Self-supervised learning (SSL) is currently one of the premier techniques to create data representations that are actionable for transfer learning in the absence of human annotations. Despite their success, the underlying geometry of these representations remains elusive, which obfuscates the quest for more robust, trustworthy, and interpretable models. In particular, mainstream SSL techniques rely on a specific deep neural network architecture with two cascaded neural networks: the encoder and the projector. When used for transfer learning, the projector is discarded since empirical results show that its representation generalizes more poorly than the encoder's. In this paper, we investigate this curious phenomenon and analyze how the strength of the data augmentation policies affects the data embedding. We discover a non-trivial relation between the encoder, the projector, and the data augmentation strength: with increasingly larger augmentation policies, the projector, rather than the encoder, is more strongly driven to become invariant to the augmentations. It does so by eliminating crucial information about the data by learning to project it into a low-dimensional space, a noisy estimate of the data manifold tangent plane in the encoder representation. This analysis is substantiated through a geometrical perspective with theoretical and empirical results.
Many application studies rely on audio DNN models pre-trained on a large-scale dataset as essential feature extractors, and they extract features from the last layers. In this study, we focus on our finding that the middle layer features of existing supervised pre-trained models are more effective than the late layer features for some tasks. We propose a simple approach to compose features effective for general-purpose applications, consisting of two steps: (1) calculating feature vectors along the time frame from middle/late layer outputs, and (2) fusing them. This approach improves the utility of frequency and channel information in downstream processes, and combines the effectiveness of middle and late layer features for different tasks. As a result, the feature vectors become effective for general purposes. In the experiments using VGGish, PANNs' CNN14, and AST on nine downstream tasks, we first show that each layer output of these models serves different tasks. Then, we demonstrate that the proposed approach significantly improves their performance and brings it to a level comparable to that of the state-of-the-art. In particular, the performance of the non-semantic speech (NOSS) tasks greatly improves, especially on Speech commands V2 with VGGish of +77.1 (14.3% to 91.4%).
Previous security research efforts orbiting around graphs have been exclusively focusing on either (de-)anonymizing the graphs or understanding the security and privacy issues of graph neural networks. Little attention has been paid to understand the privacy risks of integrating the output from graph embedding models (e.g., node embeddings) with complex downstream machine learning pipelines. In this paper, we fill this gap and propose a novel model-agnostic graph recovery attack that exploits the implicit graph structural information preserved in the embeddings of graph nodes. We show that an adversary can recover edges with decent accuracy by only gaining access to the node embedding matrix of the original graph without interactions with the node embedding models. We demonstrate the effectiveness and applicability of our graph recovery attack through extensive experiments.
The remote sensing community has identified data fusion as one of the key challenging topics of the 21st century. The subject of image fusion in two-dimensional (2D) space has been covered in several published reviews. However, the special case of 2.5D/3D Digital Elevation Model (DEM) fusion has not been addressed till date. DEM fusion is a key application of data fusion in remote sensing. It takes advantage of the complementary characteristics of multi-source DEMs to deliver a more complete, accurate and reliable elevation dataset. Although several methods for fusing DEMs have been developed, the absence of a well-rounded review has limited their proliferation among researchers and end-users. It is often required to combine knowledge from multiple studies to inform a holistic perspective and guide further research. In response, this paper provides a systematic review of DEM fusion: the pre-processing workflow, methods and applications, enhanced with a meta-analysis. Through the discussion and comparative analysis, unresolved challenges and open issues were identified, and future directions for research were proposed. This review is a timely solution and an invaluable source of information for researchers within the fields of remote sensing and spatial information science, and the data fusion community at large.
Given the high incidence and effective treatment options for liver diseases, they are of great socioeconomic importance. One of the most common methods for analyzing CT and MRI images for diagnosis and follow-up treatment is liver segmentation. Recent advances in deep learning have demonstrated encouraging results for automatic liver segmentation. Despite this, their success depends primarily on the availability of an annotated database, which is often not available because of privacy concerns. Federated Learning has been recently proposed as a solution to alleviate these challenges by training a shared global model on distributed clients without access to their local databases. Nevertheless, Federated Learning does not perform well when it is trained on a high degree of heterogeneity of image data due to multi-modal imaging, such as CT and MRI, and multiple scanner types. To this end, we propose Fednorm and its extension \fednormp, two Federated Learning algorithms that use a modality-based normalization technique. Specifically, Fednorm normalizes the features on a client-level, while Fednorm+ employs the modality information of single slices in the feature normalization. Our methods were validated using 428 patients from six publicly available databases and compared to state-of-the-art Federated Learning algorithms and baseline models in heterogeneous settings (multi-institutional, multi-modal data). The experimental results demonstrate that our methods show an overall acceptable performance, achieve Dice per patient scores up to 0.961, consistently outperform locally trained models, and are on par or slightly better than centralized models.
