arXiv subjects

Gang Chen

Publications and source records attributed to Gang Chen.

At least 181 records · Page 10Linked to original sources

arXiv2024

Localization transitions of correlated particles in nonreciprocal quasicrystals

The interplay among interaction, non-Hermiticity, and disorder opens a new avenue for engineering novel phase transitions. We here study the spectral and localization features of two interacting bosons in one-dimensional nonreciprocal quasicrystals. Specifically, by considering a quasiperiodic Hubbard lattice with nonreciprocal hoppings, we show that the interaction can lead to a mobility edge, which arises from the fact that the bound states display a much lower threshold for spectral and extended-localized transitions than scattering states. The localization transition of bound or scattering states is accompanied by a complex-real spectrum transition. Moreover, while the two-particle localized states are robust to the boundary conditions, the two-particle extended states turn into skin modes under open boundary condition. We also show the correlated dynamics to characterize these localization transitions. Finally, we reveal that the bound states can form mobility edge on their own by introducing a dimerized nonreciprocal quasicrystal. Our paper may pave the way for the study of non-Hermitian few-body physics.

cond-mat.dis-nn↗

arXiv2024

A High-Throughput FPGA Accelerator for Lightweight CNNs With Balanced Dataflow

FPGA accelerators for lightweight neural convolutional networks (LWCNNs) have recently attracted significant attention. Most existing LWCNN accelerators focus on single-Computing-Engine (CE) architecture with local optimization. However, these designs typically suffer from high on-chip/off-chip memory overhead and low computational efficiency due to their layer-by-layer dataflow and unified resource mapping mechanisms. To tackle these issues, a novel multi-CE-based accelerator with balanced dataflow is proposed to efficiently accelerate LWCNN through memory-oriented and computing-oriented optimizations. Firstly, a streaming architecture with hybrid CEs is designed to minimize off-chip memory access while maintaining a low cost of on-chip buffer size. Secondly, a balanced dataflow strategy is introduced for streaming architectures to enhance computational efficiency by improving efficient resource mapping and mitigating data congestion. Furthermore, a resource-aware memory and parallelism allocation methodology is proposed, based on a performance model, to achieve better performance and scalability. The proposed accelerator is evaluated on Xilinx ZC706 platform using MobileNetV2 and ShuffleNetV2.Implementation results demonstrate that the proposed accelerator can save up to 68.3% of on-chip memory size with reduced off-chip memory access compared to the reference design. It achieves an impressive performance of up to 2092.4 FPS and a state-of-the-art MAC efficiency of up to 94.58%, while maintaining a high DSP utilization of 95%, thus significantly outperforming current LWCNN accelerators.

cs.AR↗

arXiv2024

High-speed and High-quality Vision Reconstruction of Spike Camera with Spike Stability Theorem

Neuromorphic vision sensors, such as the dynamic vision sensor (DVS) and spike camera, have gained increasing attention in recent years. The spike camera can detect fine textures by mimicking the fovea in the human visual system, and output a high-frequency spike stream. Real-time high-quality vision reconstruction from the spike stream can build a bridge to high-level vision task applications of the spike camera. To realize high-speed and high-quality vision reconstruction of the spike camera, we propose a new spike stability theorem that reveals the relationship between spike stream characteristics and stable light intensity. Based on the spike stability theorem, two parameter-free algorithms are designed for the real-time vision reconstruction of the spike camera. To demonstrate the performances of our algorithms, two datasets (a public dataset PKU-Spike-High-Speed and a newly constructed dataset SpikeCityPCL) are used to compare the reconstruction quality and speed of various reconstruction methods. Experimental results show that, compared with the current state-of-the-art (SOTA) reconstruction methods, our reconstruction methods obtain the best tradeoff between the reconstruction quality and speed. Additionally, we design the FPGA implementation method of our algorithms to realize the real-time (running at 20,000 FPS) visual reconstruction. Our work provides new theorem and algorithm foundations for the real-time edge-end vision processing of the spike camera.

