In contrastive self-supervised learning, the common way to learn discriminative representation is to pull different augmented "views" of the same image closer while pushing all other images further apart, which has been proven to be effective. However, it is unavoidable to construct undesirable views containing different semantic concepts during the augmentation procedure. It would damage the semantic consistency of representation to pull these augmentations closer in the feature space indiscriminately. In this study, we introduce feature-level augmentation and propose a novel semantics-consistent feature search (SCFS) method to mitigate this negative effect. The main idea of SCFS is to adaptively search semantics-consistent features to enhance the contrast between semantics-consistent regions in different augmentations. Thus, the trained model can learn to focus on meaningful object regions, improving the semantic representation ability. Extensive experiments conducted on different datasets and tasks demonstrate that SCFS effectively improves the performance of self-supervised learning and achieves state-of-the-art performance on different downstream tasks.

在各种设备上部署深度学习模型已成为一个重要的话题。硬件专业化的浪潮为多维张量计算带来了一套多样化的加速度原始图。这些新的加速原始基原料以及新兴的机器学习模型带来了巨大的工程挑战。在本文中，我们提出了Tensorir，这是一种编译器抽象，用于通过这些张量计算原始素优化程序。Tensorir概括了现有机器学习编译器中使用的循环巢表示，以将张量计算作为一流的公民。最后，我们在抽象之上构建了一个端到端框架，以自动优化给定的张量计算原始图的深度学习模型。实验结果表明，Tensorir编译会自动使用给定硬件后端的张量计算原始图，并提供与跨平台的最新手工精制系统竞争性能的性能。

多模型对现实世界应用的承诺激发了可视化和理解其内部力学的研究，其最终目标是使利益相关者能够可视化模型行为，执行模型调试并促进对机器学习模型的信任。但是，现代的多模型模型通常是黑盒神经网络，这使得了解其内部力学变得具有挑战性。我们如何能在这些模型中可视化多模式相互作用的内部建模？我们的论文旨在通过提出Multiviz来填补这一空白，这是一种通过将可解释性问题分为4个阶段来分析多模型模型行为的方法：（1）单峰的重要性：每种模式如何有助于下游建模和预测，（2）交叉交叉。 - 模式相互作用：不同模态如何相互关系，（3）多模式表示：如何在决策级特征中表示单峰和跨模式的交互作用，以及（4）多模式预测：决策级特征如何组成以制造一个预言。 Multiviz旨在在不同的模式，模型，任务和研究领域进行操作。通过对6个现实世界任务的8个训练模型的实验，我们表明，Multiviz中的互补阶段共同使用户能够（1）模拟模型预测，（2）将可解释的概念分配给功能，（3）对模型错误分析执行错误分析，（4）使用错误分析到调试模型的见解。 Multiviz公开可用，将定期使用新的解释工具和指标进行更新，并欢迎社区的意见。

Masked image modeling (MIM) performs strongly in pre-training large vision Transformers (ViTs). However, small models that are critical for real-world applications cannot or only marginally benefit from this pre-training approach. In this paper, we explore distillation techniques to transfer the success of large MIM-based pre-trained models to smaller ones. We systematically study different options in the distillation framework, including distilling targets, losses, input, network regularization, sequential distillation, etc, revealing that: 1) Distilling token relations is more effective than CLS token- and feature-based distillation; 2) An intermediate layer of the teacher network as target perform better than that using the last layer when the depth of the student mismatches that of the teacher; 3) Weak regularization is preferred; etc. With these findings, we achieve significant fine-tuning accuracy improvements over the scratch MIM pre-training on ImageNet-1K classification, using all the ViT-Tiny, ViT-Small, and ViT-base models, with +4.2%/+2.4%/+1.4% gains, respectively. Our TinyMIM model of base size achieves 52.2 mIoU in AE20K semantic segmentation, which is +4.1 higher than the MAE baseline. Our TinyMIM model of tiny size achieves 79.6% top-1 accuracy on ImageNet-1K image classification, which sets a new record for small vision models of the same size and computation budget. This strong performance suggests an alternative way for developing small vision Transformer models, that is, by exploring better training methods rather than introducing inductive biases into architectures as in most previous works. Code is available at https://github.com/OliverRensu/TinyMIM.

Existing automated techniques for software documentation typically attempt to reason between two main sources of information: code and natural language. However, this reasoning process is often complicated by the lexical gap between more abstract natural language and more structured programming languages. One potential bridge for this gap is the Graphical User Interface (GUI), as GUIs inherently encode salient information about underlying program functionality into rich, pixel-based data representations. This paper offers one of the first comprehensive empirical investigations into the connection between GUIs and functional, natural language descriptions of software. First, we collect, analyze, and open source a large dataset of functional GUI descriptions consisting of 45,998 descriptions for 10,204 screenshots from popular Android applications. The descriptions were obtained from human labelers and underwent several quality control mechanisms. To gain insight into the representational potential of GUIs, we investigate the ability of four Neural Image Captioning models to predict natural language descriptions of varying granularity when provided a screenshot as input. We evaluate these models quantitatively, using common machine translation metrics, and qualitatively through a large-scale user study. Finally, we offer learned lessons and a discussion of the potential shown by multimodal models to enhance future techniques for automated software documentation.

