arXiv subjects

Yang You

Publications and source records attributed to Yang You.

At least 235 records · Page 13Linked to original sources

arXiv2020

Human Correspondence Consensus for 3D Object Semantic Understanding

Semantic understanding of 3D objects is crucial in many applications such as object manipulation. However, it is hard to give a universal definition of point-level semantics that everyone would agree on. We observe that people have a consensus on semantic correspondences between two areas from different objects, but are less certain about the exact semantic meaning of each area. Therefore, we argue that by providing human labeled correspondences between different objects from the same category instead of explicit semantic labels, one can recover rich semantic information of an object. In this paper, we introduce a new dataset named CorresPondenceNet. Based on this dataset, we are able to learn dense semantic embeddings with a novel geodesic consistency loss. Accordingly, several state-of-the-art networks are evaluated on this correspondence benchmark. We further show that CorresPondenceNet could not only boost fine-grained understanding of heterogeneous objects but also cross-object registration and partial object matching.

cs.CV↗

arXiv2020

Training EfficientNets at Supercomputer Scale: 83% ImageNet Top-1 Accuracy in One Hour

EfficientNets are a family of state-of-the-art image classification models based on efficiently scaled convolutional neural networks. Currently, EfficientNets can take on the order of days to train; for example, training an EfficientNet-B0 model takes 23 hours on a Cloud TPU v2-8 node. In this paper, we explore techniques to scale up the training of EfficientNets on TPU-v3 Pods with 2048 cores, motivated by speedups that can be achieved when training at such scales. We discuss optimizations required to scale training to a batch size of 65536 on 1024 TPU-v3 cores, such as selecting large batch optimizers and learning rate schedules as well as utilizing distributed evaluation and batch normalization techniques. Additionally, we present timing and performance benchmarks for EfficientNet models trained on the ImageNet dataset in order to analyze the behavior of EfficientNets at scale. With our optimizations, we are able to train EfficientNet on ImageNet to an accuracy of 83% in 1 hour and 4 minutes.

cs.LG↗

arXiv2020

How much progress have we made in neural network training? A New Evaluation Protocol for Benchmarking Optimizers

Many optimizers have been proposed for training deep neural networks, and they often have multiple hyperparameters, which make it tricky to benchmark their performance. In this work, we propose a new benchmarking protocol to evaluate both end-to-end efficiency (training a model from scratch without knowing the best hyperparameter) and data-addition training efficiency (the previously selected hyperparameters are used for periodically re-training the model with newly collected data). For end-to-end efficiency, unlike previous work that assumes random hyperparameter tuning, which over-emphasizes the tuning time, we propose to evaluate with a bandit hyperparameter tuning strategy. A human study is conducted to show that our evaluation protocol matches human tuning behavior better than the random search. For data-addition training, we propose a new protocol for assessing the hyperparameter sensitivity to data shift. We then apply the proposed benchmarking framework to 7 optimizers and various tasks, including computer vision, natural language processing, reinforcement learning, and graph mining. Our results show that there is no clear winner across all the tasks.

cs.LG↗

arXiv2020

The Clock and Control System for the ATLAS Liquid Argon Calorimeter Phase-I Upgrade

A Liquid-argon Trigger Digitizer Board (LTDB) is being developed to upgrade the ATLAS Liquid Argon Calorimeter Phase-I trigger electronics. The LTDB located at the front end needs to obtain the clock signals and be configured and monitored remotely from the back end. A clock and control system is being developed for the LTDB and the major functions of the system have been evaluated. The design and evaluation of the clock and control system are presented in this paper.

physics.ins-det↗

arXiv2020

LOCx2, a Low-latency, Low-overhead, 2 x 5.12-Gbps Transmitter ASIC for the ATLAS Liquid Argon Calorimeter Trigger Upgrade

In this paper, we present the design and test results of LOCx2, a transmitter ASIC for the ATLAS Liquid Argon Calorimeter trigger upgrade. LOCx2 consists of two channels and each channel encodes ADC data with an overhead of 14.3% and transmits serial data at 5.12 Gbps with a latency of less than 27.2 ns. LOCx2 is fabricated with a commercial 0.25-um Silicon-on-Sapphire CMOS technology and is packaged in a 100-pin QFN package. The power consumption of LOCx2 is about 843 mW.

physics.ins-det↗

arXiv2020

KeypointNet: A Large-scale 3D Keypoint Dataset Aggregated from Numerous Human Annotations

