Search

Algebraic Multigrid (AMG) is one of the most used iterative algorithms for solving large sparse linear equations $Ax=b$. In AMG, the coarse grid is a key component that affects the efficiency of the algorithm, the construction of which relies on the strong threshold parameter $\theta$. This parameter is generally chosen empirically, with a default value in many current AMG solvers of 0.25 for 2D problems and 0.5 for 3D problems. However, for many practical problems, the quality of the coarse grid and the efficiency of the AMG algorithm are sensitive to $\theta$; the default value is rarely optimal, and sometimes is far from it. Therefore, how to choose a better $\theta$ is an important question. In this paper, we propose a deep learning based auto-tuning method, AutoAMG($\theta$) for multiscale sparse linear equations, which are widely used in practical problems. The method uses Graph Neural Networks (GNNs) to extract matrix features, and a Multilayer Perceptron (MLP) to build the mapping between matrix features and the optimal $\theta$, which can adaptively output $\theta$ values for different matrices. Numerical experiments show that AutoAMG($\theta$) can achieve significant speedup compared to the default $\theta$ value.

This study investigates the impact of molecular thermal fluctuations on compressible decaying isotropic turbulence using the unified stochastic particle (USP) method, encompassing both two-dimensional (2D) and three-dimensional (3D) scenarios. The findings reveal that the turbulent spectra of velocity and thermodynamic variables follow the wavenumber scaling law of ${k}^{(d-1)}$ for different spatial dimensions $d$ within the high wavenumber range, indicating the impact of thermal fluctuations on small-scale turbulent statistics. With the application of Helmholtz decomposition, it is found that the thermal fluctuation spectra of solenoidal and compressible velocity components (${\vec{u}}_{s}$ and ${\vec{u}}_{c}$) follow an energy ratio of 1:1 for 2D cases, while the ratio changes to 2:1 for 3D cases. Comparisons between 3D turbulent spectra obtained through USP simulations and direct numerical simulations of the Navier-Stokes equations demonstrate that thermal fluctuations dominate the spectra at length scales comparable to the Kolmogorov length scale. Additionally, the effect of thermal fluctuations on the spectrum of ${\vec{u}}_{c}$ is significantly influenced by variations in the turbulent Mach number. We further study the impact of thermal fluctuations on the predictability of turbulence. With initial differences caused by thermal fluctuations, different flow realizations display significant disparities in velocity and thermodynamic fields at larger scales after a certain period of time, which can be characterized by "inverse error cascades". Moreover, the results suggest a strong correlation between the predictabilities of thermodynamic fields and the predictability of ${\vec{u}}_{c}$.

With the widespread use of deep neural networks(DNNs) in intelligent systems, DNN accelerators with high performance and energy efficiency are greatly demanded. As one of the feasible processing-in-memory(PIM) architectures, 3D-stacked-DRAM-based PIM(DRAM-PIM) architecture enables large-capacity memory and low-cost memory access, which is a promising solution for DNN accelerators with better performance and energy efficiency. However, the low-cost characteristics of stacked DRAM and the distributed manner of memory access and data storing require us to rebalance the hardware design and DNN mapping. In this paper, we propose NicePIM to efficiently explore the design space of hardware architecture and DNN mapping of DRAM-PIM accelerators, which consists of three key components: PIM-Tuner, PIM-Mapper and Data-Scheduler. PIM-Tuner optimizes the hardware configurations leveraging a DNN model for classifying area-compliant architectures and a deep kernel learning model for identifying better hardware parameters. PIM-Mapper explores a variety of DNN mapping configurations, including parallelism between branches of DNN, DNN layer partitioning, DRAM capacity allocation and data layout pattern in DRAM to generate high-hardware-utilization DNN mapping schemes for various hardware configurations. The Data-Scheduler employs an integer-linear-programming-based data scheduling algorithm to alleviate the inter-PIM-node communication overhead of data-sharing brought by DNN layer partitioning. Experimental results demonstrate that NicePIM can optimize hardware configurations for DRAM-PIM systems effectively and can generate high-quality DNN mapping schemes with latency and energy cost reduced by 37% and 28% on average respectively compared to the baseline method., Comment: 13 pages

