PyTorchKR Papers

최신 논문

총 716편
태그·기간 등 상세 필터

카테고리 탐색

로그인 후 이 태그 구독

arXiv:2109.15098공식 코드 · Papers with Code

Deep Homography Estimation in Dynamic Surgical Scenes for Laparoscopic Camera Motion Extraction

Current laparoscopic camera motion automation relies on rule-based approaches or only focuses on surgical tools. Imitation Learning (IL) methods could alleviate these shortcomings, but have so far been applied to oversimplified setups. Instead of extracting actions from oversimplified setups, in this work we introduce a method that allows to extract a laparoscope holder's actions from videos of laparoscopic interventions. We synthetically add camera motion to a newly acquired dataset of camera motion free da Vinci surgery image sequences through a novel homography generation algorithm. The synthetic camera motion serves as a supervisory signal for camera motion estimation that is invariant to object and tool motion. We perform an extensive evaluation of state-of-the-art (SOTA) Deep Neural Networks (DNNs) across multiple compute regimes, finding our method transfers from our camera motion free da Vinci surgery dataset to videos of laparoscopic interventions, outperforming classical homography estimation approaches in both, precision by 41%, and runtime on a CPU by 43%.

arXiv:2105.01457공식 코드 · Papers with Code

CFEAR Radarodometry - Conservative Filtering for Efficient and Accurate Radar Odometry

This paper presents the accurate, highly efficient, and learning-free method CFEAR Radarodometry for large-scale radar odometry estimation. By using a filtering technique that keeps the k strongest returns per azimuth and by additionally filtering the radar data in Cartesian space, we are able to compute a sparse set of oriented surface points for efficient and accurate scan matching. Registration is carried out by minimizing a point-to-line metric and robustness to outliers is achieved using a Huber loss. We were able to additionally reduce drift by jointly registering the latest scan to a history of keyframes and found that our odometry method generalizes to different sensor models and datasets without changing a single parameter. We evaluate our method in three widely different environments and demonstrate an improvement over spatially cross-validated state-of-the-art with an overall translation error of 1.76% in a public urban radar odometry benchmark, running at 55Hz merely on a single laptop CPU thread.

arXiv:2109.12985공식 코드 · Papers with Code

Synerise at RecSys 2021: Twitter user engagement prediction with a fast neural model

In this paper we present our 2nd place solution to ACM RecSys 2021 Challenge organized by Twitter. The challenge aims to predict user engagement for a set of tweets, offering an exceptionally large data set of 1 billion data points sampled from over four weeks of real Twitter interactions. Each data point contains multiple sources of information, such as tweet text along with engagement features, user features, and tweet features. The challenge brings the problem close to a real production environment by introducing strict latency constraints in the model evaluation phase: the average inference time for single tweet engagement prediction is limited to 6ms on a single CPU core with 64GB memory. Our proposed model relies on extensive feature engineering performed with methods such as the Efficient Manifold Density Estimator (EMDE) - our previously introduced algorithm based on Locality Sensitive Hashing method, and novel Fourier Feature Encoding, among others. In total, we create numerous features describing a user's Twitter account status and the content of a tweet. In order to adhere to the strict latency constraints, the underlying model is a simple residual feed-forward neural network. The system is a variation of our previous methods which proved successful in KDD Cup 2021, WSDM Challenge 2021, and SIGIR eCom Challenge 2020. We release the source code at: https://github.com/Synerise/recsys-challenge-2021

arXiv:2109.12979공식 코드 · Papers with Code

CT-ICP: Real-time Elastic LiDAR Odometry with Loop Closure

Multi-beam LiDAR sensors are increasingly used in robotics, particularly with autonomous cars for localization and perception tasks, both relying on the ability to build a precise map of the environment. For this, we propose a new real-time LiDAR-only odometry method called CT-ICP (for Continuous-Time ICP), completed into a full SLAM with a novel loop detection procedure. The core of this method, is the introduction of the combined continuity in the scan matching, and discontinuity between scans. It allows both the elastic distortion of the scan during the registration for increased precision, and the increased robustness to high frequency motions from the discontinuity. We build a complete SLAM on top of this odometry, using a fast pure LiDAR loop detection based on elevation image 2D matching, providing a pose graph with loop constraints. To show the robustness of the method, we tested it on seven datasets: KITTI, KITTI-raw, KITTI-360, KITTI-CARLA, ParisLuco, Newer College, and NCLT in driving and high-frequency motion scenarios. Both the CT-ICP odometry and the loop detection are made available online. CT-ICP is currently first, among those giving access to a public code, on the KITTI odometry leaderboard, with an average Relative Translation Error (RTE) of 0.59% and an average time per scan of 60ms on a CPU with a single thread.

