Tracking objects in 3D space and predicting their 6DoF pose is an essential task in computer vision. State-of-the-art approaches often rely on object texture to tackle this problem. However, while they achieve impressive results, many objects do not contain sufficient texture, violating the main underlying assumption. In the following, we thus propose ICG, a novel probabilistic tracker that fuses region and depth information and only requires the object geometry. Our method deploys correspondence lines and points to iteratively refine the pose. We also implement robust occlusion handling to improve performance in real-world settings. Experiments on the YCB-Video, OPT, and Choi datasets demonstrate that, even for textured objects, our approach outperforms the current state of the art with respect to accuracy and robustness. At the same time, ICG shows fast convergence and outstanding efficiency, requiring only 1.3 ms per frame on a single CPU core. Finally, we analyze the influence of individual components and discuss our performance compared to deep learning-based methods. The source code of our tracker is publicly available.
The demand to process vast amounts of data generated from state-of-the-art high resolution cameras has motivated novel energy-efficient on-device AI solutions. Visual data in such cameras are usually captured in the form of analog voltages by a sensor pixel array, and then converted to the digital domain for subsequent AI processing using analog-to-digital converters (ADC). Recent research has tried to take advantage of massively parallel low-power analog/digital computing in the form of near- and in-sensor processing, in which the AI computation is performed partly in the periphery of the pixel array and partly in a separate on-board CPU/accelerator. Unfortunately, high-resolution input images still need to be streamed between the camera and the AI processing unit, frame by frame, causing energy, bandwidth, and security bottlenecks. To mitigate this problem, we propose a novel Processing-in-Pixel-in-memory (P2M) paradigm, that customizes the pixel array by adding support for analog multi-channel, multi-bit convolution, batch normalization, and ReLU (Rectified Linear Units). Our solution includes a holistic algorithm-circuit co-design approach and the resulting P2M paradigm can be used as a drop-in replacement for embedding memory-intensive first few layers of convolutional neural network (CNN) models within foundry-manufacturable CMOS image sensor platforms. Our experimental results indicate that P2M reduces data transfer bandwidth from sensors and analog to digital conversions by ~21x, and the energy-delay product (EDP) incurred in processing a MobileNetV2 model on a TinyML use case for visual wake words dataset (VWW) by up to ~11x compared to standard near-processing or in-sensor implementations, without any significant drop in test accuracy.
We present the Versatile Grasp Quality Convolutional Neural Network (VGQ-CNN), a grasp quality prediction network for 6-DOF grasps. VGQ-CNN can be used when evaluating grasps for objects seen from a wide range of camera poses or mobile robots without the need to retrain the network. By defining the grasp orientation explicitly as an input to the network, VGQ-CNN can evaluate 6-DOF grasp poses, moving beyond the 4-DOF grasps used in most image-based grasp evaluation methods like GQ-CNN. To train VGQ-CNN, we generate the new Versatile Grasp dataset (VG-dset) containing 6-DOF grasps observed from a wide range of camera poses. VGQ-CNN achieves a balanced accuracy of 82.1% on our test-split while generalising to a variety of camera poses. Meanwhile, it achieves competitive performance for overhead cameras and top-grasps with a balanced accuracy of 74.2% compared to GQ-CNN's 76.6%. We also propose a modified network architecture, FAST-VGQ-CNN, that speeds up inference using a shared encoder architecture and can make 128 grasp quality predictions in 12ms on a CPU. Code and data are available at https://aucoroboticsmu.github.io/vgq-cnn/.
We present a compilation flow for the generation of CNN inference accelerators on FPGAs. The flow translates a frozen model into OpenCL kernels with the TVM compiler and uses the Intel OpenCL SDK to compile to an FPGA bitstream. We improve the quality of the generated hardware with optimizations applied to the base OpenCL kernels generated by TVM. These optimizations increase parallelism, reduce memory access latency, increase concurrency and save on-chip resources. We automate these optimizations in TVM and evaluate them by generating accelerators for LeNet-5, MobileNetV1 and ResNet-34 on an Intel Stratix~10SX. We show that the optimizations improve the performance of the generated accelerators by up to 846X over the base accelerators. The performance of the optimized accelerators is up to 4.57X better than TensorFlow on CPU, 3.83X better than single-threaded TVM and is only 0.34X compared to TVM with 56 threads. Our optimized kernels also outperform ones generated by a similar approach (that also uses high-level synthesis) while providing more functionality and flexibility. However, it underperforms an approach that utilizes hand-optimized designs. Thus, we view our approach as useful in pre-production environments that benefit from increased performance and fast prototyping, realizing the benefits of FPGAs without hardware design expertise.
