduce the Performance Law for SR models, which aims to theoretically investigate and model the relationship between model performance and data quality. Specifically, we first fit the HR and NDCG metrics to transformer-based SR models. Subsequently, we propose Approximate Entropy (ApEn) to assess data quality, presenting a more nuanced approach compared to traditional data quantity metrics. Our method enables accurate predictions across various dataset scales and model sizes, demonstrating a strong correlation in large SR models and offering insights into achieving optimal performance for any given model configuration.

cs.AI↗

arXiv2025

Generative Large Recommendation Models: Emerging Trends in LLMs for Recommendation

In the era of information overload, recommendation systems play a pivotal role in filtering data and delivering personalized content. Recent advancements in feature interaction and user behavior modeling have significantly enhanced the recall and ranking processes of these systems. With the rise of large language models (LLMs), new opportunities have emerged to further improve recommendation systems. This tutorial explores two primary approaches for integrating LLMs: LLMs-enhanced recommendations, which leverage the reasoning capabilities of general LLMs, and generative large recommendation models, which focus on scaling and sophistication. While the former has been extensively covered in existing literature, the latter remains underexplored. This tutorial aims to fill this gap by providing a comprehensive overview of generative large recommendation models, including their recent advancements, challenges, and potential research directions. Key topics include data quality, scaling laws, user behavior mining, and efficiency in training and inference. By engaging with this tutorial, participants will gain insights into the latest developments and future opportunities in the field, aiding both academic research and practical applications. The timely nature of this exploration supports the rapid evolution of recommendation systems, offering valuable guidance for researchers and practitioners alike.

cs.IR↗

arXiv2025

Provable Benefit of Annealed Langevin Monte Carlo for Non-log-concave Sampling

We consider the outstanding problem of sampling from an unnormalized density that may be non-log-concave and multimodal. To enhance the performance of simple Markov chain Monte Carlo (MCMC) methods, techniques of annealing type have been widely used. However, quantitative theoretical guarantees of these techniques are under-explored. This study takes a first step toward providing a non-asymptotic analysis of annealed MCMC. Specifically, we establish, for the first time, an oracle complexity of $\widetilde{O}\left(\frac{dβ^2{\cal A}^2}{\varepsilon^6}\right)$ for the simple annealed Langevin Monte Carlo algorithm to achieve $\varepsilon^2$ accuracy in Kullback-Leibler divergence to the target distribution $π\propto{\rm e}^{-V}$ on $\mathbb{R}^d$ with $β$-smooth potential $V$. Here, ${\cal A}$ represents the action of a curve of probability measures interpolating the target distribution $π$ and a readily sampleable distribution.

stat.ML↗

arXiv2025

Multi-granularity Interest Retrieval and Refinement Network for Long-Term User Behavior Modeling in CTR Prediction

Click-through Rate (CTR) prediction is crucial for online personalization platforms. Recent advancements have shown that modeling rich user behaviors can significantly improve the performance of CTR prediction. Current long-term user behavior modeling algorithms predominantly follow two cascading stages. The first stage retrieves subsequence related to the target item from the long-term behavior sequence, while the second stage models the relationship between the subsequence and the target item. Despite significant progress, these methods have two critical flaws. First, the retrieval query typically includes only target item information, limiting the ability to capture the user's diverse interests. Second, relational information, such as sequential and interactive information within the subsequence, is frequently overlooked. Therefore, it requires to be further mined to more accurately model user interests. To this end, we propose Multi-granularity Interest Retrieval and Refinement Network (MIRRN). Specifically, we first construct queries based on behaviors observed at different time scales to obtain subsequences, each capturing users' interest at various granularities. We then introduce an noval multi-head Fourier transformer to efficiently learn sequential and interactive information within the subsequences, leading to more accurate modeling of user interests. Finally, we employ multi-head target attention to adaptively assess the impact of these multi-granularity interests on the target item. Extensive experiments have demonstrated that MIRRN significantly outperforms state-of-the-art baselines. Furthermore, an A/B test shows that MIRRN increases the average number of listening songs by 1.32% and the average time of listening songs by 0.55% on the Huawei Music App. The implementation code is publicly available at https://github.com/USTC-StarTeam/MIRRN.