The local and global features are both essential for automatic speech recognition (ASR). Many recent methods have verified that simply combining local and global features can further promote ASR performance. However, these methods pay less attention to the interaction of local and global features, and their series architectures are rigid to reflect local and global relationships. To address these issues, this paper proposes InterFormer for interactive local and global features fusion to learn a better representation for ASR. Specifically, we combine the convolution block with the transformer block in a parallel design. Besides, we propose a bidirectional feature interaction module (BFIM) and a selective fusion module (SFM) to implement the interaction and fusion of local and global features, respectively. Extensive experiments on public ASR datasets demonstrate the effectiveness of our proposed InterFormer and its superior performance over the other Transformer and Conformer models., Comment: Accepted by Interspeech 2023

Most contemporary supervised Remote Sensing (RS) image Change Detection (CD) approaches are customized for equal-resolution bitemporal images. Real-world applications raise the need for cross-resolution change detection, aka, CD based on bitemporal images with different spatial resolutions. Given training samples of a fixed bitemporal resolution difference (ratio) between the high-resolution (HR) image and the low-resolution (LR) one, current cross-resolution methods may fit a certain ratio but lack adaptation to other resolution differences. Toward continuous cross-resolution CD, we propose scale-invariant learning to enforce the model consistently predicting HR results given synthesized samples of varying resolution differences. Concretely, we synthesize blurred versions of the HR image by random downsampled reconstructions to reduce the gap between HR and LR images. We introduce coordinate-based representations to decode per-pixel predictions by feeding the coordinate query and corresponding multi-level embedding features into an MLP that implicitly learns the shape of land cover changes, therefore benefiting recognizing blurred objects in the LR image. Moreover, considering that spatial resolution mainly affects the local textures, we apply local-window self-attention to align bitemporal features during the early stages of the encoder. Extensive experiments on two synthesized and one real-world different-resolution CD datasets verify the effectiveness of the proposed method. Our method significantly outperforms several vanilla CD methods and two cross-resolution CD methods on the three datasets both in in-distribution and out-of-distribution settings. The empirical results suggest that our method could yield relatively consistent HR change predictions regardless of varying bitemporal resolution ratios. Our code is available at \url{https://github.com/justchenhao/SILI_CD}., Comment: 21 pages, 11 figures. Accepted article by IEEE TGRS

Attention-based encoder-decoder (AED) models have shown impressive performance in ASR. However, most existing AED methods neglect to simultaneously leverage both acoustic and semantic features in decoder, which is crucial for generating more accurate and informative semantic states. In this paper, we propose an Acoustic and Semantic Cooperative Decoder (ASCD) for ASR. In particular, unlike vanilla decoders that process acoustic and semantic features in two separate stages, ASCD integrates them cooperatively. To prevent information leakage during training, we design a Causal Multimodal Mask. Moreover, a variant Semi-ASCD is proposed to balance accuracy and computational cost. Our proposal is evaluated on the publicly available AISHELL-1 and aidatatang_200zh datasets using Transformer, Conformer, and Branchformer as encoders, respectively. The experimental results show that ASCD significantly improves the performance by leveraging both the acoustic and semantic information cooperatively., Comment: Accepted by Interspeech 2023

