arXiv subjects
Publications and source records attributed to Ying Wang.
We evaluate the Hankel determinants of the convolution powers of Motzkin numbers for $r\leq 27$ by finding shifted periodic continued fractions, which arose in application of Sulanke and Xin's continued fraction method. We also conjecture some polynomial characterization of these determinants.
Large language models (LLMs) are widely used for natural language understanding and text generation. An LLM model relies on a time-consuming step called LLM decoding to generate output tokens. Several prior works focus on improving the performance of LLM decoding using parallelism techniques, such as batching and speculative decoding. State-of-the-art LLM decoding has both compute-bound and memory-bound kernels. Some prior works statically identify and map these different kernels to a heterogeneous architecture consisting of both processing-in-memory (PIM) units and computation-centric accelerators. We observe that characteristics of LLM decoding kernels (e.g., whether or not a kernel is memory-bound) can change dynamically due to parameter changes to meet user and/or system demands, making (1) static kernel mapping to PIM units and computation-centric accelerators suboptimal, and (2) one-size-fits-all approach of designing PIM units inefficient due to a large degree of heterogeneity even in memory-bound kernels. In this paper, we aim to accelerate LLM decoding while considering the dynamically changing characteristics of the kernels involved. We propose PAPI (PArallel Decoding with PIM), a PIM-enabled heterogeneous architecture that exploits dynamic scheduling of compute-bound or memory-bound kernels to suitable hardware units. PAPI has two key mechanisms: (1) online kernel characterization to dynamically schedule kernels to the most suitable hardware units at runtime and (2) a PIM-enabled heterogeneous computing system that harmoniously orchestrates both computation-centric processing units and hybrid PIM units with different computing capabilities. Our experimental results on three broadly-used LLMs show that PAPI achieves 1.8$\times$ and 11.1$\times$ speedups over a state-of-the-art heterogeneous LLM accelerator and a state-of-the-art PIM-only LLM accelerator, respectively.
Waveguide quantum electrodynamics (QED) has opened a new frontier in quantum optics, which enables the radiative coupling of distantly located emitters via the spatially extended waveguide mode. This coupling leads to modified emission dynamics and previous work has reported the observation of increased intensity correlations (an antidip) when probing the resonance response of multiple emitters. However, the interference between independent emitters has been shown to lead to a similar response. Here, we directly observe resonant energy transfer between two distant quantum emitters by recording an antidip in the intensity correlations, $g^{(2)}(τ)$, while driving only one of the emitters. Under the condition that only a single emitter is driven, the antidip in photon coincidences is a distinctive signature of emitter-emitter coupling, which enables the transfer of energy from the driven to the undriven emitter. Interestingly, the observed mechanism is a long-range and waveguide-engineered version of resonant Förster transfer, which is responsible for the transport of energy between chlorophylls in the photosynthesis. Building on the established coupling, we demonstrate collective driving of the coupled emitter pair. Specifically, we control the relative driving phase and amplitude of the emitters and apply this collective excitation scheme to selectively populate the long-lived subradiant state. This results in suppressed emission, i.e. the peculiar situation where driving two emitters as opposed to one effectively reduces the probability of photon emission. Our work presents novel emission regimes and excitation schemes for a multi-emitter waveguide QED system. These can be exploited to deterministically generate emitter-emitter entanglement and advanced photonic states providing robustness against losses for photonic quantum computation and quantum communication.
