Transformer-based pre-trained language models have significantly improved the performance of various natural language processing (NLP) tasks in the recent years. While effective and prevalent, these models are usually prohibitively large for resource-limited deployment scenarios. A thread of research has thus been working on applying network pruning techniques under the pretrain-then-finetune paradigm widely adopted in NLP. However, the existing pruning results on benchmark transformers, such as BERT, are not as remarkable as the pruning results in the literature of convolutional neural networks (CNNs). In particular, common wisdom in pruning CNN states that sparse pruning technique compresses a model more than that obtained by reducing number of channels and layers (Elsen et al., 2020; Zhu and Gupta, 2017), while existing works on sparse pruning of BERT yields inferior results than its small-dense counterparts such as TinyBERT (Jiao et al., 2020). In this work, we aim to fill this gap by studying how knowledge are transferred and lost during the pre-train, fine-tune, and pruning process, and proposing a knowledge-aware sparse pruning process that achieves significantly superior results than existing literature. We show for the first time that sparse pruning compresses a BERT model significantly more than reducing its number of channels and layers. Experiments on multiple data sets of GLUE benchmark show that our method outperforms the leading competitors with a 20-times weight/FLOPs compression and neglectable loss in prediction accuracy.
Cloth-Changing person re-identification (CC-ReID) aims at matching the same person across different locations over a long-duration, e.g., over days, and therefore inevitably meets challenge of changing clothing. In this paper, we focus on handling well the CC-ReID problem under a more challenging setting, i.e., just from a single image, which enables high-efficiency and latency-free pedestrian identify for real-time surveillance applications. Specifically, we introduce Gait recognition as an auxiliary task to drive the Image ReID model to learn cloth-agnostic representations by leveraging personal unique and cloth-independent gait information, we name this framework as GI-ReID. GI-ReID adopts a two-stream architecture that consists of a image ReID-Stream and an auxiliary gait recognition stream (Gait-Stream). The Gait-Stream, that is discarded in the inference for high computational efficiency, acts as a regulator to encourage the ReID-Stream to capture cloth-invariant biometric motion features during the training. To get temporal continuous motion cues from a single image, we design a Gait Sequence Prediction (GSP) module for Gait-Stream to enrich gait information. Finally, a high-level semantics consistency over two streams is enforced for effective knowledge regularization. Experiments on multiple image-based Cloth-Changing ReID benchmarks, e.g., LTCC, PRCC, Real28, and VC-Clothes, demonstrate that GI-ReID performs favorably against the state-of-the-arts. Codes are available at https://github.com/jinx-USTC/GI-ReID.
In this paper, a deep reinforcement learning (DRL)-based approach to the Lyapunov optimization is considered to minimize the time-average penalty while maintaining queue stability. A proper construction of state and action spaces is provided to form a proper Markov decision process (MDP) for the Lyapunov optimization. A condition for the reward function of reinforcement learning (RL) for queue stability is derived. Based on the analysis and practical RL with reward discounting, a class of reward functions is proposed for the DRL-based approach to the Lyapunov optimization. The proposed DRL-based approach to the Lyapunov optimization does not required complicated optimization at each time step and operates with general non-convex and discontinuous penalty functions. Hence, it provides an alternative to the conventional drift-plus-penalty (DPP) algorithm for the Lyapunov optimization. The proposed DRL-based approach is applied to resource allocation in edge computing systems with queue stability and numerical results demonstrate its successful operation.
The fish target detection algorithm lacks a good quality data set, and the algorithm achieves real-time detection with lower power consumption on embedded devices, and it is difficult to balance the calculation speed and identification ability. To this end, this paper collected and annotated a data set named "Aquarium Fish" of 84 fishes containing 10042 images, and based on this data set, proposed a multi-scale input fast fish target detection network (BTP-yoloV3) and its optimization method. The experiment uses Depthwise convolution to redesign the backbone of the yoloV4 network, which reduces the amount of calculation by 94.1%, and the test accuracy is 92.34%. Then, the training model is enhanced with MixUp, CutMix, and mosaic to increase the test accuracy by 1.27%; Finally, use the mish, swish, and ELU activation functions to increase the test accuracy by 0.76%. As a result, the accuracy of testing the network with 2000 fish images reached 94.37%, and the computational complexity of the network BFLOPS was only 5.47. Comparing the YoloV3~4, MobileNetV2-yoloV3, and YoloV3-tiny networks of migration learning on this data set. The results show that BTP-Yolov3 has smaller model parameters, faster calculation speed, and lower energy consumption during operation while ensuring the calculation accuracy. It provides a certain reference value for the practical application of neural network.
One of the most well-established applications of machine learning is in deciding what content to show website visitors. When observation data comes from high-velocity, user-generated data streams, machine learning methods perform a balancing act between model complexity, training time, and computational costs. Furthermore, when model freshness is critical, the training of models becomes time-constrained. Parallelized batch offline training, although horizontally scalable, is often not time-considerate or cost-effective. In this paper, we propose Lambda Learner, a new framework for training models by incremental updates in response to mini-batches from data streams. We show that the resulting model of our framework closely estimates a periodically updated model trained on offline data and outperforms it when model updates are time-sensitive. We provide theoretical proof that the incremental learning updates improve the loss-function over a stale batch model. We present a large-scale deployment on the sponsored content platform for a large social network, serving hundreds of millions of users across different channels (e.g., desktop, mobile). We address challenges and complexities from both algorithms and infrastructure perspectives, and illustrate the system details for computation, storage, and streaming production of training data.