Theoretical studies on spin-dependent transport through helical molecules with straight spiral geometry have received intense research interest in the past decade, however, the physics in circular helical molecules has still less been explored. In this work, we theoretically construct a circular single helix (CSH) possessing the chirality-induced spin-orbit coupling and contacting with two non-magnetic electrodes. Our theoretical calculations demonstrate that the spin-related transport in CSH exhibits the so-called chiral-induced spin selectivity (CISS) effect and more importantly, the CISS-reduced spin-dependent destructive quantum interference (DQI) also occurs in the CSH, without any external magnetic field or magnetic electrodes. Moreover, the changing of CSH length or the electrode positions exhibits specific patterns in the spin-polarized conductance. Particularly, the dephasing magnitude can adjust effectively these two spin-dependent effects to realize their coexistence. Additionally, the phase transition between the CISS-dependent constructive quantum interference (CQI) and DQI is also observed in the CSH. Our theoretical work puts forwards a new material plateau to explore the CISS and to exhibit the novel CISS-dependent CQI effect., Comment: 10 pages, 5 figures, 1 table

In this paper, we investigate a distributed aggregative optimization problem in a network, where each agent has its own local cost function which depends not only on the local state variable but also on an aggregated function of state variables from all agents. To accelerate the optimization process, we combine heavy ball and Nesterov's accelerated methods with distributed aggregative gradient tracking, and propose two novel algorithms named DAGT-HB and DAGT-NES for solving the distributed aggregative optimization problem. We analyse that the DAGT-HB and DAGT-NES algorithms can converge to an optimal solution at a global $\mathbf{R}-$linear convergence rate when the objective function is smooth and strongly convex, and when the parameters (e.g., step size and momentum coefficients) are selected within certain ranges. A numerical experiment on the optimal placement problem is given to verify the effectiveness and superiority of our proposed algorithms.

Pretraining on large-scale datasets can boost the performance of object detectors while the annotated datasets for object detection are hard to scale up due to the high labor cost. What we possess are numerous isolated filed-specific datasets, thus, it is appealing to jointly pretrain models across aggregation of datasets to enhance data volume and diversity. In this paper, we propose a strong framework for utilizing Multiple datasets to pretrain DETR-like detectors, termed METR, without the need for manual label spaces integration. It converts the typical multi-classification in object detection into binary classification by introducing a pre-trained language model. Specifically, we design a category extraction module for extracting potential categories involved in an image and assign these categories into different queries by language embeddings. Each query is only responsible for predicting a class-specific object. Besides, to adapt our novel detection paradigm, we propose a group bipartite matching strategy that limits the ground truths to match queries assigned to the same category. Extensive experiments demonstrate that METR achieves extraordinary results on either multi-task joint training or the pretrain & finetune paradigm. Notably, our pre-trained models have high flexible transferability and increase the performance upon various DETR-like detectors on COCO val2017 benchmark. Codes will be available after this paper is published.

The vanilla fractional order gradient descent may oscillatively converge to a region around the global minimum instead of converging to the exact minimum point, or even diverge, in the case where the objective function is strongly convex. To address this problem, a novel adaptive fractional order gradient descent (AFOGD) method and a novel adaptive fractional order accelerated gradient descent (AFOAGD) method are proposed in this paper. Inspired by the quadratic constraints and Lyapunov stability analysis from robust control theory, we establish a linear matrix inequality to analyse the convergence of our proposed algorithms. We prove that the proposed algorithms can achieve R-linear convergence when the objective function is $\textbf{L-}$smooth and $\textbf{m-}$strongly-convex. Several numerical simulations are demonstrated to verify the effectiveness and superiority of our proposed algorithms., Comment: 8pages,5 figures

Physics-informed neural networks (PINNs) have effectively been demonstrated in solving forward and inverse differential equation problems, but they are still trapped in training failures when the target functions to be approximated exhibit high-frequency or multi-scale features. In this paper, we propose to employ implicit stochastic gradient descent (ISGD) method to train PINNs for improving the stability of training process. We heuristically analyze how ISGD overcome stiffness in the gradient flow dynamics of PINNs, especially for problems with multi-scale solutions. We theoretically prove that for two-layer fully connected neural networks with large hidden nodes, randomly initialized ISGD converges to a globally optimal solution for the quadratic loss function. Empirical results demonstrate that ISGD works well in practice and compares favorably to other gradient-based optimization methods such as SGD and Adam, while can also effectively address the numerical stiffness in training dynamics via gradient descent., Comment: 17 pages, published as a conference paper at AAAI23

