最近，视觉变压器（VIT）在计算机视野中连续建立了新的里程碑，而高计算和内存成本使其在工业生产中的传播困难。修剪是一种用于硬件效率的传统模型压缩范例，已广泛应用于各种DNN结构。尽管如此，它含糊不清，如何在vit结构上进行独家修剪。考虑三个关键点：结构特征，VITS的内部数据模式和相关边缘设备部署，我们利用输入令牌稀疏性并提出了一种计算感知软修剪框架，可以在扁平的vanilla变压器上设置。和CNN型结构，例如基于池的Vit（坑）。更具体地说，我们设计了一种基于动态关注的多头令牌选择器，它是一个轻量级模块，用于自适应实例 - 明智令牌选择。我们进一步引入了一种软修剪技术，它将选择器模块生成的较少的信息令牌集成到将参与后续计算的包令牌，而不是完全丢弃。我们的框架通过我们所提出的计算感知培训策略，我们通过特定边缘设备的准确性和计算限制之间的权衡。实验结果表明，我们的框架显着降低了VIT的计算成本，同时在图像分类上保持了可比性。此外，我们的框架可以保证所识别的模型，以满足移动设备和FPGA的资源规范，甚至在移动平台上实现DEIT-T的实时执行。例如，我们的方法在移动设备上减少了DEIT-T至26毫秒的延迟（26％$ \ SIM 41％的41％），在移动设备上，在0.25％$ \ sim $ 4％的ImageNet上的前1个精度高出4％。我们的代码即将发布。

由不同形状和非线性形状变化引起的机器官的大变形，对医学图像配准产生了重大挑战。传统的注册方法需要通过特定变形模型迭代地优化目标函数以及细致的参数调谐，但在具有大变形的图像中具有有限的能力。虽然基于深度学习的方法可以从输入图像到它们各自的变形字段中的复杂映射，但它是基于回归的，并且容易被卡在局部最小值，特别是当涉及大变形时。为此，我们呈现随机策划者 - 演员 - 评论家（SPAC），这是一种新的加强学习框架，可以执行逐步登记。关键概念通过每次步骤连续地翘曲运动图像，以最终与固定图像对齐。考虑到在传统的强化学习（RL）框架中处理高维连续动作和状态空间有挑战性，我们向标准演员 - 评论家模型引入了一个新的概念“计划”，这是低维度，可以促进演员生成易于高维行动。整个框架基于无监督的培训，并以端到端的方式运行。我们在几个2D和3D医学图像数据集上评估我们的方法，其中一些包含大变形。我们的经验结果强调了我们的工作实现了一致，显着的收益和优于最先进的方法。

训练无模型的深度加强学习模型来解决图像到图像转换是困难的，因为它涉及高维连续状态和动作空间。在本文中，我们借鉴了最近的最大熵增强学习框架成功的灵感来设计用于挑战连续控制问题，在包括图像表示，产生和控制的高维连续空间上开发随机策略。这种方法的核心是随机演员 - 执行程序 - 批评者 - 评论家（SAEC），这是一个违法的演员 - 评论家模型，具有额外的excator来生成现实图像。具体地，该actor通过随机潜行动作侧重于高级表示和控制策略，以及明确地指示执行器生成用于操纵状态的低级动作。关于若干图像到图像转换任务的实验已经证明了在面对高维连续空间问题时所提出的SAEC的有效性和稳健性。

Consensus clustering aggregates partitions in order to find a better fit by reconciling clustering results from different sources/executions. In practice, there exist noise and outliers in clustering task, which, however, may significantly degrade the performance. To address this issue, we propose a novel algorithm -- robust consensus clustering that can find common ground truth among experts' opinions, which tends to be minimally affected by the bias caused by the outliers. In particular, we formalize the robust consensus clustering problem as a constraint optimization problem, and then derive an effective algorithm upon alternating direction method of multipliers (ADMM) with rigorous convergence guarantee. Our method outperforms the baselines on benchmarks. We apply the proposed method to the real-world advertising campaign segmentation and forecasting tasks using the proposed consensus clustering results based on the similarity computed via Kolmogorov-Smirnov Statistics. The accurate clustering result is helpful for building the advertiser profiles so as to perform the forecasting.

In computational advertising, a challenging problem is how to recommend the bid for advertisers to achieve the best return on investment (ROI) given budget constraint. This paper presents a bid recommendation scenario that discovers the concavity changes in click prediction curves. The recommended bid is derived based on the turning point from significant increase (i.e. concave downward) to slow increase (convex upward). Parametric learning based method is applied by solving the corresponding constraint optimization problem. Empirical studies on real-world advertising scenarios clearly demonstrate the performance gains for business metrics (including revenue increase, click increase and advertiser ROI increase).