Detecting 3D objects keypoints is of great interest to the areas of both graphics and computer vision. There have been several 2D and 3D keypoint datasets aiming to address this problem in a data-driven way. These datasets, however, either lack scalability or bring ambiguity to the definition of keypoints. Therefore, we present KeypointNet: the first large-scale and diverse 3D keypoint dataset that contains 103,450 keypoints and 8,234 3D models from 16 object categories, by leveraging numerous human annotations. To handle the inconsistency between annotations from different people, we propose a novel method to aggregate these keypoints automatically, through minimization of a fidelity loss. Finally, ten state-of-the-art methods are benchmarked on our proposed dataset. Our code and data are available on https://github.com/qq456cvb/KeypointNet.

cs.CV↗

arXiv2020

The Limit of the Batch Size

Large-batch training is an efficient approach for current distributed deep learning systems. It has enabled researchers to reduce the ImageNet/ResNet-50 training from 29 hours to around 1 minute. In this paper, we focus on studying the limit of the batch size. We think it may provide a guidance to AI supercomputer and algorithm designers. We provide detailed numerical optimization instructions for step-by-step comparison. Moreover, it is important to understand the generalization and optimization performance of huge batch training. Hoffer et al. introduced "ultra-slow diffusion" theory to large-batch training. However, our experiments show contradictory results with the conclusion of Hoffer et al. We provide comprehensive experimental results and detailed analysis to study the limitations of batch size scaling and "ultra-slow diffusion" theory. For the first time we scale the batch size on ImageNet to at least a magnitude larger than all previous work, and provide detailed studies on the performance of many state-of-the-art optimization schemes under this setting. We propose an optimization recipe that is able to improve the top-1 test accuracy by 18% compared to the baseline.

cs.LG↗

arXiv2020

Large Batch Optimization for Deep Learning: Training BERT in 76 minutes

Training large deep neural networks on massive datasets is computationally very challenging. There has been recent surge in interest in using large batch stochastic optimization methods to tackle this issue. The most prominent algorithm in this line of research is LARS, which by employing layerwise adaptive learning rates trains ResNet on ImageNet in a few minutes. However, LARS performs poorly for attention models like BERT, indicating that its performance gains are not consistent across tasks. In this paper, we first study a principled layerwise adaptation strategy to accelerate training of deep neural networks using large mini-batches. Using this strategy, we develop a new layerwise adaptive large batch optimization technique called LAMB; we then provide convergence analysis of LAMB as well as LARS, showing convergence to a stationary point in general nonconvex settings. Our empirical results demonstrate the superior performance of LAMB across various tasks such as BERT and ResNet-50 training with very little hyperparameter tuning. In particular, for BERT training, our optimizer enables use of very large batch sizes of 32868 without any degradation of performance. By increasing the batch size to the memory limit of a TPUv3 Pod, BERT training time can be reduced from 3 days to just 76 minutes (Table 1). The LAMB implementation is available at https://github.com/tensorflow/addons/blob/master/tensorflow_addons/optimizers/lamb.py

cs.LG↗

arXiv2019

Pointwise Rotation-Invariant Network with Adaptive Sampling and 3D Spherical Voxel Convolution

Point cloud analysis without pose priors is very challenging in real applications, as the orientations of point clouds are often unknown. In this paper, we propose a brand new point-set learning framework PRIN, namely, Pointwise Rotation-Invariant Network, focusing on rotation-invariant feature extraction in point clouds analysis. We construct spherical signals by Density Aware Adaptive Sampling to deal with distorted point distributions in spherical space. In addition, we propose Spherical Voxel Convolution and Point Re-sampling to extract rotation-invariant features for each point. Our network can be applied to tasks ranging from object classification, part segmentation, to 3D feature matching and label alignment. We show that, on the dataset with randomly rotated point clouds, PRIN demonstrates better performance than state-of-the-art methods without any data augmentation. We also provide theoretical analysis for the rotation-invariance achieved by our methods.

cs.CV↗

arXiv2019

Combinational Q-Learning for Dou Di Zhu

Deep reinforcement learning (DRL) has gained a lot of attention in recent years, and has been proven to be able to play Atari games and Go at or above human levels. However, those games are assumed to have a small fixed number of actions and could be trained with a simple CNN network. In this paper, we study a special class of Asian popular card games called Dou Di Zhu, in which two adversarial groups of agents must consider numerous card combinations at each time step, leading to huge number of actions. We propose a novel method to handle combinatorial actions, which we call combinational Q-learning (CQL). We employ a two-stage network to reduce action space and also leverage order-invariant max-pooling operations to extract relationships between primitive actions. Results show that our method prevails over state-of-the art methods like naive Q-learning and A3C. We develop an easy-to-use card game environments and train all agents adversarially from sractch, with only knowledge of game rules and verify that our agents are comparative to humans. Our code to reproduce all reported results will be available online.