The billion-scale Large Language Models (LLMs) need deployment on expensive server-grade GPUs with large-storage HBMs and abundant computation capability. As LLM-assisted services become popular, achieving cost-effective LLM inference on budget-friendly hardware becomes the trend. Extensive researches relocate LLM parameters from expensive GPUs to host memory. However, the restricted bandwidth between the host and GPU memory limits the inference performance. This work introduces Hermes, a budget-friendly system that leverages the near-data processing (NDP) within commodity DRAM DIMMs to enhance the performance of a single consumer-grade GPU, achieving efficient LLM inference. The inherent activation sparsity in LLMs naturally divides weight parameters into two categories, termed ``hot" and ``cold" neurons, respectively. Hot neurons, which consist of only approximately 20\% of all weight parameters, account for 80\% of the total computational load, while cold neurons make up the other 80\% of parameters but are responsible for just 20\% of the computational load. Therefore, we propose a heterogeneous computing strategy: mapping hot neurons to a single computation-efficient GPU, while offloading cold neurons to NDP-DIMMs, which offer large memory size but limited computation capabilities. Meanwhile, the dynamic nature of activation sparsity needs a real-time partition of hot/cold neurons and adaptive remapping of cold neurons across multiple NDP-DIMM modules. Therefore, we introduce a lightweight predictor optimizing real-time neuron partition and adjustment between GPU and NDP-DIMMs. We also utilize a window-based online scheduling mechanism to maintain load balance among NDP-DIMM modules. Hermes facilitates the deployment of LLaMA2-70B on consumer-grade hardware at 13.75 tokens/s and realizes an average 75.24$\times$ speedup over the state-of-the-art offloading-based inference system.
Computing-in-memory (CIM) architectures demonstrate superior performance over traditional architectures. To unleash the potential of CIM accelerators, many compilation methods have been proposed, focusing on application scheduling optimization specific to CIM. However, existing compilation methods often overlook CIM's capability to switch dynamically between compute and memory modes, which is crucial for accommodating the diverse memory and computational needs of real-world deep neural network architectures, especially the emerging large language models. To fill this gap, we introduce CMSwitch, a novel compiler to optimize resource allocation for CIM accelerators with adaptive mode-switching capabilities, thereby enhancing the performance of DNN applications. Specifically, our approach integrates the compute-memory mode switch into the CIM compilation optimization space by introducing a new hardware abstraction attribute. Then, we propose a novel compilation optimization pass that identifies the optimal network segment and the corresponding mode resource allocations using dynamic programming and mixed-integer programming. CMSwitch uses the tailored meta-operator to express the compilation result in a generalized manner. Evaluation results demonstrate that CMSwitch achieves an average speedup of 1.31$\times$ compared to existing SOTA CIM compilation works, highlighting CMSwitch's effectiveness in fully exploiting the potential of CIM processors for a wide range of real-world DNN applications.
Attention-based methods have demonstrated exceptional performance in modelling long-range dependencies on spherical cortical surfaces, surpassing traditional Geometric Deep Learning (GDL) models. However, their extensive inference time and high memory demands pose challenges for application to large datasets with limited computing resources. Inspired by the state space model in computer vision, we introduce the attention-free Vision Mamba (Vim) to spherical surfaces, presenting a domain-agnostic architecture for analyzing data on spherical manifolds. Our method achieves surface patching by representing spherical data as a sequence of triangular patches derived from a subdivided icosphere. The proposed Surface Vision Mamba (SiM) is evaluated on multiple neurodevelopmental phenotype regression tasks using cortical surface metrics from neonatal brains. Experimental results demonstrate that SiM outperforms both attention- and GDL-based methods, delivering 4.8 times faster inference and achieving 91.7% lower memory consumption compared to the Surface Vision Transformer (SiT) under the Ico-4 grid partitioning. Sensitivity analysis further underscores the potential of SiM to identify subtle cognitive developmental patterns. The code is available at https://github.com/Rongzhao-He/surface-vision-mamba.
Accurate physical activity level (PAL) classification could be beneficial for osteoarthritis (OA) management. This study examines the impact of sensor placement and deep learning models on AL classification using the Metabolic Equivalent of Task values. The results show that the addition of anankle sensor (WA) significantly improves the classification of intensity activities compared to wrist-only configuration(53% to 86.2%). The CNN-LSTM model achieves the highest accuracy (95.09%). Statistical analysis confirms multi-sensor setups outperform single-sensor configurations (p < 0.05). The WA configuration offers a balance between usability and accuracy, making it a cost-effective solution for AL monitoring, particularly in OA management.