arXiv:2109.09495공식 코드 · Papers with Code

GhostShiftAddNet: More Features from Energy-Efficient Operations

Deep convolutional neural networks (CNNs) are computationally and memory intensive. In CNNs, intensive multiplication can have resource implications that may challenge the ability for effective deployment of inference on resource-constrained edge devices. This paper proposes GhostShiftAddNet, where the motivation is to implement a hardware-efficient deep network: a multiplication-free CNN with fewer redundant features. We introduce a new bottleneck block, GhostSA, that converts all multiplications in the block to cheap operations. The bottleneck uses an appropriate number of bit-shift filters to process intrinsic feature maps, then applies a series of transformations that consist of bit-wise shifts with addition operations to generate more feature maps that fully learn to capture information underlying intrinsic features. We schedule the number of bit-shift and addition operations for different hardware platforms. We conduct extensive experiments and ablation studies with desktop and embedded (Jetson Nano) devices for implementation and measurements. We demonstrate the proposed GhostSA block can replace bottleneck blocks in the backbone of state-of-the-art networks architectures and gives improved performance on image classification benchmarks. Further, our GhostShiftAddNet can achieve higher classification accuracy with fewer FLOPs and parameters (reduced by up to 3x) than GhostNet. When compared to GhostNet, inference latency on the Jetson Nano is improved by 1.3x and 2x on the GPU and CPU respectively. Code is available open-source on \url{https://github.com/JIABI/GhostShiftAddNet}.

arXiv:2109.04030공식 코드 · Papers with Code

Bag of Tricks for Optimizing Transformer Efficiency

Improving Transformer efficiency has become increasingly attractive recently. A wide range of methods has been proposed, e.g., pruning, quantization, new architectures and etc. But these methods are either sophisticated in implementation or dependent on hardware. In this paper, we show that the efficiency of Transformer can be improved by combining some simple and hardware-agnostic methods, including tuning hyper-parameters, better design choices and training strategies. On the WMT news translation tasks, we improve the inference efficiency of a strong Transformer system by 3.80X on CPU and 2.52X on GPU. The code is publicly available at https://github.com/Lollipop321/mini-decoder-network.

arXiv:2108.13976공식 코드 · Papers with Code

WarpDrive: Extremely Fast End-to-End Deep Multi-Agent Reinforcement Learning on a GPU

Deep reinforcement learning (RL) is a powerful framework to train decision-making models in complex environments. However, RL can be slow as it requires repeated interaction with a simulation of the environment. In particular, there are key system engineering bottlenecks when using RL in complex environments that feature multiple agents with high-dimensional state, observation, or action spaces. We present WarpDrive, a flexible, lightweight, and easy-to-use open-source RL framework that implements end-to-end deep multi-agent RL on a single GPU (Graphics Processing Unit), built on PyCUDA and PyTorch. Using the extreme parallelization capability of GPUs, WarpDrive enables orders-of-magnitude faster RL compared to common implementations that blend CPU simulations and GPU models. Our design runs simulations and the agents in each simulation in parallel. It eliminates data copying between CPU and GPU. It also uses a single simulation data store on the GPU that is safely updated in-place. WarpDrive provides a lightweight Python interface and flexible environment wrappers that are easy to use and extend. Together, this allows the user to easily run thousands of concurrent multi-agent simulations and train on extremely large batches of experience. Through extensive experiments, we verify that WarpDrive provides high-throughput and scales almost linearly to many agents and parallel environments. For example, WarpDrive yields 2.9 million environment steps/second with 2000 environments and 1000 agents (at least 100x higher throughput compared to a CPU implementation) in a benchmark Tag simulation. As such, WarpDrive is a fast and extensible multi-agent RL platform to significantly accelerate research and development.

