arXiv subjects

Lei Wang

Publications and source records attributed to Lei Wang.

At least 631 records · Page 35Linked to original sources

arXiv2024

The Whole is Better than the Sum: Using Aggregated Demonstrations in In-Context Learning for Sequential Recommendation

Large language models (LLMs) have shown excellent performance on various NLP tasks. To use LLMs as strong sequential recommenders, we explore the in-context learning approach to sequential recommendation. We investigate the effects of instruction format, task consistency, demonstration selection, and number of demonstrations. As increasing the number of demonstrations in ICL does not improve accuracy despite using a long prompt, we propose a novel method called LLMSRec-Syn that incorporates multiple demonstration users into one aggregated demonstration. Our experiments on three recommendation datasets show that LLMSRec-Syn outperforms state-of-the-art LLM-based sequential recommendation methods. In some cases, LLMSRec-Syn can perform on par with or even better than supervised learning methods. Our code is publicly available at https://github.com/demoleiwang/LLMSRec_Syn.

cs.IR↗

arXiv2024

Gradient-Aware Logit Adjustment Loss for Long-tailed Classifier

In the real-world setting, data often follows a long-tailed distribution, where head classes contain significantly more training samples than tail classes. Consequently, models trained on such data tend to be biased toward head classes. The medium of this bias is imbalanced gradients, which include not only the ratio of scale between positive and negative gradients but also imbalanced gradients from different negative classes. Therefore, we propose the Gradient-Aware Logit Adjustment (GALA) loss, which adjusts the logits based on accumulated gradients to balance the optimization process. Additionally, We find that most of the solutions to long-tailed problems are still biased towards head classes in the end, and we propose a simple and post hoc prediction re-balancing strategy to further mitigate the basis toward head class. Extensive experiments are conducted on multiple popular long-tailed recognition benchmark datasets to evaluate the effectiveness of these two designs. Our approach achieves top-1 accuracy of 48.5\%, 41.4\%, and 73.3\% on CIFAR100-LT, Places-LT, and iNaturalist, outperforming the state-of-the-art method GCL by a significant margin of 3.62\%, 0.76\% and 1.2\%, respectively. Code is available at https://github.com/lt-project-repository/lt-project.

cs.CV↗

arXiv2024

Chaotic Masking Protocol for Secure Communication and Attack Detection in Remote Estimation of Cyber-Physical Systems

In remote estimation of cyber-physical systems (CPSs), sensor measurements transmitted through network may be attacked by adversaries, leading to leakage risk of privacy (e.g., the system state), and/or failure of the remote estimator. To deal with this problem, a chaotic masking protocol is proposed in this paper to secure the sensor measurements transmission. In detail, at the plant side, a chaotic dynamic system is deployed to encode the sensor measurement, and at the estimator side, an estimator estimates both states of the physical plant and the chaotic system. With this protocol, no additional secure communication links is needed for synchronization, and the masking effect can be perfectly removed when the estimator is in steady state. Furthermore, this masking protocol can deal with multiple types of attacks, i.e., eavesdropping attack, replay attack, and stealthy false data injection attack.

eess.SY↗

arXiv2024

Distributed Least-Squares Optimization Solvers with Differential Privacy

This paper studies the distributed least-squares optimization problem with differential privacy requirement of local cost functions, for which two differentially private distributed solvers are proposed. The first is established on the distributed gradient tracking algorithm, by appropriately perturbing the initial values and parameters that contain the privacy-sensitive data with Gaussian and truncated Laplacian noises, respectively. Rigorous proofs are established to show the achievable trade-off between the (ε, δ)-differential privacy and the computation accuracy. The second solver is established on the combination of the distributed shuffling mechanism and the average consensus algorithm, which enables each agent to obtain a noisy version of parameters characterizing the global gradient. As a result, the least-squares optimization problem can be eventually solved by each agent locally in such a way that any given (ε, δ)-differential privacy requirement can be preserved while the solution may be computed with the accuracy independent of the network size, which makes the latter more suitable for large-scale distributed least-squares problems. Numerical simulations are presented to show the effectiveness of both solvers.