Optimal Transport is a popular distance metric for measuring similarity between distributions. Exact algorithms for computing Optimal Transport can be slow, which has motivated the development of approximate numerical solvers (e.g. Sinkhorn method). We introduce a new and very simple combinatorial approach to find an $\varepsilon$-approximation of the OT distance. Our algorithm achieves a near-optimal execution time of $O(n^2/\varepsilon^2)$ for computing OT distance and, for the special case of the assignment problem, the execution time improves to $O(n^2/\varepsilon)$. Our algorithm is based on the push-relabel framework for min-cost flow problems. Unlike the other combinatorial approach (Lahn, Mulchandani and Raghvendra, NeurIPS 2019) which does not have a fast parallel implementation, our algorithm has a parallel execution time of $O(\log n/\varepsilon^2)$. Interestingly, unlike the Sinkhorn algorithm, our method also readily provides a compact transport plan as well as a solution to an approximate version of the dual formulation of the OT problem, both of which have numerous applications in Machine Learning. For the assignment problem, we provide both a CPU implementation as well as an implementation that exploits GPU parallelism. Experiments suggest that our algorithm is faster than the Sinkhorn algorithm, both in terms of CPU and GPU implementations, especially while computing matchings with a high accuracy.
In this work, we consider the problem of learning a perception model for monocular robot navigation using few annotated images. Using a Vision Transformer (ViT) pretrained with a label-free self-supervised method, we successfully train a coarse image segmentation model for the Duckietown environment using 70 training images. Our model performs coarse image segmentation at the 8x8 patch level, and the inference resolution can be adjusted to balance prediction granularity and real-time perception constraints. We study how best to adapt a ViT to our task and environment, and find that some lightweight architectures can yield good single-image segmentation at a usable frame rate, even on CPU. The resulting perception model is used as the backbone for a simple yet robust visual servoing agent, which we deploy on a differential drive mobile robot to perform two tasks: lane following and obstacle avoidance.
Fourier ptychography has attracted a wide range of focus for its ability of large space-bandwidth-produce, and quantative phase measurement. It is a typical computational imaging technique which refers to optimizing both the imaging hardware and reconstruction algorithms simultaneously. The data redundancy and inverse problem algorithms are the sources of FPM's excellent performance. But at the same time, this large amount of data processing and complex algorithms also greatly reduce the imaging speed. In this article, we propose a parallel Fourier ptychography reconstruction framework consisting of three levels of parallel computing parts and implemented it with both central processing unit (CPU) and compute unified device architecture (CUDA) platform. In the conventional FPM reconstruction framework, the sample image is divided into multiple sub-regions for separately processing because the illumination angles for different subregions are varied for the same LED and different subregions contain different defocus distances due to the non-planar distribution or non-ideal posture of biological sample. We first build a parallel computing sub-framework in spatial domain based on the above-mentioned characteristics. And then, by utilizing the sequential characteristics of different spectrum regions to update, a parallel computing sub-framework in the spectrum domain is carried out in our scheme. The feasibility of the proposed parallel FPM reconstruction framework is verified with different experimental results acquired with the system we built.
The rapid development of quantum computing has demonstrated many unique characteristics of quantum advantages, such as richer feature representation and more secured protection on model parameters. This work proposes a vertical federated learning architecture based on variational quantum circuits to demonstrate the competitive performance of a quantum-enhanced pre-trained BERT model for text classification. In particular, our proposed hybrid classical-quantum model consists of a novel random quantum temporal convolution (QTC) learning framework replacing some layers in the BERT-based decoder. Our experiments on intent classification show that our proposed BERT-QTC model attains competitive experimental results in the Snips and ATIS spoken language datasets. Particularly, the BERT-QTC boosts the performance of the existing quantum circuit-based language model in two text classification datasets by 1.57% and 1.52% relative improvements. Furthermore, BERT-QTC can be feasibly deployed on both existing commercial-accessible quantum computation hardware and CPU-based interface for ensuring data isolation.
