Few-shot learning aims to adapt knowledge learned from previous tasks to novel tasks with only a limited amount of labeled data. Research literature on few-shot learning exhibits great diversity, while different algorithms often excel at different fe
w-shot learning scenarios. It is therefore tricky to decide which learning strategies to use under different task conditions. Inspired by the recent success in Automated Machine Learning literature (AutoML), in this paper, we present Meta Navigator, a framework that attempts to solve the aforementioned limitation in few-shot learning by seeking a higher-level strategy and proffer to automate the selection from various few-shot learning designs. The goal of our work is to search for good parameter adaptation policies that are applied to different stages in the network for few-shot classification. We present a search space that covers many popular few-shot learning algorithms in the literature and develop a differentiable searching and decoding algorithm based on meta-learning that supports gradient-based optimization. We demonstrate the effectiveness of our searching-based method on multiple benchmark datasets. Extensive experiments show that our approach significantly outperforms baselines and demonstrates performance advantages over many state-of-the-art methods. Code and models will be made publicly available.
Large-scale, pre-trained language models (LMs) have achieved human-level performance on a breadth of language understanding tasks. However, evaluations only based on end task performance shed little light on machines true ability in language understa
nding and reasoning. In this paper, we highlight the importance of evaluating the underlying reasoning process in addition to end performance. Toward this goal, we introduce Tiered Reasoning for Intuitive Physics (TRIP), a novel commonsense reasoning dataset with dense annotations that enable multi-tiered evaluation of machines reasoning process. Our empirical results show that while large LMs can achieve high end performance, they struggle to support their predictions with valid supporting evidence. The TRIP dataset and our baseline results will motivate verifiable evaluation of commonsense reasoning and facilitate future research toward developing better language understanding and reasoning models.
We use a semi-analytic galaxy formation model to study the co-evolution of supermassive black holes (SMBHs) with their host galaxies. Although the coalescence of SMBHs is not important, the quasar-mode accretion induced by mergers plays a dominant ro
le in the growth of SMBHs. Mergers play a more important role in the growth of SMBH host galaxies than in the SMBH growth. It is the combined contribution from quasar mode accretion and mergers to the SMBH growth and the combined contribution from starburst and mergers to their host galaxy growth that determine the observed scaling relation between the SMBH masses and their host galaxy masses. We also find that mergers are more important in the growth of SMBH host galaxies compared to normal galaxies which share the same stellar mass range as the SMBH host galaxies.
Generating pre-initial conditions (or particle loads) is the very first step to set up a cosmological N-body simulation. In this work, we revisit the numerical convergence of pre-initial conditions on dark matter halo properties using a set of simula
tions which only differs in initial particle loads, i.e. grid, glass, and the newly introduced capacity constrained Voronoi tessellation (CCVT). We find that the median halo properties agree fairly well (i.e. within a convergence level of a few per cent) among simulations running from different initial loads. We also notice that for some individual haloes cross-matched among different simulations, the relative difference of their properties sometimes can be several tens of per cent. By looking at the evolution history of these poorly converged haloes, we find that they are usually merging haloes or haloes have experienced recent merger events, and their merging processes in different simulations are out-of-sync, making the convergence of halo properties become poor temporarily. We show that, comparing to the simulation starting with an anisotropic grid load, the simulation with an isotropic CCVT load converges slightly better to the simulation with a glass load, which is also isotropic. Among simulations with different pre-initial conditions, haloes in higher density environments tend to have their properties converged slightly better. Our results confirm that CCVT loads behave as well as the widely used grid and glass loads at small scales, and for the first time we quantify the convergence of two independent isotropic particle loads (i.e. glass and CCVT) on halo properties.
In this manuscript, we introduce a semi-automatic scene graph annotation tool for images, the GeneAnnotator. This software allows human annotators to describe the existing relationships between participators in the visual scene in the form of directe
d graphs, hence enabling the learning and reasoning on visual relationships, e.g., image captioning, VQA and scene graph generation, etc. The annotations for certain image datasets could either be merged in a single VG150 data-format file to support most existing models for scene graph learning or transformed into a separated annotation file for each single image to build customized datasets. Moreover, GeneAnnotator provides a rule-based relationship recommending algorithm to reduce the heavy annotation workload. With GeneAnnotator, we propose Traffic Genome, a comprehensive scene graph dataset with 1000 diverse traffic images, which in return validates the effectiveness of the proposed software for scene graph annotation. The project source code, with usage examples and sample data is available at https://github.com/Milomilo0320/A-Semi-automatic-Annotation-Software-for-Scene-Graph, under the Apache open-source license.