arXiv:2108.10470공식 코드 · Papers with Code

Isaac Gym: High Performance GPU-Based Physics Simulation For Robot Learning

Isaac Gym offers a high performance learning platform to train policies for wide variety of robotics tasks directly on GPU. Both physics simulation and the neural network policy training reside on GPU and communicate by directly passing data from physics buffers to PyTorch tensors without ever going through any CPU bottlenecks. This leads to blazing fast training times for complex robotics tasks on a single GPU with 2-3 orders of magnitude improvements compared to conventional RL training that uses a CPU based simulator and GPU for neural networks. We host the results and videos at \url{https://sites.google.com/view/isaacgym-nvidia} and isaac gym can be downloaded at \url{https://developer.nvidia.com/isaac-gym}.

arXiv:2108.08513공식 코드 · Papers with Code

Fast Passage Re-ranking with Contextualized Exact Term Matching and Efficient Passage Expansion

BERT-based information retrieval models are expensive, in both time (query latency) and computational resources (energy, hardware cost), making many of these models impractical especially under resource constraints. The reliance on a query encoder that only performs tokenization and on the pre-processing of passage representations at indexing, has allowed the recently proposed TILDE method to overcome the high query latency issue typical of BERT-based models. This however is at the expense of a lower effectiveness compared to other BERT-based re-rankers and dense retrievers. In addition, the original TILDE method is characterised by indexes with a very high memory footprint, as it expands each passage into the size of the BERT vocabulary. In this paper, we propose TILDEv2, a new model that stems from the original TILDE but that addresses its limitations. TILDEv2 relies on contextualized exact term matching with expanded passages. This requires to only store in the index the score of tokens that appear in the expanded passages (rather than all the vocabulary), thus producing indexes that are 99% smaller than those of TILDE. This matching mechanism also improves ranking effectiveness by 24%, without adding to the query latency. This makes TILDEv2 the state-of-the-art passage re-ranking method for CPU-only environments, capable of maintaining query latency below 100ms on commodity hardware.

arXiv:2108.08189공식 코드 · Papers with Code

FOX-NAS: Fast, On-device and Explainable Neural Architecture Search

Neural architecture search can discover neural networks with good performance, and One-Shot approaches are prevalent. One-Shot approaches typically require a supernet with weight sharing and predictors that predict the performance of architecture. However, the previous methods take much time to generate performance predictors thus are inefficient. To this end, we propose FOX-NAS that consists of fast and explainable predictors based on simulated annealing and multivariate regression. Our method is quantization-friendly and can be efficiently deployed to the edge. The experiments on different hardware show that FOX-NAS models outperform some other popular neural network architectures. For example, FOX-NAS matches MobileNetV2 and EfficientNet-Lite0 accuracy with 240% and 40% less latency on the edge CPU. FOX-NAS is the 3rd place winner of the 2020 Low-Power Computer Vision Challenge (LPCVC), DSP classification track. See all evaluation results at https://lpcv.ai/competitions/2020. Search code and pre-trained models are released at https://github.com/great8nctu/FOX-NAS.

arXiv:2108.07049공식 코드 · Papers with Code

Towards Efficient and Data Agnostic Image Classification Training Pipeline for Embedded Systems

Nowadays deep learning-based methods have achieved a remarkable progress at the image classification task among a wide range of commonly used datasets (ImageNet, CIFAR, SVHN, Caltech 101, SUN397, etc.). SOTA performance on each of the mentioned datasets is obtained by careful tuning of the model architecture and training tricks according to the properties of the target data. Although this approach allows setting academic records, it is unrealistic that an average data scientist would have enough resources to build a sophisticated training pipeline for every image classification task he meets in practice. This work is focusing on reviewing the latest augmentation and regularization methods for the image classification and exploring ways to automatically choose some of the most important hyperparameters: total number of epochs, initial learning rate value and it's schedule. Having a training procedure equipped with a lightweight modern CNN architecture (like bileNetV3 or EfficientNet), sufficient level of regularization and adaptive to data learning rate schedule, we can achieve a reasonable performance on a variety of downstream image classification tasks without manual tuning of parameters to each particular task. Resulting models are computationally efficient and can be deployed to CPU using the OpenVINO toolkit. Source code is available as a part of the OpenVINO Training Extensions (https://github.com/openvinotoolkit/training_extensions).

