Visual crowd counting has been recently studied as a way to enable people counting in crowd scenes from images. Albeit successful, vision-based crowd counting approaches could fail to capture informative features in extreme conditions, e.g., imaging at night and occlusion. In this work, we introduce a novel task of audiovisual crowd counting, in which visual and auditory information are integrated for counting purposes. We collect a large-scale benchmark, named auDiovISual Crowd cOunting (DISCO) dataset, consisting of 1,935 images and the corresponding audio clips, and 170,270 annotated instances. In order to fuse the two modalities, we make use of a linear feature-wise fusion module that carries out an affine transformation on visual and auditory features. Finally, we conduct extensive experiments using the proposed dataset and approach. Experimental results show that introducing auditory information can benefit crowd counting under different illumination, noise, and occlusion conditions. The dataset and code will be released. Code and data have been made available
Softening labels of training datasets with respect to data representations has been frequently used to improve the training of deep neural networks (DNNs). While such a practice has been studied as a way to leverage privileged information about the distribution of the data, a well-trained learner with soft classification outputs should be first obtained as a prior to generate such privileged information. To solve such chicken-egg problem, we propose COLAM framework that Co-Learns DNNs and soft labels through Alternating Minimization of two objectives - (a) the training loss subject to soft labels and (b) the objective to learn improved soft labels - in one end-to-end training procedure. We performed extensive experiments to compare our proposed method with a series of baselines. The experiment results show that COLAM achieves improved performance on many tasks with better testing classification accuracy. We also provide both qualitative and quantitative analyses that explain why COLAM works well.
In this paper, we introduce a novel interpreting framework that learns an interpretable model based on an ontology-based sampling technique to explain agnostic prediction models. Different from existing approaches, our algorithm considers contextual correlation among words, described in domain knowledge ontologies, to generate semantic explanations. To narrow down the search space for explanations, which is a major problem of long and complicated text data, we design a learnable anchor algorithm, to better extract explanations locally. A set of regulations is further introduced, regarding combining learned interpretable representations with anchors to generate comprehensible semantic explanations. An extensive experiment conducted on two real-world datasets shows that our approach generates more precise and insightful explanations compared with baseline approaches.
Arbitrary image style transfer is a challenging task which aims to stylize a content image conditioned on an arbitrary style image. In this task the content-style feature transformation is a critical component for a proper fusion of features. Existing feature transformation algorithms often suffer from unstable learning, loss of content and style details, and non-natural stroke patterns. To mitigate these issues, this paper proposes a parameter-free algorithm, Style Projection, for fast yet effective content-style transformation. To leverage the proposed Style Projection~component, this paper further presents a real-time feed-forward model for arbitrary style transfer, including a regularization for matching the content semantics between inputs and outputs. Extensive experiments have demonstrated the effectiveness and efficiency of the proposed method in terms of qualitative analysis, quantitative evaluation, and user study.
Text classification is one of the most important and fundamental tasks in natural language processing. Performance of this task mainly dependents on text representation learning. Currently, most existing learning frameworks mainly focus on encoding local contextual information between words. These methods always neglect to exploit global clues, such as label information, for encoding text information. In this study, we propose a label-guided learning framework LguidedLearn for text representation and classification. Our method is novel but simple that we only insert a label-guided encoding layer into the commonly used text representation learning schemas. That label-guided layer performs label-based attentive encoding to map the universal text embedding (encoded by a contextual information learner) into different label spaces, resulting in label-wise embeddings. In our proposed framework, the label-guided layer can be easily and directly applied with a contextual encoding method to perform jointly learning. Text information is encoded based on both the local contextual information and the global label clues. Therefore, the obtained text embeddings are more robust and discriminative for text classification. Extensive experiments are conducted on benchmark datasets to illustrate the effectiveness of our proposed method.
Associating sound and its producer in complex audiovisual scene is a challenging task, especially when we are lack of annotated training data. In this paper, we present a flexible audiovisual model that introduces a soft-clustering module as the audio and visual content detector, and regards the pervasive property of audiovisual concurrency as the latent supervision for inferring the correlation among detected contents. To ease the difficulty of audiovisual learning, we propose a novel curriculum learning strategy that trains the model from simple to complex scene. We show that such ordered learning procedure rewards the model the merits of easy training and fast convergence. Meanwhile, our audiovisual model can also provide effective unimodal representation and cross-modal alignment performance. We further deploy the well-trained model into practical audiovisual sound localization and separation task. We show that our localization model significantly outperforms existing methods, based on which we show comparable performance in sound separation without referring external visual supervision. Our video demo can be found at https://youtu.be/kuClfGG0cFU.
Search space is a key consideration for neural architecture search. Recently, Xie et al. (2019) found that randomly generated networks from the same distribution perform similarly, which suggests we should search for random graph distributions instead of graphs. We propose graphon as a new search space. A graphon is the limit of Cauchy sequence of graphs and a scale-free probabilistic distribution, from which graphs of different number of nodes can be drawn. By utilizing properties of the graphon space and the associated cut-distance metric, we develop theoretically motivated techniques that search for and scale up small-capacity stage-wise graphs found on small datasets to large-capacity graphs that can handle ImageNet. The scaled stage-wise graphs outperform DenseNet and randomly wired Watts-Strogatz networks, indicating the benefits of graphon theory in NAS applications.
The success of deep neural networks is clouded by two issues that largely remain open to this day: the abundance of adversarial attacks that fool neural networks with small perturbations and the lack of interpretation for the predictions they make. Empirical evidence in the literature as well as theoretical analysis on simple models suggest these two seemingly disparate issues may actually be connected, as robust models tend to be more interpretable than non-robust models. In this paper, we provide evidence for the claim that this relationship is bidirectional. Viz., models that are forced to have interpretable gradients are more robust to adversarial examples than models trained in a standard manner. With further analysis and experiments, we identify two factors behind this phenomenon, namely the suppression of the gradient and the selective use of features guided by high-quality interpretations, which explain model behaviors under various regularization and target interpretation settings.
Definition Extraction (DE) is one of the well-known topics in Information Extraction that aims to identify terms and their corresponding definitions in unstructured texts. This task can be formalized either as a sentence classification task (i.e., containing term-definition pairs or not) or a sequential labeling task (i.e., identifying the boundaries of the terms and definitions). The previous works for DE have only focused on one of the two approaches, failing to model the inter-dependencies between the two tasks. In this work, we propose a novel model for DE that simultaneously performs the two tasks in a single framework to benefit from their inter-dependencies. Our model features deep learning architectures to exploit the global structures of the input sentences as well as the semantic consistencies between the terms and the definitions, thereby improving the quality of the representation vectors for DE. Besides the joint inference between sentence classification and sequential labeling, the proposed model is fundamentally different from the prior work for DE in that the prior work has only employed the local structures of the input sentences (i.e., word-to-word relations), and not yet considered the semantic consistencies between terms and definitions. In order to implement these novel ideas, our model presents a multi-task learning framework that employs graph convolutional neural networks and predicts the dependency paths between the terms and the definitions. We also seek to enforce the consistency between the representations of the terms and definitions both globally (i.e., increasing semantic consistency between the representations of the entire sentences and the terms/definitions) and locally (i.e., promoting the similarity between the representations of the terms and the definitions).