Real-world visual recognition problems often exhibit long-tailed distributions, where the amount of data for learning in different categories shows significant imbalance. Standard classification models learned on such data distribution often make bia
sed predictions towards the head classes while generalizing poorly to the tail classes. In this paper, we present two effective modifications of CNNs to improve network learning from long-tailed distribution. First, we present a Class Activation Map Calibration (CAMC) module to improve the learning and prediction of network classifiers, by enforcing network prediction based on important image regions. The proposed CAMC module highlights the correlated image regions across data and reinforces the representations in these areas to obtain a better global representation for classification. Furthermore, we investigate the use of normalized classifiers for representation learning in long-tailed problems. Our empirical study demonstrates that by simply scaling the outputs of the classifier with an appropriate scalar, we can effectively improve the classification accuracy on tail classes without losing the accuracy of head classes. We conduct extensive experiments to validate the effectiveness of our design and we set new state-of-the-art performance on five benchmarks, including ImageNet-LT, Places-LT, iNaturalist 2018, CIFAR10-LT, and CIFAR100-LT.
We address the challenging task of few-shot segmentation in this work. It is essential for few-shot semantic segmentation to fully utilize the support information. Previous methods typically adapt masked average pooling over the support feature to ex
tract the support clues as a global vector, usually dominated by the salient part and loses some important clues. In this work, we argue that every support pixels information is desired to be transferred to all query pixels and propose a Correspondence Matching Network (CMNet) with an Optimal Transport Matching module to mine out the correspondence between the query and support images. Besides, it is important to fully utilize both local and global information from the annotated support images. To this end, we propose a Message Flow module to propagate the message along the inner-flow within the same image and cross-flow between support and query images, which greatly help enhance the local feature representations. We further address the few-shot segmentation as a multi-task learning problem to alleviate the domain gap issue between different datasets. Experiments on PASCAL VOC 2012, MS COCO, and FSS-1000 datasets show that our network achieves new state-of-the-art few-shot segmentation performance.
Development of deep learning systems for biomedical segmentation often requires access to expert-driven, manually annotated datasets. If more than a single expert is involved in the annotation of the same images, then the inter-expert agreement is no
t necessarily perfect, and no single expert annotation can precisely capture the so-called ground truth of the regions of interest on all images. Also, it is not trivial to generate a reference estimate using annotations from multiple experts. Here we present a deep neural network, defined as U-Net-and-a-half, which can simultaneously learn from annotations performed by multiple experts on the same set of images. U-Net-and-a-half contains a convolutional encoder to generate features from the input images, multiple decoders that allow simultaneous learning from image masks obtained from annotations that were independently generated by multiple experts, and a shared low-dimensional feature space. To demonstrate the applicability of our framework, we used two distinct datasets from digital pathology and radiology, respectively. Specifically, we trained two separate models using pathologist-driven annotations of glomeruli on whole slide images of human kidney biopsies (10 patients), and radiologist-driven annotations of lumen cross-sections of human arteriovenous fistulae obtained from intravascular ultrasound images (10 patients), respectively. The models based on U-Net-and-a-half exceeded the performance of the traditional U-Net models trained on single expert annotations alone, thus expanding the scope of multitask learning in the context of biomedical image segmentation.
Fundamental machine learning theory shows that different samples contribute unequally both in learning and testing processes. Contemporary studies on DNN imply that such sample difference is rooted on the distribution of intrinsic pattern information
, namely sample regularity. Motivated by the recent discovery on network memorization and generalization, we proposed a pair of sample regularity measures for both processes with a formulation-consistent representation. Specifically, cumulative binary training/generalizing loss (CBTL/CBGL), the cumulative number of correct classiffcations of the training/testing sample within training stage, is proposed to quantize the stability in memorization-generalization process; while forgetting/mal-generalizing events, i.e., the mis-classification of previously learned or generalized sample, are utilized to represent the uncertainty of sample regularity with respect to optimization dynamics. Experiments validated the effectiveness and robustness of the proposed approaches for mini-batch SGD optimization. Further applications on training/testing sample selection show the proposed measures sharing the unified computing procedure could benefit for both tasks.
In this work, we aim to address the challenging task of open set recognition (OSR). Many recent OSR methods rely on auto-encoders to extract class-specific features by a reconstruction strategy, requiring the network to restore the input image on pix
el-level. This strategy is commonly over-demanding for OSR since class-specific features are generally contained in target objects, not in all pixels. To address this shortcoming, here we discard the pixel-level reconstruction strategy and pay more attention to improving the effectiveness of class-specific feature extraction. We propose a mutual information-based method with a streamlined architecture, Maximal Mutual Information Open Set Recognition (M2IOSR). The proposed M2IOSR only uses an encoder to extract class-specific features by maximizing the mutual information between the given input and its latent features across multiple scales. Meanwhile, to further reduce the open space risk, latent features are constrained to class conditional Gaussian distributions by a KL-divergence loss function. In this way, a strong function is learned to prevent the network from mapping different observations to similar latent features and help the network extract class-specific features with desired statistical characteristics. The proposed method significantly improves the performance of baselines and achieves new state-of-the-art results on several benchmarks consistently.