arXiv:2108.05818공식 코드 · Papers with Code

PatrickStar: Parallel Training of Pre-trained Models via Chunk-based Memory Management

The pre-trained model (PTM) is revolutionizing Artificial Intelligence (AI) technology. However, the hardware requirement of PTM training is prohibitively high, making it a game for a small proportion of people. Therefore, we proposed PatrickStar system to lower the hardware requirements of PTMs and make them accessible to everyone. PatrickStar uses the CPU-GPU heterogeneous memory space to store the model data. Different from existing works, we organize the model data in memory chunks and dynamically distribute them in the heterogeneous memory. Guided by the runtime memory statistics collected in a warm-up iteration, chunks are orchestrated efficiently in heterogeneous memory and generate lower CPU-GPU data transmission volume and higher bandwidth utilization. Symbiosis with the Zero Redundancy Optimizer, PatrickStar scales to multiple GPUs on multiple nodes. % using data parallelism. The system can train tasks on bigger models and larger batch sizes, which cannot be accomplished by existing works. Experimental results show that PatrickStar extends model scales 2.27 and 2.5 times of DeepSpeed, and consistently exhibits significantly higher execution speed. PatricStar also successfully runs the 175B GPT3 training task on a 32 GPU cluster. Our code is publicly available at https://github.com/Tencent/PatrickStar.

arXiv:2108.00644공식 코드 · Papers with Code

Jointly Optimizing Query Encoder and Product Quantization to Improve Retrieval Performance

Recently, Information Retrieval community has witnessed fast-paced advances in Dense Retrieval (DR), which performs first-stage retrieval with embedding-based search. Despite the impressive ranking performance, previous studies usually adopt brute-force search to acquire candidates, which is prohibitive in practical Web search scenarios due to its tremendous memory usage and time cost. To overcome these problems, vector compression methods have been adopted in many practical embedding-based retrieval applications. One of the most popular methods is Product Quantization (PQ). However, although existing vector compression methods including PQ can help improve the efficiency of DR, they incur severely decayed retrieval performance due to the separation between encoding and compression. To tackle this problem, we present JPQ, which stands for Joint optimization of query encoding and Product Quantization. It trains the query encoder and PQ index jointly in an end-to-end manner based on three optimization strategies, namely ranking-oriented loss, PQ centroid optimization, and end-to-end negative sampling. We evaluate JPQ on two publicly available retrieval benchmarks. Experimental results show that JPQ significantly outperforms popular vector compression methods. Compared with previous DR models that use brute-force search, JPQ almost matches the best retrieval performance with 30x compression on index size. The compressed index further brings 10x speedup on CPU and 2x speedup on GPU in query latency.

arXiv:2107.06397공식 코드 · Papers with Code

SurgeonAssist-Net: Towards Context-Aware Head-Mounted Display-Based Augmented Reality for Surgical Guidance

We present SurgeonAssist-Net: a lightweight framework making action-and-workflow-driven virtual assistance, for a set of predefined surgical tasks, accessible to commercially available optical see-through head-mounted displays (OST-HMDs). On a widely used benchmark dataset for laparoscopic surgical workflow, our implementation competes with state-of-the-art approaches in prediction accuracy for automated task recognition, and yet requires 7.4x fewer parameters, 10.2x fewer floating point operations per second (FLOPS), is 7.0x faster for inference on a CPU, and is capable of near real-time performance on the Microsoft HoloLens 2 OST-HMD. To achieve this, we make use of an efficient convolutional neural network (CNN) backbone to extract discriminative features from image data, and a low-parameter recurrent neural network (RNN) architecture to learn long-term temporal dependencies. To demonstrate the feasibility of our approach for inference on the HoloLens 2 we created a sample dataset that included video of several surgical tasks recorded from a user-centric point-of-view. After training, we deployed our model and cataloged its performance in an online simulated surgical scenario for the prediction of the current surgical task. The utility of our approach is explored in the discussion of several relevant clinical use-cases. Our code is publicly available at https://github.com/doughtmw/surgeon-assist-net.

