We show that training a multi-headed self-attention-based deep network to predict deleted, information-dense 2-8 Hz speech modulations over a 1.5-second section of a speech utterance is an effective way to make machines learn to extract speech modulations using time-domain contextual information. Our work exhibits that, once trained on large volumes of unlabelled data, the outputs of the self-attention layers vary in time with a modulation peak at 4 Hz. These pre-trained layers can be used to initialize parts of an Automatic Speech Recognition system to reduce its reliance on labeled speech data greatly.
Few-shot learning is a challenging problem since only a few examples are provided to recognize a new class. Several recent studies exploit additional semantic information, e.g. text embeddings of class names, to address the issue of rare samples through combining semantic prototypes with visual prototypes. However, these methods still suffer from the spurious visual features learned from the rare support samples, resulting in limited benefits. In this paper, we propose a novel Semantic Prompt (SP) approach for few-shot learning. Instead of the naive exploitation of semantic information for remedying classifiers, we explore leveraging semantic information as prompts to tune the visual feature extraction network adaptively. Specifically, we design two complementary mechanisms to insert semantic prompts into the feature extractor: one is to enable the interaction between semantic prompts and patch embeddings along the spatial dimension via self-attention, another is to supplement visual features with the transformed semantic prompts along the channel dimension. By combining these two mechanisms, the feature extractor presents a better ability to attend to the class-specific features and obtains more generalized image representations with merely a few support samples. Through extensive experiments on four datasets, the proposed approach achieves promising results, improving the 1-shot learning accuracy by 3.67% on average.
Visual simultaneous localization and mapping (SLAM) systems face challenges in detecting loop closure under the circumstance of large viewpoint changes. In this paper, we present an object-based loop closure detection method based on the spatial layout and semanic consistency of the 3D scene graph. Firstly, we propose an object-level data association approach based on the semantic information from semantic labels, intersection over union (IoU), object color, and object embedding. Subsequently, multi-view bundle adjustment with the associated objects is utilized to jointly optimize the poses of objects and cameras. We represent the refined objects as a 3D spatial graph with semantics and topology. Then, we propose a graph matching approach to select correspondence objects based on the structure layout and semantic property similarity of vertices' neighbors. Finally, we jointly optimize camera trajectories and object poses in an object-level pose graph optimization, which results in a globally consistent map. Experimental results demonstrate that our proposed data association approach can construct more accurate 3D semantic maps, and our loop closure method is more robust than point-based and object-based methods in circumstances with large viewpoint changes.
Low-light image enhancement (LLIE) investigates how to improve illumination and produce normal-light images. The majority of existing methods improve low-light images via a global and uniform manner, without taking into account the semantic information of different regions. Without semantic priors, a network may easily deviate from a region's original color. To address this issue, we propose a novel semantic-aware knowledge-guided framework (SKF) that can assist a low-light enhancement model in learning rich and diverse priors encapsulated in a semantic segmentation model. We concentrate on incorporating semantic knowledge from three key aspects: a semantic-aware embedding module that wisely integrates semantic priors in feature representation space, a semantic-guided color histogram loss that preserves color consistency of various instances, and a semantic-guided adversarial loss that produces more natural textures by semantic priors. Our SKF is appealing in acting as a general framework in LLIE task. Extensive experiments show that models equipped with the SKF significantly outperform the baselines on multiple datasets and our SKF generalizes to different models and scenes well. The code is available at Semantic-Aware-Low-Light-Image-Enhancement.
The strength of machine learning models stems from their ability to learn complex function approximations from data; however, this strength also makes training deep neural networks challenging. Notably, the complex models tend to memorize the training data, which results in poor regularization performance on test data. The regularization techniques such as L1, L2, dropout, etc. are proposed to reduce the overfitting effect; however, they bring in additional hyperparameters tuning complexity. These methods also fall short when the inter-class similarity is high due to the underlying data distribution, leading to a less accurate model. In this paper, we present a novel approach to regularize the models by leveraging the information-rich latent embeddings and their high intra-class correlation. We create phantom embeddings from a subset of homogenous samples and use these phantom embeddings to decrease the inter-class similarity of instances in their latent embedding space. The resulting models generalize better as a combination of their embedding and regularize them without requiring an expensive hyperparameter search. We evaluate our method on two popular and challenging image classification datasets (CIFAR and FashionMNIST) and show how our approach outperforms the standard baselines while displaying better training behavior.