The traditional vocoders have the advantages of high synthesis efficiency, strong interpretability, and speech editability, while the neural vocoders have the advantage of high synthesis quality. To combine the advantages of two vocoders, inspired by the traditional deterministic plus stochastic model, this paper proposes a novel neural vocoder named NeuralDPS which can retain high speech quality and acquire high synthesis efficiency and noise controllability. Firstly, this framework contains four modules: a deterministic source module, a stochastic source module, a neural V/UV decision module and a neural filter module. The input required by the vocoder is just the spectral parameter, which avoids the error caused by estimating additional parameters, such as F0. Secondly, to solve the problem that different frequency bands may have different proportions of deterministic components and stochastic components, a multiband excitation strategy is used to generate a more accurate excitation signal and reduce the neural filter's burden. Thirdly, a method to control noise components of speech is proposed. In this way, the signal-to-noise ratio (SNR) of speech can be adjusted easily. Objective and subjective experimental results show that our proposed NeuralDPS vocoder can obtain similar performance with the WaveNet and it generates waveforms at least 280 times faster than the WaveNet vocoder. It is also 28% faster than WaveGAN's synthesis efficiency on a single CPU core. We have also verified through experiments that this method can effectively control the noise components in the predicted speech and adjust the SNR of speech. Examples of generated speech can be found at https://hairuo55.github.io/NeuralDPS.
In this paper, we bring forward the use of the recently developed Signature Transform as a way to measure the similarity between image distributions and provide detailed acquaintance and extensive evaluations. We are the first to pioneer RMSE and MAE Signature, along with log-signature as an alternative to measure GAN convergence, a problem that has been extensively studied. We are also forerunners to introduce analytical measures based on statistics to study the goodness of fit of the GAN sample distribution that are both efficient and effective. Current GAN measures involve lots of computation normally done at the GPU and are very time consuming. In contrast, we diminish the computation time to the order of seconds and computation is done at the CPU achieving the same level of goodness. Lastly, a PCA adaptive t-SNE approach, which is novel in this context, is also proposed for data visualization.
Ant Colony Optimization (ACO) is a family of nature-inspired metaheuristics often applied to finding approximate solutions to difficult optimization problems. Despite being significantly faster than exact methods, the ACOs can still be prohibitively slow, especially if compared to basic problem-specific heuristics. As recent research has shown, it is possible to significantly improve the performance through algorithm refinements and careful parallel implementation benefiting from multi-core CPUs and dedicated accelerators. In this paper, we present a novel ACO variant, namely the Focused ACO (FACO). One of the core elements of the FACO is a mechanism for controlling the number of differences between a newly constructed and a selected previous solution. The mechanism results in a more focused search process, allowing to find improvements while preserving the quality of the existing solution. An additional benefit is a more efficient integration with a problem-specific local search. Computational study based on a range of the Traveling Salesman Problem instances shows that the FACO outperforms the state-of-the-art ACOs when solving large TSP instances. Specifically, the FACO required less than an hour of an 8-core commodity CPU time to find high-quality solutions (within 1% from the best-known results) for TSP Art Instances ranging from 100000 to 200000 nodes.
In recent years, due to a higher demand for portable devices, which provide restricted amounts of processing capacity and battery power, the need for energy and time efficient hard- and software solutions has increased. Preliminary estimations of time and energy consumption can thus be valuable to improve implementations and design decisions. To this end, this paper presents a method to estimate the time and energy consumption of a given software solution, without having to rely on the use of a traditional Cycle Accurate Simulator (CAS). Instead, we propose to utilize a combination of high-level functional simulation with a mechanistic extension to include non-functional properties: Instruction counts from virtual execution are multiplied with corresponding specific energies and times. By evaluating two common image processing algorithms on an FPGA-based CPU, where a mean relative estimation error of 3% is achieved for cacheless systems, we show that this estimation tool can be a valuable aid in the development of embedded processor architectures. The tool allows the developer to reach well-suited design decisions regarding the optimal processor hardware configuration for a given algorithm at an early stage in the design process.