cs.IR↗

arXiv2025

Organometallic-Inorganic Hybrid MXenes with Tunable Superconductivity

Ti-based two-dimensional transition-metal carbides (MXenes) have attracted attention due to their superior properties and are being explored across various applications1,2. Despite their versatile properties, superconductivity has never been demonstrated, not even predicted, for this important group of 2D materials. In this work, we have introduced an electrochemical intercalation protocol to construct versatile organometallic-inorganic hybrid MXenes and achieved tunable superconductivity in the metallocene-modified layered crystals. Through structural editing of MXene matrix at atomic scale and meticulously modulated intercalation route, Ti3C2Tx intercalated with metallocene species exhibits a superconductive transition temperature (Tc) of 10.2 K. Guest intercalation induced electron filling and strain engineering are responsible for the emerging superconductivity in this intrinsically non-superconducting material. Theoretically, simulated electron-phonon interaction effects further elucidate the nature of the changes in Tc. Furthermore, the Tc of crafted artificial superlattices beyond Ti-based MXenes have been predicted, offering a general strategy for engineering superconductivity and magnetism in layered hybrid materials.

cond-mat.supr-con↗

arXiv2025

Equivalence of Noether charge and Hilbert action boundary term formulas for the black hole entropy in $F(R_{abcd})$ gravity theory

By working with the covariant phase space formalism, we have shown that not only can the Hamiltonian conjugate to a Killing vector field $ξ$ be expressed as the sum of the associated Noether charge and $ξ$ contracted with the Hilbert action boundary term for $F(R_{abcd})$ gravity, but also be written as its contraction with another $ξ$ independent tensor field. With this, we have proven the equivalence of Noether charge and Hilbert action boundary term formulae for the stationary black hole entropy in $F(R_{abcd})$ gravity, which is further substantiated by our explicit computation using both formulae.

hep-th↗

arXiv2025

FuXi-$α$: Scaling Recommendation Model with Feature Interaction Enhanced Transformer

Inspired by scaling laws and large language models, research on large-scale recommendation models has gained significant attention. Recent advancements have shown that expanding sequential recommendation models to large-scale recommendation models can be an effective strategy. Current state-of-the-art sequential recommendation models primarily use self-attention mechanisms for explicit feature interactions among items, while implicit interactions are managed through Feed-Forward Networks (FFNs). However, these models often inadequately integrate temporal and positional information, either by adding them to attention weights or by blending them with latent representations, which limits their expressive power. A recent model, HSTU, further reduces the focus on implicit feature interactions, constraining its performance. We propose a new model called FuXi-$α$ to address these issues. This model introduces an Adaptive Multi-channel Self-attention mechanism that distinctly models temporal, positional, and semantic features, along with a Multi-stage FFN to enhance implicit feature interactions. Our offline experiments demonstrate that our model outperforms existing models, with its performance continuously improving as the model size increases. Additionally, we conducted an online A/B test within the Huawei Music app, which showed a $4.76\%$ increase in the average number of songs played per user and a $5.10\%$ increase in the average listening duration per user. Our code has been released at https://github.com/USTC-StarTeam/FuXi-alpha.

cs.IR↗

arXiv2025

A globally divergence-free entropy stable nodal DG method for conservative ideal MHD equations

We propose an arbitrarily high-order globally divergence-free entropy stable nodal discontinuous Galerkin (DG) method to directly solve the conservative form of the ideal MHD equations using appropriate quadrature rules. The method ensures a globally divergence-free magnetic field by updating it at interfaces with a constraint-preserving formulation [5] and employing a novel least-squares reconstruction technique. Leveraging this property, the semi-discrete nodal DG scheme is proven to be entropy stable. To handle the problems with strong shocks, we introduce a novel limiting strategy that suppresses unphysical oscillations while preserving the globally divergence-free property. Numerical experiments verify the accuracy and efficacy of our method.

