Human activity recognition (HAR) using wearable sensors has benefited much less from recent advances in Machine Learning than fields such as computer vision and natural language processing. This is to a large extent due to the lack of large scale repositories of labeled training data. In our research we aim to facilitate the use of online videos, which exists in ample quantity for most activities and are much easier to label than sensor data, to simulate labeled wearable motion sensor data. In previous work we already demonstrate some preliminary results in this direction focusing on very simple, activity specific simulation models and a single sensor modality (acceleration norm)\cite{10.1145/3341162.3345590}. In this paper we show how we can train a regression model on generic motions for both accelerometer and gyro signals and then apply it to videos of the target activities to generate synthetic IMU data (acceleration and gyro norms) that can be used to train and/or improve HAR models. We demonstrate that systems trained on simulated data generated by our regression model can come to within around 10% of the mean F1 score of a system trained on real sensor data. Furthermore we show that by either including a small amount of real sensor data for model calibration or simply leveraging the fact that (in general) we can easily generate much more simulated data from video than we can collect in terms of real sensor data the advantage of real sensor data can be eventually equalized.

Here we build a hierarchical modular network called Language guided machine action (LGMA), whose modules process information stream mimicking human cortical network that allows to achieve multiple general tasks such as language guided action, intention decomposition and mental simulation before action execution etc. LGMA contains 3 main systems: (1) primary sensory system that multimodal sensory information of vision, language and sensorimotor. (2) association system involves and Broca modules to comprehend and synthesize language, BA14/40 module to translate between sensorimotor and language, midTemporal module to convert between language and vision, and superior parietal lobe to integrate attended visual object and arm state into cognitive map for future spatial actions. Pre-supplementary motor area (pre-SMA) can converts high level intention into sequential atomic actions, while SMA can integrate these atomic actions, current arm and attended object state into sensorimotor vector to apply corresponding torques on arm via pre-motor and primary motor of arm to achieve the intention. The high-level executive system contains PFC that does explicit inference and guide voluntary action based on language, while BG is the habitual action control center.

The brain is a highly reconfigurable machine capable of task-specific adaptations. The brain continually rewires itself for a more optimal configuration to solve problems. We propose a novel strategic synthesis algorithm for feedforward networks that draws directly from the brain's behaviours when learning. The proposed approach analyses the network and ranks weights based on their magnitude. Unlike existing approaches that advocate random selection, we select highly performing nodes as starting points for new edges and exploit the Gaussian distribution over the weights to select corresponding endpoints. The strategy aims only to produce useful connections and result in a smaller residual network structure. The approach is complemented with pruning to further the compression. We demonstrate the techniques to deep feedforward networks. The residual sub-networks that are formed from the synthesis approaches in this work form common sub-networks with similarities up to ~90%. Using pruning as a complement to the strategic synthesis approach, we observe improvements in compression.

Time series are all around in real-world applications. However, unexpected accidents for example broken sensors or missing of the signals will cause missing values in time series, making the data hard to be utilized. It then does harm to the downstream applications such as traditional classification or regression, sequential data integration and forecasting tasks, thus raising the demand for data imputation. Currently, time series data imputation is a well-studied problem with different categories of methods. However, these works rarely take the temporal relations among the observations and treat the time series as normal structured data, losing the information from the time data. In recent, deep learning models have raised great attention. Time series methods based on deep learning have made progress with the usage of models like RNN, since it captures time information from data. In this paper, we mainly focus on time series imputation technique with deep learning methods, which recently made progress in this field. We will review and discuss their model architectures, their pros and cons as well as their effects to show the development of the time series imputation methods.

The major driver of global warming has been identified as the anthropogenic release of greenhouse gas (GHG) emissions from industrial activities. The quantitative monitoring of these emissions is mandatory to fully understand their effect on the Earth's climate and to enforce emission regulations on a large scale. In this work, we investigate the possibility to detect and quantify industrial smoke plumes from globally and freely available multi-band image data from ESA's Sentinel-2 satellites. Using a modified ResNet-50, we can detect smoke plumes of different sizes with an accuracy of 94.3%. The model correctly ignores natural clouds and focuses on those imaging channels that are related to the spectral absorption from aerosols and water vapor, enabling the localization of smoke. We exploit this localization ability and train a U-Net segmentation model on a labeled sub-sample of our data, resulting in an Intersection-over-Union (IoU) metric of 0.608 and an overall accuracy for the detection of any smoke plume of 94.0%; on average, our model can reproduce the area covered by smoke in an image to within 5.6%. The performance of our model is mostly limited by occasional confusion with surface objects, the inability to identify semi-transparent smoke, and human limitations to properly identify smoke based on RGB-only images. Nevertheless, our results enable us to reliably detect and qualitatively estimate the level of smoke activity in order to monitor activity in industrial plants across the globe. Our data set and code base are publicly available.