cs.LG↗

arXiv2019

Large-Batch Training for LSTM and Beyond

Large-batch training approaches have enabled researchers to utilize large-scale distributed processing and greatly accelerate deep-neural net (DNN) training. For example, by scaling the batch size from 256 to 32K, researchers have been able to reduce the training time of ResNet50 on ImageNet from 29 hours to 2.2 minutes (Ying et al., 2018). In this paper, we propose a new approach called linear-epoch gradual-warmup (LEGW) for better large-batch training. With LEGW, we are able to conduct large-batch training for both CNNs and RNNs with the Sqrt Scaling scheme. LEGW enables Sqrt Scaling scheme to be useful in practice and as a result we achieve much better results than the Linear Scaling learning rate scheme. For LSTM applications, we are able to scale the batch size by a factor of 64 without losing accuracy and without tuning the hyper-parameters. For CNN applications, LEGW is able to achieve the same accuracy even as we scale the batch size to 32K. LEGW works better than previous large-batch auto-tuning techniques. LEGW achieves a 5.3X average speedup over the baselines for four LSTM-based applications on the same hardware. We also provide some theoretical explanations for LEGW.

cs.LG↗

arXiv2018

Accurate, Fast and Scalable Kernel Ridge Regression on Parallel and Distributed Systems

We propose two new methods to address the weak scaling problems of KRR: the Balanced KRR (BKRR) and K-means KRR (KKRR). These methods consider alternative ways to partition the input dataset into p different parts, generating p different models, and then selecting the best model among them. Compared to a conventional implementation, KKRR2 (optimized version of KKRR) improves the weak scaling efficiency from 0.32% to 38% and achieves a 591times speedup for getting the same accuracy by using the same data and the same hardware (1536 processors). BKRR2 (optimized version of BKRR) achieves a higher accuracy than the current fastest method using less training time for a variety of datasets. For the applications requiring only approximate solutions, BKRR2 improves the weak scaling efficiency to 92% and achieves 3505 times speedup (theoretical speedup: 4096 times).

cs.DC↗

arXiv2018

ImageNet Training in Minutes

Finishing 90-epoch ImageNet-1k training with ResNet-50 on a NVIDIA M40 GPU takes 14 days. This training requires 10^18 single precision operations in total. On the other hand, the world's current fastest supercomputer can finish 2 * 10^17 single precision operations per second (Dongarra et al 2017, https://www.top500.org/lists/2017/06/). If we can make full use of the supercomputer for DNN training, we should be able to finish the 90-epoch ResNet-50 training in one minute. However, the current bottleneck for fast DNN training is in the algorithm level. Specifically, the current batch size (e.g. 512) is too small to make efficient use of many processors. For large-scale DNN training, we focus on using large-batch data-parallelism synchronous SGD without losing accuracy in the fixed epochs. The LARS algorithm (You, Gitman, Ginsburg, 2017, arXiv:1708.03888) enables us to scale the batch size to extremely large case (e.g. 32K). We finish the 100-epoch ImageNet training with AlexNet in 11 minutes on 1024 CPUs. About three times faster than Facebook's result (Goyal et al 2017, arXiv:1706.02677), we finish the 90-epoch ImageNet training with ResNet-50 in 20 minutes on 2048 KNLs without losing accuracy. State-of-the-art ImageNet training speed with ResNet-50 is 74.9% top-1 test accuracy in 15 minutes. We got 74.9% top-1 test accuracy in 64 epochs, which only needs 14 minutes. Furthermore, when we increase the batch size to above 16K, our accuracy is much higher than Facebook's on corresponding batch sizes. Our source code is available upon request.

cs.CV↗

arXiv2017

Large Batch Training of Convolutional Networks

A common way to speed up training of large convolutional networks is to add computational units. Training is then performed using data-parallel synchronous Stochastic Gradient Descent (SGD) with mini-batch divided between computational units. With an increase in the number of nodes, the batch size grows. But training with large batch size often results in the lower model accuracy. We argue that the current recipe for large batch training (linear learning rate scaling with warm-up) is not general enough and training may diverge. To overcome this optimization difficulties we propose a new training algorithm based on Layer-wise Adaptive Rate Scaling (LARS). Using LARS, we scaled Alexnet up to a batch size of 8K, and Resnet-50 to a batch size of 32K without loss in accuracy.