Battery life is one of the major limitations to mobile device use, which makes research on energy efficient soft- and hardware an important task. This paper investigates the energy required by a CPU when decoding compressed bitstream videos on mobile platforms. A model is derived that describes the energy consumption of the new HEVC decoder for intra coded videos. We show that the relative estimation error of the model is smaller than 3.2% and that the model can be used to build encoders aiming at minimizing decoding energy.
The huge demand for computation in artificial intelligence (AI) is driving unparalleled investments in hardware and software systems for AI. This leads to an explosion in the number of specialized hardware devices, which are now offered by major cloud vendors. By hiding the low-level complexity through a tensor-based interface, tensor computation runtimes (TCRs) such as PyTorch allow data scientists to efficiently exploit the exciting capabilities offered by the new hardware. In this paper, we explore how database management systems can ride the wave of innovation happening in the AI space. We design, build, and evaluate Tensor Query Processor (TQP): TQP transforms SQL queries into tensor programs and executes them on TCRs. TQP is able to run the full TPC-H benchmark by implementing novel algorithms for relational operators on the tensor routines. At the same time, TQP can support various hardware while only requiring a fraction of the usual development effort. Experiments show that TQP can improve query execution time by up to 10$\times$ over specialized CPU- and GPU-only systems. Finally, TQP can accelerate queries mixing ML predictions and SQL end-to-end, and deliver up to 9$\times$ speedup over CPU baselines.
We present two methods of real time magnitude spectrogram inversion: streaming Griffin Lim(GL) and streaming MelGAN. We demonstrate the impact of looking ahead on perceptual quality of MelGAN. As little as one hop size (12.5ms) of lookahead is able to significantly improve perceptual quality in comparison to its causal version. We compare streaming GL with the streaming MelGAN and show different trade-offs in terms of perceptual quality, on-device latency, algorithmic delay, memory footprint and noise sensitivity. For fair quality assessment of the GL approach, we use input log magnitude spectrogram without mel transformation. We evaluate presented real time spectrogram inversion approaches on clean, noisy and atypical speech. We specified conditions when streaming GL has comparable quality with MelGAN: noisy audio and no mel transformation. Streaming GL is 2.4x faster than real time on the ARM CPU of a Pixel4 and it uses 4.5x times less memory than MelGAN.
Machine learning algorithms must be able to efficiently cope with massive data sets. Therefore, they have to scale well on any modern system and be able to exploit the computing power of accelerators independent of their vendor. In the field of supervised learning, Support Vector Machines (SVMs) are widely used. However, even modern and optimized implementations such as LIBSVM or ThunderSVM do not scale well for large non-trivial dense data sets on cutting-edge hardware: Most SVM implementations are based on Sequential Minimal Optimization, an optimized though inherent sequential algorithm. Hence, they are not well-suited for highly parallel GPUs. Furthermore, we are not aware of a performance portable implementation that supports CPUs and GPUs from different vendors. We have developed the PLSSVM library to solve both issues. First, we resort to the formulation of the SVM as a least squares problem. Training an SVM then boils down to solving a system of linear equations for which highly parallel algorithms are known. Second, we provide a hardware independent yet efficient implementation: PLSSVM uses different interchangeable backends--OpenMP, CUDA, OpenCL, SYCL--supporting modern hardware from various vendors like NVIDIA, AMD, or Intel on multiple GPUs. PLSSVM can be used as a drop-in replacement for LIBSVM. We observe a speedup on CPUs of up to 10 compared to LIBSVM and on GPUs of up to 14 compared to ThunderSVM. Our implementation scales on many-core CPUs with a parallel speedup of 74.7 on up to 256 CPU threads and on multiple GPUs with a parallel speedup of 3.71 on four GPUs. The code, utility scripts, and documentation are all available on GitHub: https://github.com/SC-SGS/PLSSVM.