math.NA↗

arXiv2024

Image Classification with Deep Reinforcement Active Learning

Deep learning is currently reaching outstanding performances on different tasks, including image classification, especially when using large neural networks. The success of these models is tributary to the availability of large collections of labeled training data. In many real-world scenarios, labeled data are scarce, and their hand-labeling is time, effort and cost demanding. Active learning is an alternative paradigm that mitigates the effort in hand-labeling data, where only a small fraction is iteratively selected from a large pool of unlabeled data, and annotated by an expert (a.k.a oracle), and eventually used to update the learning models. However, existing active learning solutions are dependent on handcrafted strategies that may fail in highly variable learning environments (datasets, scenarios, etc). In this work, we devise an adaptive active learning method based on Markov Decision Process (MDP). Our framework leverages deep reinforcement learning and active learning together with a Deep Deterministic Policy Gradient (DDPG) in order to dynamically adapt sample selection strategies to the oracle's feedback and the learning environment. Extensive experiments conducted on three different image classification benchmarks show superior performances against several existing active learning strategies.

cs.CV↗

arXiv2024

Quantum Electronics on Quantum Liquids and Solids

Nonpolar atoms or molecules with low particle mass and weak inter-particle interactions can form quantum liquids and solids (QLS) at low temperatures. Excess electrons naturally bind to the surfaces of QLS in a vacuum, exhibiting unique quantum electronic behaviors in two and lower dimensions. This article reviews the historical development and recent progress in this field. Key topics include collective and individual electron transport on liquid helium, solid neon, and solid hydrogen; theoretical proposals and experimental efforts toward single-electron qubits on superfluid helium; the recent experimental realization of single-electron charge qubits on solid neon; and related theoretical calculations. Finally, we discuss and envision future exploration of quantum electronics in heterogeneous QLS systems.

quant-ph↗

arXiv2024

Scaling New Frontiers: Insights into Large Recommendation Models

Recommendation systems are essential for filtering data and retrieving relevant information across various applications. Recent advancements have seen these systems incorporate increasingly large embedding tables, scaling up to tens of terabytes for industrial use. However, the expansion of network parameters in traditional recommendation models has plateaued at tens of millions, limiting further benefits from increased embedding parameters. Inspired by the success of large language models (LLMs), a new approach has emerged that scales network parameters using innovative structures, enabling continued performance improvements. A significant development in this area is Meta's generative recommendation model HSTU, which illustrates the scaling laws of recommendation systems by expanding parameters to thousands of billions. This new paradigm has achieved substantial performance gains in online experiments. In this paper, we aim to enhance the understanding of scaling laws by conducting comprehensive evaluations of large recommendation models. Firstly, we investigate the scaling laws across different backbone architectures of the large recommendation models. Secondly, we conduct comprehensive ablation studies to explore the origins of these scaling laws. We then further assess the performance of HSTU, as the representative of large recommendation models, on complex user behavior modeling tasks to evaluate its applicability. Notably, we also analyze its effectiveness in ranking tasks for the first time. Finally, we offer insights into future directions for large recommendation models. Supplementary materials for our research are available on GitHub at https://github.com/USTC-StarTeam/Large-Recommendation-Models.