math.OC↗

arXiv2024

Ions-induced Epitaxial Growth of Perovskite Nanocomposites for Highly Efficient Light-Emitting Diodes with EQE Exceeding 30%

Metal halide perovskites, a class of cost-effective semiconductor materials, are of great interest for modern and upcoming display technologies that prioritize the light-emitting diodes (LEDs) with high efficiency and excellent color purity. The prevailing approach to achieving efficient luminescence from pervoskites is enhancing exciton binding effect and confining carriers by reducing their dimensionality or grain size. However, splitting pervoskite lattice into smaller ones generates abundant boundaries in solid films and results in more surface trap states, needing exact passivation to suppress trap-assisted nonradiative losses. Here, an ions-induced heteroepitaxial growth method is employed to assembe perovskite lattices with different structures into large-sized grains to produce lattice-anchored nanocomposites for efficient LEDs with high color purity. This approach enables the nanocomposite thin films, composed of three-dimensional (3D) CsPbBr3 and its variant of zero-dimensional (0D) Cs4PbBr6, to feature significant low trap-assisted nonradiative recombination, enhanced light out-coupling with a corrugated surface, and well-balanced charge carrier transport. Based on the resultant 3D/0D perovskite nanocomposites, we demonstrate the perovskite LEDs achieving an remarkable external quantum efficiency of 31.0% at the emission peak of 521 nm with a narrow full width at half-maximum of only 18 nm. This research introduces a novel approach to the development of well-assembled nanocomposites for perovskite LEDs, demonstrating high efficiency comparable to that of state-of-the-art organic LEDs.

cond-mat.mtrl-sci↗

arXiv2024

Revisiting wrong sign Yukawa coupling of type II two-Higgs-doublet model in light of the recent LHC data

In light of the recent LHC Higgs data, we examine the parameter space of type II two-Higgs-doublet model in which the 125 GeV Higgs has the wrong sign Yukawa couplings. Combining related theoretical and experimental limits, we find that the LHC Higgs data exclude most of the parameter space of the wrong sign Yukawa coupling. For $m_H=$ 600 GeV, the allowed samples are mainly distributed in several corners and narrow bands of m_A<20 GeV, 30 GeV<m_A<120 GeV, 240 GeV<m_A<300 GeV, 380 GeV <m_A<430 GeV, and 480 GeV<m_A<550 GeV. For m_A=600 GeV, m_H is required to be less than 470 GeV. The light pseudo-scalar with a mass of 20 GeV is still allowed in case of the wrong sign Yukawa coupling of 125 GeV Higgs.

hep-ph↗

arXiv2024

The lepton flavor violating decays $Z\to l_i l_j$ in the simplest little Higgs model

In the simplest little Higgs model the new flavor-changing interactions between heavy neutrinos and the Standard Model leptons can generate contributions to some lepton flavor violating decays of $Z$-boson at one-loop level, such as $Z \to τ^{\pm}μ^{\mp}$, $Z\to τ^{\pm}e^{\mp}$, and $Z \to μ^{\pm}e^{\mp}$. We examine the decay modes, and find that the branching ratios can reach $10^{-7}$ for the three decays, which should be accessible at the Giga$Z$ option of the ILC.

hep-ph↗

arXiv2024

Analyzing and Reducing Catastrophic Forgetting in Parameter Efficient Tuning

Existing research has shown that large language models (LLMs) exhibit remarkable performance in language understanding and generation. However, when LLMs are continuously fine-tuned on complex and diverse domain-specific downstream tasks, the inference performance on historical tasks decreases dramatically, which is known as a catastrophic forgetting problem. A trade-off needs to be kept between learning plasticity and memory stability. Plenty of existing works have explored strategies like memory replay, regularization and parameter isolation, but little is known about the geometric connection of various adjacent minima in the continual LLMs fine-tuning scenarios. In this work, we investigate the geometric connections of different minima through the lens of mode connectivity, which means different minima can be connected by a low-loss valley. Through extensive experiments, we uncover the mode connectivity phenomenon in the LLMs continual learning scenario and find that it can strike a balance between plasticity and stability. Building upon these findings, we propose a simple yet effective method called Interpolation-based LoRA (I-LoRA), which constructs a dual-memory experience replay framework based on LoRA parameter interpolations. Extensive experiments and analysis on eight domain-specific CL benchmarks demonstrate that I-LoRA consistently show significant improvement over the previous state-of-the-art approaches with up to $11\%$ performance gains, providing a strong baseline and insights for future research on the large language model continual learning problem. Our code is available at \url{https://github.com/which47/LLMCL}.