Anomaly detection is one of the most active research areas in various critical domains, such as healthcare, fintech, and public security. However, little attention has been paid to scholarly data, i.e., anomaly detection in a citation network. Citation is considered as one of the most crucial metrics to evaluate the impact of scientific research, which may be gamed in multiple ways. Therefore, anomaly detection in citation networks is of significant importance to identify manipulation and inflation of citations. To address this open issue, we propose a novel deep graph learning model, namely GLAD (Graph Learning for Anomaly Detection), to identify anomalies in citation networks. GLAD incorporates text semantic mining to network representation learning by adding both node attributes and link attributes via graph neural networks. It exploits not only the relevance of citation contents but also hidden relationships between papers. Within the GLAD framework, we propose an algorithm called CPU (Citation PUrpose) to discover the purpose of citation based on citation texts. The performance of GLAD is validated through a simulated anomalous citation dataset. Experimental results demonstrate the effectiveness of GLAD on the anomalous citation detection task.
Recent studies in Learning to Rank have shown the possibility to effectively distill a neural network from an ensemble of regression trees. This result leads neural networks to become a natural competitor of tree-based ensembles on the ranking task. Nevertheless, ensembles of regression trees outperform neural models both in terms of efficiency and effectiveness, particularly when scoring on CPU. In this paper, we propose an approach for speeding up neural scoring time by applying a combination of Distillation, Pruning and Fast Matrix multiplication. We employ knowledge distillation to learn shallow neural networks from an ensemble of regression trees. Then, we exploit an efficiency-oriented pruning technique that performs a sparsification of the most computationally-intensive layers of the neural network that is then scored with optimized sparse matrix multiplication. Moreover, by studying both dense and sparse high performance matrix multiplication, we develop a scoring time prediction model which helps in devising neural network architectures that match the desired efficiency requirements. Comprehensive experiments on two public learning-to-rank datasets show that neural networks produced with our novel approach are competitive at any point of the effectiveness-efficiency trade-off when compared with tree-based ensembles, providing up to 4x scoring time speed-up without affecting the ranking quality.
Deep Learning (DL) has shown impressive performance in many mobile applications. Most existing works have focused on reducing the computational and resource overheads of running Deep Neural Networks (DNN) inference on resource-constrained mobile devices. However, the other aspect of DNN operations, i.e. training (forward and backward passes) on smartphone GPUs, has received little attention thus far. To this end, we conduct an initial analysis to examine the feasibility of on-device training on smartphones using mobile GPUs. We first employ the open-source mobile DL framework (MNN) and its OpenCL backend for running compute kernels on GPUs. Next, we observed that training on CPUs is much faster than on GPUs and identified two possible bottlenecks related to this observation: (i) computation and (ii) memory bottlenecks. To solve the computation bottleneck, we optimize the OpenCL backend's kernels, showing 2x improvements (40-70 GFLOPs) over CPUs (15-30 GFLOPs) on the Snapdragon 8 series processors. However, we find that the full DNN training is still much slower on GPUs than on CPUs, indicating that memory bottleneck plays a significant role in the lower performance of GPU over CPU. The data movement takes almost 91% of training time due to the low bandwidth. Lastly, based on the findings and failures during our investigation, we present limitations and practical guidelines for future directions.
We propose an efficient distributed out-of-memory implementation of the Non-negative Matrix Factorization (NMF) algorithm for heterogeneous high-performance-computing (HPC) systems. The proposed implementation is based on prior work on NMFk, which can perform automatic model selection and extract latent variables and patterns from data. In this work, we extend NMFk by adding support for dense and sparse matrix operation on multi-node, multi-GPU systems. The resulting algorithm is optimized for out-of-memory (OOM) problems where the memory required to factorize a given matrix is greater than the available GPU memory. Memory complexity is reduced by batching/tiling strategies, and sparse and dense matrix operations are significantly accelerated with GPU cores (or tensor cores when available). Input/Output (I/O) latency associated with batch copies between host and device is hidden using CUDA streams to overlap data transfers and compute asynchronously, and latency associated with collective communications (both intra-node and inter-node) is reduced using optimized NVIDIA Collective Communication Library NCCL based communicators. Benchmark results show significant improvement, from 32X to 76x speedup, with the new implementation using GPUs over the CPU-based NMFk. Good weak scaling was demonstrated on up to 4096 multi-GPU cluster nodes with approximately 25,000 GPUs when decomposing a dense 340 Terabyte-size matrix and an 11 Exabyte-size sparse matrix of density 10e-6.