cs.CV↗

arXiv2024

SeSeMI: Secure Serverless Model Inference on Sensitive Data

Model inference systems are essential for implementing end-to-end data analytics pipelines that deliver the benefits of machine learning models to users. Existing cloud-based model inference systems are costly, not easy to scale, and must be trusted in handling the models and user request data. Serverless computing presents a new opportunity, as it provides elasticity and fine-grained pricing. Our goal is to design a serverless model inference system that protects models and user request data from untrusted cloud providers. It offers high performance and low cost, while requiring no intrusive changes to the current serverless platforms. To realize our goal, we leverage trusted hardware. We identify and address three challenges in using trusted hardware for serverless model inference. These challenges arise from the high-level abstraction of serverless computing, the performance overhead of trusted hardware, and the characteristics of model inference workloads. We present SeSeMI, a secure, efficient, and cost-effective serverless model inference system. It adds three novel features non-intrusively to the existing serverless infrastructure and nothing else.The first feature is a key service that establishes secure channels between the user and the serverless instances, which also provides access control to models and users' data. The second is an enclave runtime that allows one enclave to process multiple concurrent requests. The final feature is a model packer that allows multiple models to be executed by one serverless instance. We build SeSeMI on top of Apache OpenWhisk, and conduct extensive experiments with three popular machine learning models. The results show that SeSeMI achieves low latency and low cost at scale for realistic workloads.

cs.CR↗

arXiv2024

Unsupervised Cross-Domain Regression for Fine-grained 3D Game Character Reconstruction

With the rise of the ``metaverse'' and the rapid development of games, it has become more and more critical to reconstruct characters in the virtual world faithfully. The immersive experience is one of the most central themes of the ``metaverse'', while the reducibility of the avatar is the crucial point. Meanwhile, the game is the carrier of the metaverse, in which players can freely edit the facial appearance of the game character. In this paper, we propose a simple but powerful cross-domain framework that can reconstruct fine-grained 3D game characters from single-view images in an end-to-end manner. Different from the previous methods, which do not resolve the cross-domain gap, we propose an effective regressor that can greatly reduce the discrepancy between the real-world domain and the game domain. To figure out the drawbacks of no ground truth, our unsupervised framework has accomplished the knowledge transfer of the target domain. Additionally, an innovative contrastive loss is proposed to solve the instance-wise disparity, which keeps the person-specific details of the reconstructed character. In contrast, an auxiliary 3D identity-aware extractor is activated to make the results of our model more impeccable. Then a large set of physically meaningful facial parameters is generated robustly and exquisitely. Experiments demonstrate that our method yields state-of-the-art performance in 3D game character reconstruction.

cs.CV↗

arXiv2024

Precision Enhancement in Spatial Measurement by Introducing Squeezed Light into Weak Value Amplification

The precision enhancement is demonstrated in an optical spatial measurement based on weak value amplification (WVA) system and split-like detection, by injecting a TEM$_{10}$ squeezed vacuum beam. It is the first time to experimentally realize high-precision optical spatial measurement beyond the shot noise limit by using squeezing-assisted WVA. Based on the WVA technique, which can amplify the signal by increasing the number of photons that injected into the interferometer, squeezed beam injection can reduce the noise level and can further improve the signal-to-noise ratio(SNR). As a result, a SNR improvement of 2dB, i.e., 1.3 time precision enhancement, can be achieved, by using a 2dB squeezed beam injection and 2.6\% postselection probability in WVA process. Our work provides an effective method to accomplish higher precision in quantum spatial measurement, which has potential applications in gravitational wave interferometer calibration, super-resolution quantum imaging, etc.

quant-ph↗

arXiv2024

A hybridizable discontinuous Galerkin method for the indefinite time-harmonic Maxwell equations

In this paper, we aim to develop a hybridizable discontinuous Galerkin (HDG) method for the indefinite time-harmonic Maxwell equations with the perfectly conducting boundary in the three-dimensional space. First, we derive the wavenumber explicit regularity result, which plays an important role in the error analysis for the HDG method. Second, we prove a discrete inf-sup condition which holds for all positive mesh size $h$, for all wavenumber $k$, and for general domain $Ω$. Then, we establish the optimal order error estimates of the underlying HDG method with constant independent of the wavenumber. The theoretical results are confirmed by numerical experiments.