cs.LG↗

arXiv2024

Global well-posedness and long time behavior of 2D MHD equations with partial dissipation in half space

In this paper, we obtain the low order global well-posedness and the asymptotic behavior of solution of 2D MHD problem with partial dissipation in half space with non-slip boundary condition. When magnetic field equal zero, the system be reduced to partial dissipation Navier-Stokes equation, so this result also implies the stabilizing effects of magnetic field in electrically conducting fluids. We use the resolvent estimate method to obtain the long time behavior for the solution of weak diffusion system, which is not necessary to prove global well-posedness.

math.AP↗

arXiv2024

How to coadd images: II. Anti-aliasing and PSF deconvolution

We have developed a novel method for co-adding multiple under-sampled images that combines the iteratively reweighted least squares and divide-and-conquer algorithms. Our approach not only allows for the anti-aliasing of the images but also enables PSF deconvolution, resulting in enhanced restoration of extended sources, the highest PSNR, and reduced ringing artefacts. To test our method, we conducted numerical simulations that replicated observation runs of the CSST/VST telescope and compared our results to those obtained using previous algorithms. The simulation showed that our method outperforms previous approaches in several ways, such as restoring the profile of extended sources and minimizing ringing artefacts. Additionally, because our method relies on the inherent advantages of least squares fitting, it is more versatile and does not depend on the local uniformity hypothesis for the PSF. However, the new method consumes much more computation than the other approaches.

astro-ph.IM↗

arXiv2024

PIMSYN: Synthesizing Processing-in-memory CNN Accelerators

Processing-in-memory architectures have been regarded as a promising solution for CNN acceleration. Existing PIM accelerator designs rely heavily on the experience of experts and require significant manual design overhead. Manual design cannot effectively optimize and explore architecture implementations. In this work, we develop an automatic framework PIMSYN for synthesizing PIM-based CNN accelerators, which greatly facilitates architecture design and helps generate energyefficient accelerators. PIMSYN can automatically transform CNN applications into execution workflows and hardware construction of PIM accelerators. To systematically optimize the architecture, we embed an architectural exploration flow into the synthesis framework, providing a more comprehensive design space. Experiments demonstrate that PIMSYN improves the power efficiency by several times compared with existing works. PIMSYN can be obtained from https://github.com/lixixi-jook/PIMSYN-NN.

cs.AR↗

arXiv2024

The Era of 1-bit LLMs: All Large Language Models are in 1.58 Bits

Recent research, such as BitNet, is paving the way for a new era of 1-bit Large Language Models (LLMs). In this work, we introduce a 1-bit LLM variant, namely BitNet b1.58, in which every single parameter (or weight) of the LLM is ternary {-1, 0, 1}. It matches the full-precision (i.e., FP16 or BF16) Transformer LLM with the same model size and training tokens in terms of both perplexity and end-task performance, while being significantly more cost-effective in terms of latency, memory, throughput, and energy consumption. More profoundly, the 1.58-bit LLM defines a new scaling law and recipe for training new generations of LLMs that are both high-performance and cost-effective. Furthermore, it enables a new computation paradigm and opens the door for designing specific hardware optimized for 1-bit LLMs.

cs.CL↗

arXiv2024

Quality at the Tail of Machine Learning Inference

Machine learning inference should be subject to stringent inference time constraints while ensuring high inference quality, especially in safety-critical (e.g., autonomous driving) and mission-critical (e.g., emotion recognition) contexts. Neglecting either aspect can lead to severe consequences, such as loss of life and property damage. Many studies lack a comprehensive consideration of these metrics, leading to incomplete or misleading evaluations. The study unveils a counterintuitive revelation: deep learning inference quality exhibits fluctuations due to inference time. To depict this phenomenon, the authors coin a new term, "tail quality," providing a more comprehensive evaluation, and overcoming conventional metric limitations. Moreover, the research proposes an initial evaluation framework to analyze factors affecting quality fluctuations, facilitating the prediction of the potential distribution of inference quality. The effectiveness of the evaluation framework is validated through experiments conducted on deep learning models for three different tasks across four systems.