cs.IR↗

arXiv2024

Controlled angular momentum injection in a magnetically levitated He II droplet

The morphology of rotating viscous classical liquid droplets has been extensively studied and is well understood. However, our understanding of rotating superfluid droplets remains limited. For instance, superfluid $^4$He (He II) can carry angular momentum through two distinct mechanisms: the formation of an array of quantized vortex lines, which induce flows resembling classical solid-body rotation, and surface traveling deformation modes associated with irrotational internal flows. These two mechanisms can result in significantly different droplet morphologies, and it remains unclear how the injected angular momentum is partitioned between them. To investigate this complex problem experimentally, one must first levitate an isolated He II droplet using techniques such as magnetic levitation. However, an outstanding challenge lies in effectively injecting angular momentum into the levitated droplet. In this paper, we describe a magneto-optical cryostat system designed to levitate He II droplets and present the design of a time-dependent, non-axially symmetric electric driving system. Based on our numerical simulations, this system should enable controlled angular momentum injection into the droplet. This study lays the foundation for future investigations into the morphology of rotating He II droplets.

cond-mat.other↗

arXiv2024

Advances in understanding vacuum break dynamics in liquid helium-cooled tubes for accelerator beamline applications

Understanding air propagation and condensation following a catastrophic vacuum break in particle accelerator beamlines cooled by liquid helium is essential for ensuring operational safety. This review summarizes experimental and theoretical work conducted in our cryogenics lab to address this issue. Systematic measurements were performed to study nitrogen gas propagation in uniform copper tubes cooled by both normal liquid helium (He I) and superfluid helium (He II). These experiments revealed a nearly exponential deceleration of the gas front, with stronger deceleration observed in He II-cooled tubes. To interpret these results, a one-dimensional (1D) theoretical model was developed, incorporating gas dynamics, heat transfer, and condensation mechanisms. The model successfully reproduced key experimental observations in the uniform tube system. However, recent experiments involving a bulky copper cavity designed to mimic the geometry of a superconducting radio-frequency (SRF) cavity revealed strong anisotropic flow patterns of nitrogen gas within the cavity, highlighting limitations in extrapolating results from simplified tube geometries to real accelerator beamlines. To address these complexities, we outline plans for systematic studies using tubes with multiple bulky cavities and the development of a two-dimensional (2D) model to simulate gas dynamics in these more intricate configurations. These efforts aim to provide a comprehensive understanding of vacuum breaks in particle accelerators and improve predictive capabilities for their operational safety.

physics.acc-ph↗

arXiv2024

Notes on solution phase space and BTZ black hole

In this paper, we use the solution phase space approach based on the covariant phase space formalism to compute the conserved charges of the BTZ black hole, namely mass, angular momentum, and entropy. Furthermore, we discuss the first law of the BTZ black hole and the Smarr relation. For completeness, outer horizon and inner horizon cases have been all included. Additionally, the results of the three-dimensional Kerr-dS spacetime have also been obtained. Our results are consistent with previous investigations. Considering the simplicity of the circumstances, we have presented the most detailed possible information, with the aim of facilitating research in related fields.

gr-qc↗

arXiv2024

High-throughput 3D shape completion of potato tubers on a harvester

Potato yield is an important metric for farmers to further optimize their cultivation practices. Potato yield can be estimated on a harvester using an RGB-D camera that can estimate the three-dimensional (3D) volume of individual potato tubers. A challenge, however, is that the 3D shape derived from RGB-D images is only partially completed, underestimating the actual volume. To address this issue, we developed a 3D shape completion network, called CoRe++, which can complete the 3D shape from RGB-D images. CoRe++ is a deep learning network that consists of a convolutional encoder and a decoder. The encoder compresses RGB-D images into latent vectors that are used by the decoder to complete the 3D shape using the deep signed distance field network (DeepSDF). To evaluate our CoRe++ network, we collected partial and complete 3D point clouds of 339 potato tubers on an operational harvester in Japan. On the 1425 RGB-D images in the test set (representing 51 unique potato tubers), our network achieved a completion accuracy of 2.8 mm on average. For volumetric estimation, the root mean squared error (RMSE) was 22.6 ml, and this was better than the RMSE of the linear regression (31.1 ml) and the base model (36.9 ml). We found that the RMSE can be further reduced to 18.2 ml when performing the 3D shape completion in the center of the RGB-D image. With an average 3D shape completion time of 10 milliseconds per tuber, we can conclude that CoRe++ is both fast and accurate enough to be implemented on an operational harvester for high-throughput potato yield estimation. CoRe++'s high-throughput and accurate processing allows it to be applied to other tuber, fruit and vegetable crops, thereby enabling versatile, accurate and real-time yield monitoring in precision agriculture. Our code, network weights and dataset are publicly available at https://github.com/UTokyo-FieldPhenomics-Lab/corepp.git.