arXiv:2107.05050공식 코드 · Papers with Code

Neural Waveshaping Synthesis

We present the Neural Waveshaping Unit (NEWT): a novel, lightweight, fully causal approach to neural audio synthesis which operates directly in the waveform domain, with an accompanying optimisation (FastNEWT) for efficient CPU inference. The NEWT uses time-distributed multilayer perceptrons with periodic activations to implicitly learn nonlinear transfer functions that encode the characteristics of a target timbre. Once trained, a NEWT can produce complex timbral evolutions by simple affine transformations of its input and output signals. We paired the NEWT with a differentiable noise synthesiser and reverb and found it capable of generating realistic musical instrument performances with only 260k total model parameters, conditioned on F0 and loudness features. We compared our method to state-of-the-art benchmarks with a multi-stimulus listening test and the Fr\'echet Audio Distance and found it performed competitively across the tested timbral domains. Our method significantly outperformed the benchmarks in terms of generation speed, and achieved real-time performance on a consumer CPU, both with and without FastNEWT, suggesting it is a viable basis for future creative sound design tools.

arXiv:2107.01779공식 코드 · Papers with Code

Depth Quality-Inspired Feature Manipulation for Efficient RGB-D Salient Object Detection

RGB-D salient object detection (SOD) recently has attracted increasing research interest by benefiting conventional RGB SOD with extra depth information. However, existing RGB-D SOD models often fail to perform well in terms of both efficiency and accuracy, which hinders their potential applications on mobile devices and real-world problems. An underlying challenge is that the model accuracy usually degrades when the model is simplified to have few parameters. To tackle this dilemma and also inspired by the fact that depth quality is a key factor influencing the accuracy, we propose a novel depth quality-inspired feature manipulation (DQFM) process, which is efficient itself and can serve as a gating mechanism for filtering depth features to greatly boost the accuracy. DQFM resorts to the alignment of low-level RGB and depth features, as well as holistic attention of the depth stream to explicitly control and enhance cross-modal fusion. We embed DQFM to obtain an efficient light-weight model called DFM-Net, where we also design a tailored depth backbone and a two-stage decoder for further efficiency consideration. Extensive experimental results demonstrate that our DFM-Net achieves state-of-the-art accuracy when comparing to existing non-efficient models, and meanwhile runs at 140ms on CPU (2.2$\times$ faster than the prior fastest efficient model) with only $\sim$8.5Mb model size (14.9% of the prior lightest). Our code will be available at https://github.com/zwbx/DFM-Net.

arXiv:2106.15335공식 코드 · Papers with Code

Privacy Budget Scheduling

Machine learning (ML) models trained on personal data have been shown to leak information about users. Differential privacy (DP) enables model training with a guaranteed bound on this leakage. Each new model trained with DP increases the bound on data leakage and can be seen as consuming part of a global privacy budget that should not be exceeded. This budget is a scarce resource that must be carefully managed to maximize the number of successfully trained models. We describe PrivateKube, an extension to the popular Kubernetes datacenter orchestrator that adds privacy as a new type of resource to be managed alongside other traditional compute resources, such as CPU, GPU, and memory. The abstractions we design for the privacy resource mirror those defined by Kubernetes for traditional resources, but there are also major differences. For example, traditional compute resources are replenishable while privacy is not: a CPU can be regained after a model finishes execution while privacy budget cannot. This distinction forces a re-design of the scheduler. We present DPF (Dominant Private Block Fairness) -- a variant of the popular Dominant Resource Fairness (DRF) algorithm -- that is geared toward the non-replenishable privacy resource but enjoys similar theoretical properties as DRF. We evaluate PrivateKube and DPF on microbenchmarks and an ML workload on Amazon Reviews data. Compared to existing baselines, DPF allows training more models under the same global privacy guarantee. This is especially true for DPF over R\'enyi DP, a highly composable form of DP.