Dynamic Mode Decomposition (DMD) has received increasing research attention due to its capability to analyze and model complex dynamical systems. However, it faces challenges in computational efficiency, noise sensitivity, and difficulty adhering to physical laws, which negatively affect its performance. Addressing these issues, we present Online Physics-informed DMD (OPIDMD), a novel adaptation of DMD into a convex optimization framework. This approach not only ensures convergence to a unique global optimum, but also enhances the efficiency and accuracy of modeling dynamical systems in an online setting. Leveraging the Bayesian DMD framework, we propose a probabilistic interpretation of Physics-informed DMD (piDMD), examining the impact of physical constraints on the DMD linear operator. Further, we implement online proximal gradient descent and formulate specific algorithms to tackle problems with different physical constraints, enabling real-time solutions across various scenarios. Compared with existing algorithms such as Exact DMD, Online DMD, and piDMD, OPIDMD achieves the best prediction performance in short-term forecasting, e.g. an $R^2$ value of 0.991 for noisy Lorenz system. The proposed method employs a time-varying linear operator, offering a promising solution for the real-time simulation and control of complex dynamical systems.
The intersection of medical imaging and artificial intelligence has become an important research direction in intelligent medical treatment, particularly in the analysis of medical images using deep learning for clinical diagnosis. Despite the advances, existing keyframe classification methods lack extraction of time series features, while ultrasonic video classification based on three-dimensional convolution requires uniform frame numbers across patients, resulting in poor feature extraction efficiency and model classification performance. This study proposes a novel video classification method based on CNN and LSTM, introducing NLP's long and short sentence processing scheme into video classification for the first time. The method reduces CNN-extracted image features to 1x512 dimension, followed by sorting and compressing feature vectors for LSTM training. Specifically, feature vectors are sorted by patient video frame numbers and populated with padding value 0 to form variable batches, with invalid padding values compressed before LSTM training to conserve computing resources. Experimental results demonstrate that our variable-frame CNNLSTM method outperforms other approaches across all metrics, showing improvements of 3-6% in F1 score and 1.5% in specificity compared to keyframe methods. The variable-frame CNNLSTM also achieves better accuracy and precision than equal-frame CNNLSTM. These findings validate the effectiveness of our approach in classifying variable-frame ultrasound videos and suggest potential applications in other medical imaging modalities.
In this paper, our goal is to establish the Sobolev space associated to the partial harmonic oscillator. Based on its heat kernel estimate, we firstly give the definition of the fractional powers of the partial harmonic oscillator $$\AH=-\partial_ρ^2-Δ_x+|x|^2,$$ and show that its negative powers are well defined on $L^p(\mathbb R^{d+1})$ for $p\in [1,\infty]$. We then define associated Riesz transforms and show that they are bounded on classical Sobolev spaces by the calculus of symbols. Secondly, by a factorization of the operator $\AH$, we define two families of Sobolev spaces with positive integer indices, and show the equivalence between them by the boundedness of Riesz transforms. Moreover, the adapted symbolic calculus also implies the boundedness of Riesz type transforms on the Sobolev spaces associated to the partial harmonic oscillator $\AH$. Lastly, as applications of our results, we obtain the revised Hardy-Littlewood-Sobolev inequality, the Gagliardo-Nirenberg-Sobolev inequality, and Hardy's inequality in the potential space $L_{\AH}^{α, p}$.