Diffusion probabilistic models have been successful in generating high-quality and diverse images. However, traditional models, whose input and output are high-resolution images, suffer from excessive memory requirements, making them less practical for edge devices. Previous approaches for generative adversarial networks proposed a patch-based method that uses positional encoding and global content information. Nevertheless, designing a patch-based approach for diffusion probabilistic models is non-trivial. In this paper, we resent a diffusion probabilistic model that generates images on a patch-by-patch basis. We propose two conditioning methods for a patch-based generation. First, we propose position-wise conditioning using one-hot representation to ensure patches are in proper positions. Second, we propose Global Content Conditioning (GCC) to ensure patches have coherent content when concatenated together. We evaluate our model qualitatively and quantitatively on CelebA and LSUN bedroom datasets and demonstrate a moderate trade-off between maximum memory consumption and generated image quality. Specifically, when an entire image is divided into 2 x 2 patches, our proposed approach can reduce the maximum memory consumption by half while maintaining comparable image quality.
Federated Learning (FL) is a novel machine learning framework, which enables multiple distributed devices cooperatively to train a shared model scheduled by a central server while protecting private data locally. However, the non-independent-and-identically-distributed (Non-IID) data samples and frequent communication across participants may significantly slow down the convergent rate and increase communication costs. To achieve fast convergence, we ameliorate the conventional local updating rule by introducing the aggregated gradients at each local update epoch, and propose an adaptive learning rate algorithm that further takes the deviation of local parameter and global parameter into consideration. The above adaptive learning rate design requires all clients' local information including the local parameters and gradients, which is challenging as there is no communication during the local update epochs. To obtain a decentralized adaptive learning rate for each client, we utilize the mean field approach by introducing two mean field terms to estimate the average local parameters and gradients respectively, which does not require the clients to exchange their local information with each other at each local epoch. Numerical results show that our proposed framework is superior to the state-of-art FL schemes in both model accuracy and convergent rate for IID and Non-IID datasets.
Searching on bipartite graphs is basal and versatile to many real-world Web applications, e.g., online recommendation, database retrieval, and query-document searching. Given a query node, the conventional approaches rely on the similarity matching with the vectorized node embeddings in the continuous Euclidean space. To efficiently manage intensive similarity computation, developing hashing techniques for graph structured data has recently become an emerging research direction. Despite the retrieval efficiency in Hamming space, prior work is however confronted with catastrophic performance decay. In this work, we investigate the problem of hashing with Graph Convolutional Network on bipartite graphs for effective Top-N search. We propose an end-to-end Bipartite Graph Convolutional Hashing approach, namely BGCH, which consists of three novel and effective modules: (1) adaptive graph convolutional hashing, (2) latent feature dispersion, and (3) Fourier serialized gradient estimation. Specifically, the former two modules achieve the substantial retention of the structural information against the inevitable information loss in hash encoding; the last module develops Fourier Series decomposition to the hashing function in the frequency domain mainly for more accurate gradient estimation. The extensive experiments on six real-world datasets not only show the performance superiority over the competing hashing-based counterparts, but also demonstrate the effectiveness of all proposed model components contained therein.
Breast cancer is a significant public health concern and early detection is critical for triaging high risk patients. Sequential screening mammograms can provide important spatiotemporal information about changes in breast tissue over time. In this study, we propose a deep learning architecture called RADIFUSION that utilizes sequential mammograms and incorporates a linear image attention mechanism, radiomic features, a new gating mechanism to combine different mammographic views, and bilateral asymmetry-based finetuning for breast cancer risk assessment. We evaluate our model on a screening dataset called Cohort of Screen-Aged Women (CSAW) dataset. Based on results obtained on the independent testing set consisting of 1,749 women, our approach achieved superior performance compared to other state-of-the-art models with area under the receiver operating characteristic curves (AUCs) of 0.905, 0.872 and 0.866 in the three respective metrics of 1-year AUC, 2-year AUC and > 2-year AUC. Our study highlights the importance of incorporating various deep learning mechanisms, such as image attention, radiomic features, gating mechanism, and bilateral asymmetry-based fine-tuning, to improve the accuracy of breast cancer risk assessment. We also demonstrate that our model's performance was enhanced by leveraging spatiotemporal information from sequential mammograms. Our findings suggest that RADIFUSION can provide clinicians with a powerful tool for breast cancer risk assessment.
We study human pose estimation in extremely low-light images. This task is challenging due to the difficulty of collecting real low-light images with accurate labels, and severely corrupted inputs that degrade prediction quality significantly. To address the first issue, we develop a dedicated camera system and build a new dataset of real low-light images with accurate pose labels. Thanks to our camera system, each low-light image in our dataset is coupled with an aligned well-lit image, which enables accurate pose labeling and is used as privileged information during training. We also propose a new model and a new training strategy that fully exploit the privileged information to learn representation insensitive to lighting conditions. Our method demonstrates outstanding performance on real extremely low light images, and extensive analyses validate that both of our model and dataset contribute to the success.