Over the past decades the machine and deep learning community has celebrated great achievements in challenging tasks such as image classification. The deep architecture of artificial neural networks together with the plenitude of available data makes it possible to describe highly complex relations. Yet, it is still impossible to fully capture what the deep learning model has learned and to verify that it operates fairly and without creating bias, especially in critical tasks, for instance those arising in the medical field. One example for such a task is the detection of distinct facial expressions, called Action Units, in facial images. Considering this specific task, our research aims to provide transparency regarding bias, specifically in relation to gender and skin color. We train a neural network for Action Unit classification and analyze its performance quantitatively based on its accuracy and qualitatively based on heatmaps. A structured review of our results indicates that we are able to detect bias. Even though we cannot conclude from our results that lower classification performance emerged solely from gender and skin color bias, these biases must be addressed, which is why we end by giving suggestions on how the detected bias can be avoided.

Graphs play an important role in many applications. Recently, Graph Neural Networks (GNNs) have achieved promising results in graph analysis tasks. Some state-of-the-art GNN models have been proposed, e.g., Graph Convolutional Networks (GCNs), Graph Attention Networks (GATs), etc. Despite these successes, most of the GNNs only have shallow structure. This causes the low expressive power of the GNNs. To fully utilize the power of the deep neural network, some deep GNNs have been proposed recently. However, the design of deep GNNs requires significant architecture engineering. In this work, we propose a method to automate the deep GNNs design. In our proposed method, we add a new type of skip connection to the GNNs search space to encourage feature reuse and alleviate the vanishing gradient problem. We also allow our evolutionary algorithm to increase the layers of GNNs during the evolution to generate deeper networks. We evaluate our method in the graph node classification task. The experiments show that the GNNs generated by our method can obtain state-of-the-art results in Cora, Citeseer, Pubmed and PPI datasets.

Fake news often involves semantic manipulations across modalities such as image, text, location etc and requires the development of multimodal semantic forensics for its detection. Recent research has centered the problem around images, calling it image repurposing -- where a digitally unmanipulated image is semantically misrepresented by means of its accompanying multimodal metadata such as captions, location, etc. The image and metadata together comprise a multimedia package. The problem setup requires algorithms to perform multimodal semantic forensics to authenticate a query multimedia package using a reference dataset of potentially related packages as evidences. Existing methods are limited to using a single evidence (retrieved package), which ignores potential performance improvement from the use of multiple evidences. In this work, we introduce a novel graph neural network based model for multimodal semantic forensics, which effectively utilizes multiple retrieved packages as evidences and is scalable with the number of evidences. We compare the scalability and performance of our model against existing methods. Experimental results show that the proposed model outperforms existing state-of-the-art algorithms with an error reduction of up to 25%.

We present a novel way for self-supervised video representation learning by: (a) decoupling the learning objective into two contrastive subtasks respectively emphasizing spatial and temporal features, and (b) performing it hierarchically to encourage multi-scale understanding. Motivated by their effectiveness in supervised learning, we first introduce spatial-temporal feature learning decoupling and hierarchical learning to the context of unsupervised video learning. In particular, our method directs the network to separately capture spatial and temporal features on the basis of contrastive learning via manipulating augmentations as regularization, and further solve such proxy tasks hierarchically by optimizing towards a compound contrastive loss. Experiments show that our proposed Hierarchically Decoupled Spatial-Temporal Contrast (HDC) achieves substantial gains over directly learning spatial-temporal features as a whole and significantly outperforms other state-of-the-art unsupervised methods on downstream action recognition benchmarks on UCF101 and HMDB51. We will release our code and pretrained weights.

Autonomous navigation consists in an agent being able to navigate without human intervention or supervision, it affects both high level planning and low level control. Navigation is at the crossroad of multiple disciplines, it combines notions of computer vision, robotics and control. This work aimed at creating, in a simulation, a navigation pipeline whose transfer to the real world could be done with as few efforts as possible. Given the limited time and the wide range of problematic to be tackled, absolute navigation performances while important was not the main objective. The emphasis was rather put on studying the sim2real gap which is one the major bottlenecks of modern robotics and autonomous navigation. To design the navigation pipeline four main challenges arise; environment, localization, navigation and planning. The iGibson simulator is picked for its photo-realistic textures and physics engine. A topological approach to tackle space representation was picked over metric approaches because they generalize better to new environments and are less sensitive to change of conditions. The navigation pipeline is decomposed as a localization module, a planning module and a local navigation module. These modules utilize three different networks, an image representation extractor, a passage detector and a local policy. The laters are trained on specifically tailored tasks with some associated datasets created for those specific tasks. Localization is the ability for the agent to localize itself against a specific space representation. It must be reliable, repeatable and robust to a wide variety of transformations. Localization is tackled as an image retrieval task using a deep neural network trained on an auxiliary task as a feature descriptor extractor. The local policy is trained with behavioral cloning from expert trajectories gathered with ROS navigation stack.