math.NA↗

arXiv2024

Reexamination of evaporation from horizontal surfaces with implications for solar interfacial evaporation experiments

To explain reported solar interfacial-evaporation rates from porous materials beyond an apparent 100% efficiency using the thermal evaporation mechanism, many publications hypothesize that intermediate water inside porous materials have a reduced latent heat. Key supporting evidence is that water-only surfaces have lower dark evaporation rates than porous evaporators, with the ratio of the two rates taken as the latent heat reduction. Through simulations and experiments, we present benchmark evaporation rates of water and show that reported differences in natural evaporation are likely due to experimental error from recessed evaporating surfaces rather than from reduced latent heat. A few millimeters recession of the evaporating surface can drop evaporation rates over 50% due to a stagnant air layer, suggesting that the comparative experiments are prone to error and the latent heat reduction hypothesis cannot be substantiated. Our results indicate that new mechanistic directions need to be pursued to understand superthermal evaporation.

physics.app-ph↗

arXiv2024

Agent Centric Operating System -- a Comprehensive Review and Outlook for Operating System

The operating system (OS) is the backbone of modern computing, providing essential services and managing resources for computer hardware and software. This review paper offers an in-depth analysis of operating systems' evolution, current state, and prospects. We begin with an overview of the concept and significance of operating systems in the digital era. In the second section, we delve into the existing released operating systems, examining their architectures, functionalities, and the ecosystems they support. We then explore recent advances in OS evolution, highlighting innovations in real-time processing, distributed computing, and security. The third section focuses on the new era of operating systems, discussing emerging trends like the Internet of Things (IoT), cloud computing, and artificial intelligence (AI) integration. We also consider the challenges and opportunities presented by these developments. This review concludes with a synthesis of the current landscape and a forward-looking discussion on the future trajectories of operating systems, including open issues and areas ripe for further research and innovation. Finally, we put forward a new OS architecture.

cs.DC↗

arXiv2024

TableGPT2: A Large Multimodal Model with Tabular Data Integration

The emergence of models like GPTs, Claude, LLaMA, and Qwen has reshaped AI applications, presenting vast new opportunities across industries. Yet, the integration of tabular data remains notably underdeveloped, despite its foundational role in numerous real-world domains. This gap is critical for three main reasons. First, database or data warehouse data integration is essential for advanced applications; second, the vast and largely untapped resource of tabular data offers immense potential for analysis; and third, the business intelligence domain specifically demands adaptable, precise solutions that many current LLMs may struggle to provide. In response, we introduce TableGPT2, a model rigorously pre-trained and fine-tuned with over 593.8K tables and 2.36M high-quality query-table-output tuples, a scale of table-related data unprecedented in prior research. This extensive training enables TableGPT2 to excel in table-centric tasks while maintaining strong general language and coding abilities. One of TableGPT2's key innovations is its novel table encoder, specifically designed to capture schema-level and cell-level information. This encoder strengthens the model's ability to handle ambiguous queries, missing column names, and irregular tables commonly encountered in real-world applications. Similar to visual language models, this pioneering approach integrates with the decoder to form a robust large multimodal model. We believe the results are compelling: over 23 benchmarking metrics, TableGPT2 achieves an average performance improvement of 35.20% in the 7B model and 49.32% in the 72B model over prior benchmark-neutral LLMs, with robust general-purpose capabilities intact.

cs.LG↗

arXiv2024

One-loop Gravitational Bremsstrahlung and Waveforms from a Heavy-Mass Effective Field Theory

Using a heavy-mass effective field theory (HEFT), we study gravitational-wave emission in the scattering of two spinless black holes or neutron stars of arbitrary masses at next-to-leading order in the Post-Minkowskian expansion. We compute the contributions to the one-loop scattering amplitude with four scalars and one graviton which are relevant to the calculation of the waveforms, also presenting expressions of classical tree-level amplitudes with four scalars and up to two radiated gravitons. The latter are obtained using a novel on-shell recursion relation for classical amplitudes with four scalars and an arbitrary number of gravitons. Our one-loop five-point amplitude is expressed in terms of a single family of master integrals with the principal value prescription for linearised massive propagators, which we evaluate using differential equations. In our HEFT approach all hyper-classical iterations and quantum corrections to the amplitude are dropped at the diagrammatic level, thereby computing directly contributions to classical physics. Our result exhibits the expected factorisation of infrared divergences, the correct soft limits, and highly nontrivial cancellations of spurious poles. Finally, using our amplitude result we compute numerically the corresponding next-to-leading corrections to the spectral waveforms and the far-field time-domain waveforms using the Newman-Penrose scalar $Ψ_4$.