Video instance segmentation is a challenging task that extends image instance segmentation to the video domain. Existing methods either rely only on single-frame information for the detection and segmentation subproblems or handle tracking as a separate post-processing step, which limit their capability to fully leverage and share useful spatial-temporal information for all the subproblems. In this paper, we propose a novel graph-neural-network (GNN) based method to handle the aforementioned limitation. Specifically, graph nodes representing instance features are used for detection and segmentation while graph edges representing instance relations are used for tracking. Both inter and intra-frame information is effectively propagated and shared via graph updates and all the subproblems (i.e. detection, segmentation and tracking) are jointly optimized in an unified framework. The performance of our method shows great improvement on the YoutubeVIS validation dataset compared to existing methods and achieves 35.2% AP with a ResNet-50 backbone, operating at 22 FPS. Code is available at http://github.com/lucaswithai/visgraph.git .
Recognizing user location with WiFi fingerprints is a popular approach for accurate indoor positioning problems. In this work, our goal is to interpret WiFi fingerprints into actual user locations. However, WiFi fingerprint data can be very high dimensional in some cases, we need to find a good representation of the input data for the learning task first. Otherwise, using neural networks will suffer from severe overfitting. In this work, we solve this issue by combining the Information Bottleneck method and Variational Inference. Based on these two approaches, we propose a Variational Information Bottleneck model for accurate indoor positioning. The proposed model consists of an encoder structure and a predictor structure. The encoder is to find a good representation in the input data for the learning task. The predictor is to use the latent representation to predict the final output. To enhance the generalization of our model, we also adopt the Dropout technique for each hidden layer of the decoder. We conduct the validation experiments on a real-world dataset. We also compare the proposed model to other existing methods so as to quantify the performances of our method.
Stylized image captioning as presented in prior work aims to generate captions that reflect characteristics beyond a factual description of the scene composition, such as sentiments. Such prior work relies on given sentiment identifiers, which are used to express a certain global style in the caption, e.g. positive or negative, however without taking into account the stylistic content of the visual scene. To address this shortcoming, we first analyze the limitations of current stylized captioning datasets and propose COCO attribute-based augmentations to obtain varied stylized captions from COCO annotations. Furthermore, we encode the stylized information in the latent space of a Variational Autoencoder; specifically, we leverage extracted image attributes to explicitly structure its sequential latent space according to different localized style characteristics. Our experiments on the Senticap and COCO datasets show the ability of our approach to generate accurate captions with diversity in styles that are grounded in the image.
Understanding an information forager's actions during interaction is very important for the study of interactive information retrieval. Although information spread in uncertain information space is substantially complex due to the high entanglement of users interacting with information objects~(text, image, etc.). However, an information forager, in general, accompanies a piece of information (information diet) while searching (or foraging) alternative contents, typically subject to decisive uncertainty. Such types of uncertainty are analogous to measurements in quantum mechanics which follow the uncertainty principle. In this paper, we discuss information seeking as a reinforcement learning task. We then present a reinforcement learning-based framework to model forager exploration that treats the information forager as an agent to guide their behaviour. Also, our framework incorporates the inherent uncertainty of the foragers' action using the mathematical formalism of quantum mechanics.
This study presents the applicability of conventional deep recurrent neural networks (RNN) to predict path-dependent plasticity associated with material heterogeneity and anisotropy. Although the architecture of RNN possesses inductive biases toward information over time, it is still challenging to learn the path-dependent material behavior as a function of the loading path considering the change from elastic to elastoplastic regimes. Our attempt is to develop a simple machine-learning-based model that can replicate elastoplastic behaviors considering material heterogeneity and anisotropy. The basic Long-Short Term Memory Unit (LSTM) is adopted for the modeling of plasticity in the two-dimensional space by enhancing the inductive bias toward the past information through manipulating input variables. Our results find that a single LSTM based model can capture the J2 plasticity responses under both monotonic and arbitrary loading paths provided the material heterogeneity. The proposed neural network architecture is then used to model elastoplastic responses of a two-dimensional transversely anisotropic material associated with computational homogenization (FE2). It is also found that a single LSTM model can be used to accurately and effectively capture the path-dependent responses of heterogeneous and anisotropic microstructures under arbitrary mechanical loading conditions.