cs.LG↗

arXiv2024

Design of a Teleoperated Robotic Bronchoscopy System for Peripheral Pulmonary Lesion Biopsy

Bronchoscopy with transbronchial biopsy is a minimally invasive and effective method for early lung cancer intervention. Robot-assisted bronchoscopy offers improved precision, spatial flexibility, and reduced risk of cross-infection. This paper introduces a novel teleoperated robotic bronchoscopy system and a three-stage procedure designed for robot-assisted bronchoscopy. The robotic mechanism enables a clinical practice similar to traditional bronchoscopy, augmented by the control of a novel variable stiffness catheter for tissue sampling. A rapid prototype of the robotic system has been fully developed and validated through in-vivo experiments. The results demonstrate the potential of the proposed robotic bronchoscopy system and variable stiffness catheter in enhancing accuracy and safety during bronchoscopy procedures.

physics.med-ph↗

arXiv2024

2D excitation information by MPS method on infinite helixes

Understanding the excitation spectrum in two-dimensional quantum many-body systems has long been a challenging task. We present an approach by introducing an excitation ansatz based on an infinite matrix product state (MPS) on a helix structure. With the canonical form of MPS states, we can accurately extract key properties such as energy, degeneracy, spectrum weight, and scaling behavior of low-energy excited states simultaneously. To validate the effectiveness of this method, we begin by applying it to the critical point of the transverse-field Ising model. The extracted scaling exponent of the energy gap closely aligns with the conformal bootstrap results. Subsequently, we apply this approach to the $J_1$-$J_2$ Heisenberg model on a square lattice. We discover that the degeneracy of lowest-energy excitations serves as a reliable metric for distinguishing different phases. The phase boundary identified by our method is consistent with some of the previous findings. The present method provides a promising avenue for studying the excitation spectrum of two-dimensional quantum many-body systems.

cond-mat.str-el↗

arXiv2024

Mastering the Game of Guandan with Deep Reinforcement Learning and Behavior Regulating

Games are a simplified model of reality and often serve as a favored platform for Artificial Intelligence (AI) research. Much of the research is concerned with game-playing agents and their decision making processes. The game of Guandan (literally, "throwing eggs") is a challenging game where even professional human players struggle to make the right decision at times. In this paper we propose a framework named GuanZero for AI agents to master this game using Monte-Carlo methods and deep neural networks. The main contribution of this paper is about regulating agents' behavior through a carefully designed neural network encoding scheme. We then demonstrate the effectiveness of the proposed framework by comparing it with state-of-the-art approaches.

cs.AI↗

arXiv2024

Locally Optimal Eigenvectors of Regular Simplex Tensors

Identifying locally optimal solutions is an important issue given an optimization model. In this paper, we focus on a special class of symmetric tensors termed regular simplex tensors, which is a newly-emerging concept, and investigate its local optimality of the related constrained nonconvex optimization model. This is proceeded by checking the first-order and second-order necessary condition sequentially. Some interesting directions concerning the regular simplex tensors, including the robust eigenpairs checking and other potential issues, are discussed in the end for future work.

math.OC↗

arXiv2024

Equivariant neural network for Green's functions of molecules and materials

The many-body Green's function provides access to electronic properties beyond density functional theory level in ab inito calculations. In this manuscript, we propose a deep learning framework for predicting the finite-temperature Green's function in atomic orbital space, aiming to achieve a balance between accuracy and efficiency. By predicting the self-energy matrices in Lehmann representation using an equivariant message passing neural network, our method respects its analytical property and the $E(3)$ equivariance. The Green's function is obtained from the predicted self-energy through Dyson equation with target total number of electrons. We present proof-of-concept benchmark results for both molecules and simple periodic systems, showing that our method is able to provide accurate estimate of physical observables such as energy and density of states based on the predicted Green's function.

physics.chem-ph↗