Taste peptides have emerged as promising natural flavoring agents attributed to their unique organoleptic properties, high safety profile, and potential health benefits. However, the de novo identification of taste peptides derived from animal, plant, or microbial sources remains a time-consuming and resource-intensive process, significantly impeding their widespread application in the food industry. Here, we present TastePepAI, a comprehensive artificial intelligence framework for customized taste peptide design and safety assessment. As the key element of this framework, a loss-supervised adaptive variational autoencoder (LA-VAE) is implemented to efficiently optimizes the latent representation of sequences during training and facilitates the generation of target peptides with desired taste profiles. Notably, our model incorporates a novel taste-avoidance mechanism, allowing for selective flavor exclusion. Subsequently, our in-house developed toxicity prediction algorithm (SpepToxPred) is integrated in the framework to undergo rigorous safety evaluation of generated peptides. Using this integrated platform, we successfully identified 73 peptides exhibiting sweet, salty, and umami, significantly expanding the current repertoire of taste peptides. This work demonstrates the potential of TastePepAI in accelerating taste peptide discovery for food applications and provides a versatile framework adaptable to broader peptide engineering challenges.
Heart rate variability (HRV) is widely recognized as a valuable biomarker for assessing autonomic cardiac regulation. Pulse rate variability (PRV) is a common surrogate of HRV given the wide usability of PPG in commercially available devices. However, there is no clear conclusion on whether PRV can replace HRV given their different physiological mechanisms. This study evaluates the interchangeability of young adults HRV and PRV during supine-to-stand (STS) tests which are known as common posture transitions in daily life monitoring. Fifteen features from time, frequency and nonlinear domains were extracted from both electrocardiography and PPG signals. Paired t-tests and Wilcoxon signed-rank tests examined the difference between the extracted HRV and PRV features during supine, transition and standing phases separately. One feature showed significant difference in the supine phase, and this discrepancy increased to four in the transition and standing phases. These findings suggested that PRV is different from HRV in the STS tests, despite the fact that both metrics can reflect the sympathetic activation triggered by the posture changes.
Kullback--Leibler (KL) divergence is a fundamental measure of the dissimilarity between two probability distributions, but it can become unstable in high-dimensional settings due to its sensitivity to mismatches in distributional support. To address robustness limitations, we propose a novel Quantum-Inspired Fidelity-based Divergence (QIF), leveraging quantum information principles yet efficiently computable on classical hardware. Compared to KL divergence, QIF demonstrates improved numerical stability under partial or near-disjoint support conditions, thereby reducing the need for extensive regularization in specific scenarios. Moreover, QIF admits well-defined theoretical bounds and continuous similarity measures. Building on this, we introduce a novel regularization method, QR-Drop, which utilizes QIF to improve generalization in machine learning models. Empirical results show that QR-Drop effectively mitigates overfitting and outperforms state-of-the-art methods.
Artificial Intelligence (AI) has the potential to revolutionize diagnosis and segmentation in medical imaging. However, development and clinical implementation face multiple challenges including limited data availability, lack of generalizability, and the necessity to incorporate multi-modal data effectively. A foundation model, which is a large-scale pre-trained AI model, offers a versatile base that can be adapted to a variety of specific tasks and contexts. Here, we present VIsualization and Segmentation Masked AutoEncoder (VIS-MAE), novel model weights specifically designed for medical imaging. Specifically, VIS-MAE is trained on a dataset of 2.5 million unlabeled images from various modalities (CT, MR, PET,X-rays, and ultrasound), using self-supervised learning techniques. It is then adapted to classification and segmentation tasks using explicit labels. VIS-MAE has high label efficiency, outperforming several benchmark models in both in-domain and out-of-domain applications. In addition, VIS-MAE has improved label efficiency as it can achieve similar performance to other models with a reduced amount of labeled training data (50% or 80%) compared to other pre-trained weights. VIS-MAE represents a significant advancement in medical imaging AI, offering a generalizable and robust solution for improving segmentation and classification tasks while reducing the data annotation workload. The source code of this work is available at https://github.com/lzl199704/VIS-MAE.