Falls have become more frequent in recent years, which has been harmful for senior citizens.Therefore detecting falls have become important and several data sets and machine learning model have been introduced related to fall detection. In this project report, a human fall detection method is proposed using a multi modality approach. We used the UP-FALL detection data set which is collected by dozens of volunteers using different sensors and two cameras. We use wrist sensor with acclerometer data keeping labels to binary classification, namely fall and no fall from the data set.We used fusion of camera and sensor data to increase performance. The experimental results shows that using only wrist data as compared to multi sensor for binary classification did not impact the model prediction performance for fall detection.

Bit-serial architectures can handle Neural Networks (NNs) with different weight precisions, achieving higher resource efficiency compared with bit-parallel architectures. Besides, the weights contain abundant zero bits owing to the fault tolerance of NNs, indicating that bit sparsity of NNs can be further exploited for performance improvement. However, the irregular proportion of zero bits in each weight causes imbalanced workloads in the Processing Element (PE) array, which degrades performance or induces overhead for sparse processing. Thus, this paper proposed a bit-sparsity quantization method to maintain the bit sparsity ratio of each weight to no more than a certain value for balancing workloads, with little accuracy loss. Then, we co-designed a sparse bit-serial architecture, called Bit-balance, to improve overall performance, supporting weight-bit sparsity and adaptive bitwidth computation. The whole design was implemented with 65nm technology at 1 GHz and performs at 326-, 30-, 56-, and 218-frame/s for AlexNet, VGG-16, ResNet-50, and GoogleNet respectively. Compared with sparse bit-serial accelerator, Bitlet, Bit-balance achieves 1.8x~2.7x energy efficiency (frame/J) and 2.1x~3.7x resource efficiency (frame/mm2)., Comment: 10 pages 17 figures

For satellite images, the presence of clouds presents a problem as clouds obscure more than half to two-thirds of the ground information. This problem causes many issues for reliability in a noise-free environment to communicate data and other applications that need seamless monitoring. Removing the clouds from the images while keeping the background pixels intact can help address the mentioned issues. Recently, deep learning methods have become popular for researching cloud removal by demonstrating promising results, among which Generative Adversarial Networks (GAN) have shown considerably better performance. In this project, we aim to address cloud removal from satellite images using AttentionGAN and then compare our results by reproducing the results obtained using traditional GANs and auto-encoders. We use RICE dataset. The outcome of this project can be used to develop applications that require cloud-free satellite images. Moreover, our results could be helpful for making further research improvements.

Modern field programmable gate array(FPGA) can be partially dynamically reconfigurable with heterogeneous resources distributed on the chip. And FPGA-based partially dynamically reconfigurable system(FPGA-PDRS) can be used to accelerate computing and improve computing flexibility. However, the traditional design of FPGA-PDRS is based on manual design. Implementing the automation of FPGA-PDRS needs to solve the problems of task modules partitioning, scheduling, and floorplanning on heterogeneous resources. Existing works only partly solve problems for the automation process of FPGA-PDRS or model homogeneous resource for FPGA-PDRS. To better solve the problems in the automation process of FPGA-PDRS and narrow the gap between algorithm and application, in this paper, we propose a complete workflow including three parts, pre-processing to generate the list of task modules candidate shapes according to the resources requirements, exploration process to search the solution of task modules partitioning, scheduling, and floorplanning, and post-optimization to improve the success rate of floorplan. Experimental results show that, compared with state-of-the-art work, the proposed complete workflow can improve performance by 18.7\%, reduce communication cost by 8.6\%, on average, with improving the resources reuse rate of the heterogeneous resources on the chip. And based on the solution generated by the exploration process, the post-optimization can improve the success rate of the floorplan by 14\%.