Scene graph generation (SGG) aims to predict graph-structured descriptions of input images, in the form of objects and relationships between them. This task is becoming increasingly useful for progress at the interface of vision and language. Here, it is important - yet challenging - to perform well on novel (zero-shot) or rare (few-shot) compositions of objects and relationships. In this paper, we identify two key issues that limit such generalization. Firstly, we show that the standard loss used in this task is unintentionally a function of scene graph density. This leads to the neglect of individual edges in large sparse graphs during training, even though these contain diverse few-shot examples that are important for generalization. Secondly, the frequency of relationships can create a strong bias in this task, such that a blind model predicting the most frequent relationship achieves good performance. Consequently, some state-of-the-art models exploit this bias to improve results. We show that such models can suffer the most in their ability to generalize to rare compositions, evaluating two different models on the Visual Genome dataset and its more recent, improved version, GQA. To address these issues, we introduce a density-normalized edge loss, which provides more than a two-fold improvement in certain generalization metrics. Compared to other works in this direction, our enhancements require only a few lines of code and no added computational cost. We also highlight the difficulty of accurately evaluating models using existing metrics, especially on zero/few shots, and introduce a novel weighted metric.
Human-computer interactive systems that rely on machine learning are becoming paramount to the lives of millions of people who use digital assistants on a daily basis. Yet, further advances are limited by the availability of data and the cost of acquiring new samples. One way to address this problem is by improving the sample efficiency of current approaches. As a solution path, we present a model-based reinforcement learning algorithm for an interactive dialogue task. We build on commonly used actor-critic methods, adding an environment model and planner that augments a learning agent to learn the model of the environment dynamics. Our results show that, on a simulation that mimics the interactive task, our algorithm requires 70 times fewer samples, compared to the baseline of commonly used model-free algorithm, and demonstrates 2~times better performance asymptotically. Moreover, we introduce a novel contribution of computing a soft planner policy and further updating a model-free policy yielding a less computationally expensive model-free agent as good as the model-based one. This model-based architecture serves as a foundation that can be extended to other human-computer interactive tasks allowing further advances in this direction.
We consider the problem of distance metric learning (DML), where the task is to learn an effective similarity measure between images. We revisit ProxyNCA and incorporate several enhancements. We find that low temperature scaling is a performance-critical component and explain why it works. Besides, we also discover that Global Max Pooling works better in general when compared to Global Average Pooling. Additionally, our proposed fast moving proxies also addresses small gradient issue of proxies, and this component synergizes well with low temperature scaling and Global Max Pooling. Our enhanced model, called ProxyNCA++, achieves a 22.9 percentage point average improvement of Recall@1 across four different zero-shot retrieval datasets compared to the original ProxyNCA algorithm. Furthermore, we achieve state-of-the-art results on the CUB200, Cars196, Sop, and InShop datasets, achieving Recall@1 scores of 72.2, 90.1, 81.4, and 90.9, respectively.
In Digital Pathology (DP), labeled data is generally very scarce due to the requirement that medical experts provide annotations. We address this issue by learning transferable features from weakly labeled data, which are collected from various parts of the body and are organized by non-medical experts. In this paper, we show that features learned from such weakly labeled datasets are indeed transferable and allow us to achieve highly competitive patch classification results on the colorectal cancer (CRC) dataset [1] and the PatchCamelyon (PCam) dataset [2] while using an order of magnitude less labeled data.
Deep neural networks require collecting and annotating large amounts of data to train successfully. In order to alleviate the annotation bottleneck, we propose a novel self-supervised representation learning approach for spatiotemporal features extracted from videos. We introduce Skip-Clip, a method that utilizes temporal coherence in videos, by training a deep model for future clip order ranking conditioned on a context clip as a surrogate objective for video future prediction. We show that features learned using our method are generalizable and transfer strongly to downstream tasks. For action recognition on the UCF101 dataset, we obtain 51.8% improvement over random initialization and outperform models initialized using inflated ImageNet parameters. Skip-Clip also achieves results competitive with state-of-the-art self-supervision methods.
As the availability and importance of temporal interaction data--such as email communication--increases, it becomes increasingly important to understand the underlying structure that underpins these interactions. Often these interactions form a multigraph, where we might have multiple interactions between two entities. Such multigraphs tend to be sparse yet structured, and their distribution often evolves over time. Existing statistical models with interpretable parameters can capture some, but not all, of these properties. We propose a dynamic nonparametric model for interaction multigraphs that combines the sparsity of edge-exchangeable multigraphs with dynamic clustering patterns that tend to reinforce recent behavioral patterns. We show that our method yields improved held-out likelihood over stationary variants, and impressive predictive performance against a range of state-of-the-art dynamic graph models.
Graphs evolving over time are a natural way to represent data in many domains, such as social networks, bioinformatics, physics and finance. Machine learning methods for graphs, which leverage such data for various prediction tasks, have seen a recent surge of interest and capability. In practice, ground truth edges between nodes in these graphs can be unknown or suboptimal, which hurts the quality of features propagated through the network. Building on recent progress in modeling temporal graphs and learning latent graphs, we extend two methods, Dynamic Representation (DyRep) and Neural Relational Inference (NRI), for the task of dynamic link prediction. We explore the effect of learning temporal attention edges using NRI without requiring the ground truth graph. In experiments on the Social Evolution dataset, we show semantic interpretability of learned attention, often outperforming the baseline DyRep model that uses a ground truth graph to compute attention. In addition, we consider functions acting on pairs of nodes, which are used to predict link or edge representations. We demonstrate that in all cases, our bilinear transformation is superior to feature concatenation, typically employed in prior work. Source code is available at https://github.com/uoguelph-mlrg/LDG.
Graph Convolutional Networks (GCNs) are a class of general models that can learn from graph structured data. Despite being general, GCNs are admittedly inferior to convolutional neural networks (CNNs) when applied to vision tasks, mainly due to the lack of domain knowledge that is hardcoded into CNNs, such as spatially oriented translation invariant filters. However, a great advantage of GCNs is the ability to work on irregular inputs, such as superpixels of images. This could significantly reduce the computational cost of image reasoning tasks. Another key advantage inherent to GCNs is the natural ability to model multirelational data. Building upon these two promising properties, in this work, we show best practices for designing GCNs for image classification; in some cases even outperforming CNNs on the MNIST, CIFAR-10 and PASCAL image datasets.
Conditional Generative Adversarial Networks (cGANs) are finding increasingly widespread use in many application domains. Despite outstanding progress, quantitative evaluation of such models often involves multiple distinct metrics to assess different desirable properties such as image quality, intra-conditioning diversity, and conditional consistency, making model benchmarking challenging. In this paper, we propose the Frechet Joint Distance (FJD), which implicitly captures the above mentioned properties in a single metric. FJD is defined as the Frechet Distance of the joint distribution of images and conditionings, making it less sensitive to the often limited per-conditioning sample size. As a result, it scales more gracefully to stronger forms of conditioning such as pixel-wise or multi-modal conditioning. We evaluate FJD on a modified version of the dSprite dataset as well as on the large scale COCO-Stuff dataset, and consistently highlight its benefits when compared to currently established metrics. Moreover, we use the newly introduced metric to compare existing cGAN-based models, with varying conditioning strengths, and show that FJD can be used as a promising single metric for model benchmarking.