cs.CV↗

arXiv2024

Particle Levitation Velocimetry for boundary layer measurements in high Reynolds number liquid helium turbulence

Understanding boundary layer flows in high Reynolds number (Re) turbulence is crucial for advancing fluid dynamics in a wide range of applications, from improving aerodynamic efficiency in aviation to optimizing energy systems in industrial processes. However, generating such flows requires complex, power-intensive large-scale facilities. Furthermore, the use of local probes, such as hot wires and pressure sensors, often introduces disturbances due to the necessary support structures, compromising measurement accuracy. In this paper, we present a solution that leverages the vanishingly small viscosity of liquid helium to produce high Re flows, combined with an innovative Particle Levitation Velocimetry (PLV) system for precise flow-field measurements. This PLV system uses magnetically levitated superconducting micro-particles to measure the near-wall velocity field in liquid helium. Through comprehensive theoretical analysis, we demonstrate that the PLV system enables quantitative measurements of the velocity boundary layer over a wall unit range of $44\le y^{+}\le 4400$, with a spatial resolution that, depending on the particle size, can reach down to about 10~$μ$m. This development opens new avenues for exploring turbulence structures and correlations within the thin boundary layer that would be otherwise difficult to achieve.

physics.flu-dyn↗

arXiv2024

Enhancing CTR Prediction in Recommendation Domain with Search Query Representation

Many platforms, such as e-commerce websites, offer both search and recommendation services simultaneously to better meet users' diverse needs. Recommendation services suggest items based on user preferences, while search services allow users to search for items before providing recommendations. Since users and items are often shared between the search and recommendation domains, there is a valuable opportunity to enhance the recommendation domain by leveraging user preferences extracted from the search domain. Existing approaches either overlook the shift in user intention between these domains or fail to capture the significant impact of learning from users' search queries on understanding their interests. In this paper, we propose a framework that learns from user search query embeddings within the context of user preferences in the recommendation domain. Specifically, user search query sequences from the search domain are used to predict the items users will click at the next time point in the recommendation domain. Additionally, the relationship between queries and items is explored through contrastive learning. To address issues of data sparsity, the diffusion model is incorporated to infer positive items the user will select after searching with certain queries in a denoising manner, which is particularly effective in preventing false positives. Effectively extracting this information, the queries are integrated into click-through rate prediction in the recommendation domain. Experimental analysis demonstrates that our model outperforms state-of-the-art models in the recommendation domain.

cs.IR↗

arXiv2024

Plug-and-Play Controllable Generation for Discrete Masked Models

This article makes discrete masked models for the generative modeling of discrete data controllable. The goal is to generate samples of a discrete random variable that adheres to a posterior distribution, satisfies specific constraints, or optimizes a reward function. This methodological development enables broad applications across downstream tasks such as class-specific image generation and protein design. Existing approaches for controllable generation of masked models typically rely on task-specific fine-tuning or additional modifications, which can be inefficient and resource-intensive. To overcome these limitations, we propose a novel plug-and-play framework based on importance sampling that bypasses the need for training a conditional score. Our framework is agnostic to the choice of control criteria, requires no gradient information, and is well-suited for tasks such as posterior sampling, Bayesian inverse problems, and constrained generation. We demonstrate the effectiveness of our approach through extensive experiments, showcasing its versatility across multiple domains, including protein design.

cs.LG↗