cs.CV↗

arXiv2017

Scaling Deep Learning on GPU and Knights Landing clusters

The speed of deep neural networks training has become a big bottleneck of deep learning research and development. For example, training GoogleNet by ImageNet dataset on one Nvidia K20 GPU needs 21 days. To speed up the training process, the current deep learning systems heavily rely on the hardware accelerators. However, these accelerators have limited on-chip memory compared with CPUs. To handle large datasets, they need to fetch data from either CPU memory or remote processors. We use both self-hosted Intel Knights Landing (KNL) clusters and multi-GPU clusters as our target platforms. From an algorithm aspect, current distributed machine learning systems are mainly designed for cloud systems. These methods are asynchronous because of the slow network and high fault-tolerance requirement on cloud systems. We focus on Elastic Averaging SGD (EASGD) to design algorithms for HPC clusters. Original EASGD used round-robin method for communication and updating. The communication is ordered by the machine rank ID, which is inefficient on HPC clusters. First, we redesign four efficient algorithms for HPC systems to improve EASGD's poor scaling on clusters. Async EASGD, Async MEASGD, and Hogwild EASGD are faster \textcolor{black}{than} their existing counterparts (Async SGD, Async MSGD, and Hogwild SGD, resp.) in all the comparisons. Finally, we design Sync EASGD, which ties for the best performance among all the methods while being deterministic. In addition to the algorithmic improvements, we use some system-algorithm codesign techniques to scale up the algorithms. By reducing the percentage of communication from 87% to 14%, our Sync EASGD achieves 5.3x speedup over original EASGD on the same platform. We get 91.5% weak scaling efficiency on 4253 KNL cores, which is higher than the state-of-the-art implementation.

cs.DC↗

arXiv2017

Resource Allocation for a Full-Duplex Base Station Aided OFDMA System

Exploiting full-duplex (FD) technology on base stations (BSs) is a promising solution to enhancing the system performance. Motivated by this, we revisit a full-duplex base station (FD-BS) aided OFDMA system, which consists of one BS, several uplink/downlink users and multiple subcarriers. A joint 3-dimensional (3D) mapping scheme among subcarriers, down-link users (DUEs), uplink users (UUEs) is considered as well as an associated power allocation optimization. In detail, we first decompose the complex 3D mapping problem into three 2-dimensional sub ones and solve them by using the iterative Hungarian method, respectively. Then based on the Lagrange dual method, we sequentially solve the power allocation and 3- dimensional mapping problem by fixing a dual point. Finally, the optimal solution can be obtained by utilizing the sub-gradient method. Unlike existing work that only solves either 3D mapping or power allocation problem but with a high computation complexity, we tackle both of them and have successfully reduced computation complexity from exponential to polynomial order. Numerical simulations are conducted to verify the proposed scheme.

cs.IT↗

arXiv2017

A Joint Resource Allocation Scheme for Multi-User Full-Duplex OFDMA Systems

Full-duplex is a promising technology and great system performance enhancement can be achieved especially when we apply it to base stations. In this paper, we consider a Multi-User OFDMA network consisting of one Full-duplex base station, several uplink users, and a number of downlink users. A joint subcarrier, downlink user and uplink user mapping and power allocation optimization problem is investigated. In detail, we first decompose the 3-dimensional mapping problem into three 2-dimensional sub-problems and solve them by iteratively using classical Hungarian Method. Then, based on the dual method, we sequentially solve the power allocation and 3-dimensional mapping problem at each dual point. And the optimal solution to the dual problem is derived by using sub-gradient method. Unlike existing methods that only solve one sub-problem but with a high computation complexity, we tackle both of them and successfully reducing computation complexity from exponential to polynomial order. Numerical simulations are conducted to verify the proposed system.

cs.IT↗

arXiv2016

A Vertical Channel Model of Molecular Communication based on Alcohol Molecules

The study of Molecular Communication(MC) is more and more prevalence, and channel model of MC plays an important role in the MC System. Since different propagation environment and modulation techniques produce different channel model, most of the research about MC are in horizontal direction,but in nature the communications between nano machines are in short range and some of the information transportation are in the vertical direction, such as transpiration of plants, biological pump in ocean, and blood transportation from heart to brain. Therefore, this paper we propose a vertical channel model which nano-machines communicate with each other in the vertical direction based on pure diffusion. We first propose a vertical molecular communication model, we mainly considered the gravity as the factor, though the channel model is also affected by other main factors, such as the flow of the medium, the distance between the transmitter and the receiver, the delay or sensitivity of the transmitter and the receiver. Secondly, we set up a test-bed for this vertical channel model, in order to verify the difference between the theory result and the experiment data. At last, we use the data we get from the experiment and the non-linear least squares method to get the parameters to make our channel model more accurate.

cs.ET↗