In cost-per-click (CPC) or cost-per-impression (CPM) advertising campaigns, advertisers always run the risk of spending the budget without getting enough conversions. Moreover, the bidding on advertising inventory has few connections with propensity one that can reach to target cost-per-acquisition (tCPA) goals. To address this problem, this paper presents a bid optimization scenario to achieve the desired tCPA goals for advertisers. In particular, we build the optimization engine to make a decision by solving the rigorously formalized constrained optimization problem, which leverages the bid landscape model learned from rich historical auction data using non-parametric learning. The proposed model can naturally recommend the bid that meets the advertisers' expectations by making inference over advertisers' historical auction behaviors, which essentially deals with the data challenges commonly faced by bid landscape modeling: incomplete logs in auctions, and uncertainty due to the variation and fluctuations in advertising bidding behaviors. The bid optimization model outperforms the baseline methods on real-world campaigns, and has been applied into a wide range of scenarios for performance improvement and revenue liftup.

We propose a new neural network design paradigm Reversible Column Network (RevCol). The main body of RevCol is composed of multiple copies of subnetworks, named columns respectively, between which multi-level reversible connections are employed. Such architectural scheme attributes RevCol very different behavior from conventional networks: during forward propagation, features in RevCol are learned to be gradually disentangled when passing through each column, whose total information is maintained rather than compressed or discarded as other network does. Our experiments suggest that CNN-style RevCol models can achieve very competitive performances on multiple computer vision tasks such as image classification, object detection and semantic segmentation, especially with large parameter budget and large dataset. For example, after ImageNet-22K pre-training, RevCol-XL obtains 88.2% ImageNet-1K accuracy. Given more pre-training data, our largest model RevCol-H reaches 90.0% on ImageNet-1K, 63.8% APbox on COCO detection minival set, 61.0% mIoU on ADE20k segmentation. To our knowledge, it is the best COCO detection and ADE20k segmentation result among pure (static) CNN models. Moreover, as a general macro architecture fashion, RevCol can also be introduced into transformers or other neural networks, which is demonstrated to improve the performances in both computer vision and NLP tasks. We release code and models at https://github.com/megvii-research/RevCol

We address the theoretical and practical problems related to the trajectory generation and tracking control of tail-sitter UAVs. Theoretically, we focus on the differential flatness property with full exploitation of actual UAV aerodynamic models, which lays a foundation for generating dynamically feasible trajectory and achieving high-performance tracking control. We have found that a tail-sitter is differentially flat with accurate aerodynamic models within the entire flight envelope, by specifying coordinate flight condition and choosing the vehicle position as the flat output. This fundamental property allows us to fully exploit the high-fidelity aerodynamic models in the trajectory planning and tracking control to achieve accurate tail-sitter flights. Particularly, an optimization-based trajectory planner for tail-sitters is proposed to design high-quality, smooth trajectories with consideration of kinodynamic constraints, singularity-free constraints and actuator saturation. The planned trajectory of flat output is transformed to state trajectory in real-time with consideration of wind in environments. To track the state trajectory, a global, singularity-free, and minimally-parameterized on-manifold MPC is developed, which fully leverages the accurate aerodynamic model to achieve high-accuracy trajectory tracking within the whole flight envelope. The effectiveness of the proposed framework is demonstrated through extensive real-world experiments in both indoor and outdoor field tests, including agile SE(3) flight through consecutive narrow windows requiring specific attitude and with speed up to 10m/s, typical tail-sitter maneuvers (transition, level flight and loiter) with speed up to 20m/s, and extremely aggressive aerobatic maneuvers (Wingover, Loop, Vertical Eight and Cuban Eight) with acceleration up to 2.5g.

Traditional multilingual neural machine translation (MNMT) uses a single model to translate all directions. However, with the increasing scale of language pairs, simply using a single model for massive MNMT brings new challenges: parameter tension and large computations. In this paper, we revisit multi-way structures by assigning an individual branch for each language (group). Despite being a simple architecture, it is challenging to train de-centralized models due to the lack of constraints to align representations from all languages. We propose a localized training recipe to map different branches into a unified space, resulting in an efficient detachable model, Lego-MT. For a fair comparison, we collect data from OPUS and build the first large-scale open-source translation benchmark covering 7 language-centric data, each containing 445 language pairs. Experiments show that Lego-MT (1.2B) brings gains of more than 4 BLEU while outperforming M2M-100 (12B) (We will public all training data, models, and checkpoints)

Despite the surprising few-shot performance of in-context learning (ICL), it is still a common practice to randomly sample examples to serve as context. This paper advocates a new principle for ICL: self-adaptive in-context learning. The self-adaption mechanism is introduced to help each sample find an in-context example permutation (i.e., selection and ordering) that can derive the correct prediction, thus maximizing performance. To validate the effectiveness of self-adaptive ICL, we propose a general select-then-rank framework and instantiate it with new selection and ranking algorithms. Upon extensive evaluation on eight different NLP datasets, our self-adaptive ICL method achieves a 40% relative improvement over the common practice setting. Further analysis reveals the enormous potential of self-adaptive ICL that it might be able to close the gap between ICL and finetuning given more advanced algorithms. Our code is released to facilitate future research in this area: https://github.com/Shark-NLP/self-adaptive-ICL