We consider the infinite-horizon, discrete-time full-information control problem. Motivated by learning theory, as a criterion for controller design we focus on regret, defined as the difference between the LQR cost of a causal controller (that has only access to past and current disturbances) and the LQR cost of a clairvoyant one (that has also access to future disturbances). In the full-information setting, there is a unique optimal non-causal controller that in terms of LQR cost dominates all other controllers. Since the regret itself is a function of the disturbances, we consider the worst-case regret over all possible bounded energy disturbances, and propose to find a causal controller that minimizes this worst-case regret. The resulting controller has the interpretation of guaranteeing the smallest possible regret compared to the best non-causal controller, no matter what the future disturbances are. We show that the regret-optimal control problem can be reduced to a Nehari problem, i.e., to approximate an anticausal operator with a causal one in the operator norm. In the state-space setting, explicit formulas for the optimal regret and for the regret-optimal controller (in both the causal and the strictly causal settings) are derived. The regret-optimal controller is the sum of the classical $H_2$ state-feedback law and a finite-dimensional controller obtained from the Nehari problem. The controller construction simply requires the solution to the standard LQR Riccati equation, in addition to two Lyapunov equations. Simulations over a range of plants demonstrates that the regret-optimal controller interpolates nicely between the $H_2$ and the $H_\infty$ optimal controllers, and generally has $H_2$ and $H_\infty$ costs that are simultaneously close to their optimal values. The regret-optimal controller thus presents itself as a viable option for control system design.
Semantic image synthesis, translating semantic layouts to photo-realistic images, is a one-to-many mapping problem. Though impressive progress has been recently made, diverse semantic synthesis that can efficiently produce semantic-level multimodal results, still remains a challenge. In this paper, we propose a novel diverse semantic image synthesis framework from the perspective of semantic class distributions, which naturally supports diverse generation at semantic or even instance level. We achieve this by modeling class-level conditional modulation parameters as continuous probability distributions instead of discrete values, and sampling per-instance modulation parameters through instance-adaptive stochastic sampling that is consistent across the network. Moreover, we propose prior noise remapping, through linear perturbation parameters encoded from paired references, to facilitate supervised training and exemplar-based instance style control at test time. Extensive experiments on multiple datasets show that our method can achieve superior diversity and comparable quality compared to state-of-the-art methods. Code will be available at \url{https://github.com/tzt101/INADE.git}
Adequate blood supply is critical for normal brain function. Brain vasculature dysfunctions such as stalled blood flow in cerebral capillaries are associated with cognitive decline and pathogenesis in Alzheimer's disease. Recent advances in imaging technology enabled generation of high-quality 3D images that can be used to visualize stalled blood vessels. However, localization of stalled vessels in 3D images is often required as the first step for downstream analysis, which can be tedious, time-consuming and error-prone, when done manually. Here, we describe a deep learning-based approach for automatic detection of stalled capillaries in brain images based on 3D convolutional neural networks. Our networks employed custom 3D data augmentations and were used weight transfer from pre-trained 2D models for initialization. We used an ensemble of several 3D models to produce the winning submission to the Clog Loss: Advance Alzheimer's Research with Stall Catchers machine learning competition that challenged the participants with classifying blood vessels in 3D image stacks as stalled or flowing. In this setting, our approach outperformed other methods and demonstrated state-of-the-art results, achieving 0.85 Matthews correlation coefficient, 85% sensitivity, and 99.3% specificity. The source code for our solution is made publicly available.
Low-light image enhancement, such as recovering color and texture details from low-light images, is a complex and vital task. For automated driving, low-light scenarios will have serious implications for vision-based applications. To address this problem, we propose a real-time unsupervised generative adversarial network (GAN) containing multiple discriminators, i.e. a multi-scale discriminator, a texture discriminator, and a color discriminator. These distinct discriminators allow the evaluation of images from different perspectives. Further, considering that different channel features contain different information and the illumination is uneven in the image, we propose a feature fusion attention module. This module combines channel attention with pixel attention mechanisms to extract image features. Additionally, to reduce training time, we adopt a shared encoder for the generator and the discriminator. This makes the structure of the model more compact and the training more stable. Experiments indicate that our method is superior to the state-of-the-art methods in qualitative and quantitative evaluations, and significant improvements are achieved for both autopilot positioning and detection results.
Persistent homology (PH) is an approach to topological data analysis (TDA) that computes multi-scale topologically invariant properties of high-dimensional data that are robust to noise. While PH has revealed useful patterns across various applications, computational requirements have limited applications to small data sets of a few thousand points. We present Dory, an efficient and scalable algorithm that can compute the persistent homology of large data sets. Dory uses significantly less memory than published algorithms and also provides significant reductions in the computation time compared to most algorithms. It scales to process data sets with millions of points. As an application, we compute the PH of the human genome at high resolution as revealed by a genome-wide Hi-C data set. Results show that the topology of the human genome changes significantly upon treatment with auxin, a molecule that degrades cohesin, corroborating the hypothesis that cohesin plays a crucial role in loop formation in DNA.