Text clustering and topic extraction are two important tasks in text mining. Usually, these two tasks are performed separately. For topic extraction to facilitate clustering, we can first project texts into a topic space and then perform a clustering algorithm to obtain clusters. To promote topic extraction by clustering, we can first obtain clusters with a clustering algorithm and then extract cluster-specific topics. However, this naive strategy ignores the fact that text clustering and topic extraction are strongly correlated and follow a chicken-and-egg relationship. Performing them separately fails to make them mutually benefit each other to achieve the best overall performance. In this paper, we propose an unsupervised text clustering and topic extraction framework (ClusTop) which integrates text clustering and topic extraction into a unified framework and can achieve high-quality clustering result and extract topics from each cluster simultaneously. Our framework includes four components: enhanced language model training, dimensionality reduction, clustering and topic extraction, where the enhanced language model can be viewed as a bridge between clustering and topic extraction. On one hand, it provides text embeddings with a strong cluster structure which facilitates effective text clustering; on the other hand, it pays high attention on the topic related words for topic extraction because of its self-attention architecture. Moreover, the training of enhanced language model is unsupervised. Experiments on two datasets demonstrate the effectiveness of our framework and provide benchmarks for different model combinations in this framework.

Cognitive Computing (COC) aims to build highly cognitive machines with low computational resources that respond in real-time. However, scholarly literature shows varying research areas and various interpretations of COC. This calls for a cohesive architecture that delineates the nature of COC. We argue that if Herbert Simon considered the design science is the science of artificial, cognitive systems are the products of cognitive science or 'the newest science of the artificial'. Therefore, building a conceptual basis for COC is an essential step into prospective cognitive computing-based systems. This paper proposes an architecture of COC through analyzing the literature on COC using a myriad of statistical analysis methods. Then, we compare the statistical analysis results with previous qualitative analysis results to confirm our findings. The study also comprehensively surveys the recent research on COC to identify the state of the art and connect the advances in varied research disciplines in COC. The study found that there are three underlaying computing paradigms, Von-Neuman, Neuromorphic Engineering and Quantum Computing, that comprehensively complement the structure of cognitive computation. The research discuss possible applications and open research directions under the COC umbrella.

Reading comprehension of legal text can be a particularly challenging task due to the length and complexity of legal clauses and a shortage of expert-annotated datasets. To address this challenge, we introduce the Merger Agreement Understanding Dataset (MAUD), an expert-annotated reading comprehension dataset based on the American Bar Association's 2021 Public Target Deal Points Study, with over 39,000 examples and over 47,000 total annotations. Our fine-tuned Transformer baselines show promising results, with models performing well above random on most questions. However, on a large subset of questions, there is still room for significant improvement. As the only expert-annotated merger agreement dataset, MAUD is valuable as a benchmark for both the legal profession and the NLP community.

Robotic teleoperation is a key technology for a wide variety of applications. It allows sending robots instead of humans in remote, possibly dangerous locations while still using the human brain with its enormous knowledge and creativity, especially for solving unexpected problems. A main challenge in teleoperation consists of providing enough feedback to the human operator for situation awareness and thus create full immersion, as well as offering the operator suitable control interfaces to achieve efficient and robust task fulfillment. We present a bimanual telemanipulation system consisting of an anthropomorphic avatar robot and an operator station providing force and haptic feedback to the human operator. The avatar arms are controlled in Cartesian space with a direct mapping of the operator movements. The measured forces and torques on the avatar side are haptically displayed to the operator. We developed a predictive avatar model for limit avoidance which runs on the operator side, ensuring low latency. The system was successfully evaluated during the ANA Avatar XPRIZE competition semifinals. In addition, we performed in lab experiments and carried out a small user study with mostly untrained operators.

In this work, we tackle two vital tasks in automated driving systems, i.e., driver intent prediction and risk object identification from egocentric images. Mainly, we investigate the question: what would be good road scene-level representations for these two tasks? We contend that a scene-level representation must capture higher-level semantic and geometric representations of traffic scenes around ego-vehicle while performing actions to their destinations. To this end, we introduce the representation of semantic regions, which are areas where ego-vehicles visit while taking an afforded action (e.g., left-turn at 4-way intersections). We propose to learn scene-level representations via a novel semantic region prediction task and an automatic semantic region labeling algorithm. Extensive evaluations are conducted on the HDD and nuScenes datasets, and the learned representations lead to state-of-the-art performance for driver intention prediction and risk object identification.