hep-th↗

arXiv2024

Revisiting CNNs for Trajectory Similarity Learning

Similarity search is a fundamental but expensive operator in querying trajectory data, due to its quadratic complexity of distance computation. To mitigate the computational burden for long trajectories, neural networks have been widely employed for similarity learning and each trajectory is encoded as a high-dimensional vector for similarity search with linear complexity. Given the sequential nature of trajectory data, previous efforts have been primarily devoted to the utilization of RNNs or Transformers. In this paper, we argue that the common practice of treating trajectory as sequential data results in excessive attention to capturing long-term global dependency between two sequences. Instead, our investigation reveals the pivotal role of local similarity, prompting a revisit of simple CNNs for trajectory similarity learning. We introduce ConvTraj, incorporating both 1D and 2D convolutions to capture sequential and geo-distribution features of trajectories, respectively. In addition, we conduct a series of theoretical analyses to justify the effectiveness of ConvTraj. Experimental results on four real-world large-scale datasets demonstrate that ConvTraj achieves state-of-the-art accuracy in trajectory similarity search. Owing to the simple network structure of ConvTraj, the training and inference speed on the Porto dataset with 1.6 million trajectories are increased by at least $240$x and $2.16$x, respectively. The source code and dataset can be found at \textit{\url{https://github.com/Proudc/ConvTraj}}.

cs.AI↗

arXiv2024

Powering In-Database Dynamic Model Slicing for Structured Data Analytics

Relational database management systems (RDBMS) are widely used for the storage of structured data. To derive insights beyond statistical aggregation, we typically have to extract specific subdatasets from the database using conventional database operations, and then apply deep neural networks (DNN) training and inference on these subdatasets in a separate analytics system. The process can be prohibitively expensive, especially when there are various subdatasets extracted for different analytical purposes. This calls for efficient in-database support of advanced analytical methods. In this paper, we introduce LEADS, a novel SQL-aware dynamic model slicing technique to customize models for specified SQL queries. LEADS improves the predictive modeling of structured data via the mixture of experts (MoE) and maintains efficiency by a SQL-aware gating network. At the core of LEADS is the construction of a general model with multiple expert sub-models trained over the database. The MoE scales up the modeling capacity, enhances effectiveness, and preserves efficiency by activating necessary experts via the SQL-aware gating network during inference. To support in-database analytics, we build an inference extension that integrates LEADS onto PostgreSQL. Our extensive experiments on real-world datasets demonstrate that LEADS consistently outperforms the baseline models, and the in-database inference extension delivers a considerable reduction in inference latency compared to traditional solutions.

cs.DB↗

arXiv2024

MoNTA: Accelerating Mixture-of-Experts Training with Network-Traffc-Aware Parallel Optimization

The Mixture of Experts (MoE) is an advanced model architecture in the industry that combines multiple specialized expert models from various domains into a single supermodel. This approach enables the model to scale without significantly increasing the computational costs of training and inference, while maximizing model performance. However, current distributed training frameworks do not consider the ultimate optimization of communication, especially for large base models. This paper proposes a network-traffic-aware parallel optimization method that selects the optimal parallel strategy based on the communication volume, and the training cluster's inter-node and intra-node network topologies. Compared to the DeepSpeed, MoNTA achieves an 8x increase in AllToAll communication performance under 8-card tensor parallelism. Compared to the baseline, training a 2x70B model using 16 A800 cards, with an 8K sequence, results in a 13% overall latency performance improvement. Project Page: https://github.com/EnflameTechnology/DeepSpeed.