arXiv:2106.08150공식 코드 · Papers with Code

MetaCache-GPU: Ultra-Fast Metagenomic Classification

The cost of DNA sequencing has dropped exponentially over the past decade, making genomic data accessible to a growing number of scientists. In bioinformatics, localization of short DNA sequences (reads) within large genomic sequences is commonly facilitated by constructing index data structures which allow for efficient querying of substrings. Recent metagenomic classification pipelines annotate reads with taxonomic labels by analyzing their $k$-mer histograms with respect to a reference genome database. CPU-based index construction is often performed in a preprocessing phase due to the relatively high cost of building irregular data structures such as hash maps. However, the rapidly growing amount of available reference genomes establishes the need for index construction and querying at interactive speeds. In this paper, we introduce MetaCache-GPU -- an ultra-fast metagenomic short read classifier specifically tailored to fit the characteristics of CUDA-enabled accelerators. Our approach employs a novel hash table variant featuring efficient minhash fingerprinting of reads for locality-sensitive hashing and their rapid insertion using warp-aggregated operations. Our performance evaluation shows that MetaCache-GPU is able to build large reference databases in a matter of seconds, enabling instantaneous operability, while popular CPU-based tools such as Kraken2 require over an hour for index construction on the same data. In the context of an ever-growing number of reference genomes, MetaCache-GPU is the first metagenomic classifier that makes analysis pipelines with on-demand composition of large-scale reference genome sets practical. The source code is publicly available at https://github.com/muellan/metacache .

arXiv:2106.07551공식 코드 · Papers with Code

MALib: A Parallel Framework for Population-based Multi-agent Reinforcement Learning

Population-based multi-agent reinforcement learning (PB-MARL) refers to the series of methods nested with reinforcement learning (RL) algorithms, which produces a self-generated sequence of tasks arising from the coupled population dynamics. By leveraging auto-curricula to induce a population of distinct emergent strategies, PB-MARL has achieved impressive success in tackling multi-agent tasks. Despite remarkable prior arts of distributed RL frameworks, PB-MARL poses new challenges for parallelizing the training frameworks due to the additional complexity of multiple nested workloads between sampling, training and evaluation involved with heterogeneous policy interactions. To solve these problems, we present MALib, a scalable and efficient computing framework for PB-MARL. Our framework is comprised of three key components: (1) a centralized task dispatching model, which supports the self-generated tasks and scalable training with heterogeneous policy combinations; (2) a programming architecture named Actor-Evaluator-Learner, which achieves high parallelism for both training and sampling, and meets the evaluation requirement of auto-curriculum learning; (3) a higher-level abstraction of MARL training paradigms, which enables efficient code reuse and flexible deployments on different distributed computing paradigms. Experiments on a series of complex tasks such as multi-agent Atari Games show that MALib achieves throughput higher than 40K FPS on a single machine with $32$ CPU cores; 5x speedup than RLlib and at least 3x speedup than OpenSpiel in multi-agent training tasks. MALib is publicly available at https://github.com/sjtu-marl/malib.

arXiv:2106.06882공식 코드 · Papers with Code

Sparse PointPillars: Maintaining and Exploiting Input Sparsity to Improve Runtime on Embedded Systems

Bird's Eye View (BEV) is a popular representation for processing 3D point clouds, and by its nature is fundamentally sparse. Motivated by the computational limitations of mobile robot platforms, we create a fast, high-performance BEV 3D object detector that maintains and exploits this input sparsity to decrease runtimes over non-sparse baselines and avoids the tradeoff between pseudoimage area and runtime. We present results on KITTI, a canonical 3D detection dataset, and Matterport-Chair, a novel Matterport3D-derived chair detection dataset from scenes in real furnished homes. We evaluate runtime characteristics using a desktop GPU, an embedded ML accelerator, and a robot CPU, demonstrating that our method results in significant detection speedups (2X or more) for embedded systems with only a modest decrease in detection quality. Our work represents a new approach for practitioners to optimize models for embedded systems by maintaining and exploiting input sparsity throughout their entire pipeline to reduce runtime and resource usage while preserving detection performance.