Through the Rossiter-McLaughlin effect, several hot Jupiters have been found to exhibit spin-orbit misalignment, and even retrograde orbits. The high obliquity observed in these planets can be attributed to two primary formation mechanisms, as summarized in the existing literature. First, the host star's spin becomes misaligned with the planetary disk during the late stages of star formation, primarily due to chaotic accretion and magnetic interactions between the star and the planetary disk. Second, the orbital inclination of an individual planet can be excited by dynamical processes such as planet-planet scattering, the Lidov-Kozai cycle, and secular chaos within the framework of Newtonian mechanics. This study introduces a third mechanism, where, within the framework of general relativity, the post-Newtonian spin-orbit coupling term induces precession of the host star's spin around the orbital angular momentum. The orbital inclination, relative to a reference plane, can expand the range of deviation in the spatial orientation of the bodies' spins from the plane's normal. The varying amplitude and period of spin precession for both the star and the planet are derived theoretically, and the results, which can be applied without restriction, agree well with numerical simulations.
Recent advances of large language models in the field of Verilog generation have raised several ethical and security concerns, such as code copyright protection and dissemination of malicious code. Researchers have employed watermarking techniques to identify codes generated by large language models. However, the existing watermarking works fail to protect RTL code copyright due to the significant syntactic and semantic differences between RTL code and software code in languages such as Python. This paper proposes a hardware watermarking framework RTLMarker that embeds watermarks into RTL code and deeper into the synthesized netlist. We propose a set of rule-based Verilog code transformations , ensuring the watermarked RTL code's syntactic and semantic correctness. In addition, we consider an inherent tradeoff between watermark transparency and watermark effectiveness and jointly optimize them. The results demonstrate RTLMarker's superiority over the baseline in RTL code watermarking.
Cloud computing has revolutionized the provisioning of computing resources, offering scalable, flexible, and on-demand services to meet the diverse requirements of modern applications. At the heart of efficient cloud operations are job scheduling and resource management, which are critical for optimizing system performance and ensuring timely and cost-effective service delivery. However, the dynamic and heterogeneous nature of cloud environments presents significant challenges for these tasks, as workloads and resource availability can fluctuate unpredictably. Traditional approaches, including heuristic and meta-heuristic algorithms, often struggle to adapt to these real-time changes due to their reliance on static models or predefined rules. Deep Reinforcement Learning (DRL) has emerged as a promising solution to these challenges by enabling systems to learn and adapt policies based on continuous observations of the environment, facilitating intelligent and responsive decision-making. This survey provides a comprehensive review of DRL-based algorithms for job scheduling and resource management in cloud computing, analyzing their methodologies, performance metrics, and practical applications. We also highlight emerging trends and future research directions, offering valuable insights into leveraging DRL to advance both job scheduling and resource management in cloud computing.
Electrochemical nitrate reduction reaction(NO3RR)offers a sustainable route for ambient ammonia synthesis. While metal-nitrogen-carbon (M-N-C) single-atom catalysts have emerged as promising candidates for NO3RR, the structure-activity relations underlying their catalytic behavior remain to be elucidated. Through systematic analysis of reported experimental data and pH-field coupled microkinetic modelling on a reversible hydrogen electrode (RHE) scale, we reveal that the coordination-dependent activity originates from distinct scaling relations governed by metal-intermediate interactions. M-N-Pyrrolic catalysts demonstrate higher turnover frequencies for ammonia production, whereas M-N-Pyridinic catalysts exhibit broader activity ranges across the activity volcano plot. Meanwhile, the adsorption and protonation of nitrate, which is a step often dismissed and/or assumed to be simultaneous in many previous reports, is identified to be the rate-determining step (RDS) in NO3RR. Remarkably, our subsequent experimental validation confirms the theoretical predictions under both neutral and alkaline conditions. This study offers a comprehensive mechanistic framework for interpreting the electrocatalytic activity of M-N-C catalysts in NO3RR, showing that a classical thermodynamic limiting-potential model is not sufficiently accurate to capture the RDS and the catalytic performance trends of different materials (even on M-N-Pyrrolic and M-N-Pyridinic catalysts). These findings provide brand new insights into the reaction mechanism of NO3RR and establish fundamental design principles for electrocatalytic ammonia synthesis.