cs.LG↗

arXiv2024

Real-time Stereo-based 3D Object Detection for Streaming Perception

The ability to promptly respond to environmental changes is crucial for the perception system of autonomous driving. Recently, a new task called streaming perception was proposed. It jointly evaluate the latency and accuracy into a single metric for video online perception. In this work, we introduce StreamDSGN, the first real-time stereo-based 3D object detection framework designed for streaming perception. StreamDSGN is an end-to-end framework that directly predicts the 3D properties of objects in the next moment by leveraging historical information, thereby alleviating the accuracy degradation of streaming perception. Further, StreamDSGN applies three strategies to enhance the perception accuracy: (1) A feature-flow-based fusion method, which generates a pseudo-next feature at the current moment to address the misalignment issue between feature and ground truth. (2) An extra regression loss for explicit supervision of object motion consistency in consecutive frames. (3) A large kernel backbone with a large receptive field for effectively capturing long-range spatial contextual features caused by changes in object positions. Experiments on the KITTI Tracking dataset show that, compared with the strong baseline, StreamDSGN significantly improves the streaming average precision by up to 4.33%. Our code is available at https://github.com/weiyangdaren/streamDSGN-pytorch.

cs.CV↗

arXiv2024

Nonreciprocal reflection of mid-infrared light by highly doped InAs at low magnetic fields

We report an experimental observation of room-temperature nonreciprocal reflection of mid-infrared light from planar highly doped InAs surfaces at low magnetic fields ranging from 0.07 T to 0.16 T. Using ellipsometry, we demonstrate that the amplitude ratio and phase shift of reflected light are nonreciprocal in the Voigt configuration. We also demonstrate using Fourier-transform infrared spectroscopy that the nonreciprocal reflectance contrast (the difference in reflectance in opposite directions) increases with the magnitude of the magnetic field for p-polarized light. Our work is a step toward the practical implementation of nonreciprocal thermal emitters and absorbers and applications such as remote magnetic field sensing.

physics.optics↗

arXiv2024

Using PT-symmetric Qubits to Break the Tradeoff Between Fidelity and the Degree of Quantum Entanglement

A noteworthy discovery is that the minimal evolution time is smaller for parity-time ($\mathcal{PT}$) symmetric systems compared to Hermitian setups. Moreover, there is a significant acceleration of two-qubit quantum entanglement preparation near the exceptional point (EP), or spectral coalescence, within such system. Nevertheless, an important problem often overlooked for quantum EP-based devices is their fidelity, greatly affected by the process of dissipation or post-selection, creating an inherent trade-off relation between the degree of entanglement and fidelity. Our study demonstrates that this limitation can be effectively overcome by harnessing an active $\mathcal{PT}$-symmetric system, which possesses balanced gain and loss, enabling maximal entanglement with rapid speed, high fidelity, and greater resilience to non-resonant errors. This new approach can efficiently prepare multi-qubit entanglement and use not only bipartite but also tripartite entanglement, as illustrative examples, even when the precise gain-loss balance is not strictly maintained. Our analytical findings are in excellent agreement with numerical simulations, confirming the potential of truly $\mathcal{PT}$-devices as a powerful tool for creating and engineering diverse quantum resources for applications in quantum information technology

quant-ph↗

arXiv2024

Machine Learning for Raman Spectroscopy-based Cyber-Marine Fish Biochemical Composition Analysis

The rapid and accurate detection of biochemical compositions in fish is a crucial real-world task that facilitates optimal utilization and extraction of high-value products in the seafood industry. Raman spectroscopy provides a promising solution for quickly and non-destructively analyzing the biochemical composition of fish by associating Raman spectra with biochemical reference data using machine learning regression models. This paper investigates different regression models to address this task and proposes a new design of Convolutional Neural Networks (CNNs) for jointly predicting water, protein, and lipids yield. To the best of our knowledge, we are the first to conduct a successful study employing CNNs to analyze the biochemical composition of fish based on a very small Raman spectroscopic dataset. Our approach combines a tailored CNN architecture with the comprehensive data preparation procedure, effectively mitigating the challenges posed by extreme data scarcity. The results demonstrate that our CNN can significantly outperform two state-of-the-art CNN models and multiple traditional machine learning models, paving the way for accurate and automated analysis of fish biochemical composition.

cs.LG↗