Projected Stochastic Gradient Descent with Quantum Annealed Binary Gradients

We present, QP-SBGD, a novel layer-wise stochastic optimiser tailored towards training neural networks with binary weights, known as binary neural networks (BNNs), on quantum hardware. BNNs reduce the computational requirements and energy consumption of deep learning models with minimal loss in accuracy. However, training them in practice remains to be an open challenge. Most known BNN-optimisers either rely on projected updates or binarise weights post-training. Instead, QP-SBGD approximately maps the gradient onto binary variables, by solving a quadratic constrained binary optimisation. Under practically reasonable assumptions, we show that this update rule converges with a rate of $\mathcal{O}(1 / \sqrt{T})$. Moreover, we show how the $\mathcal{NP}$-hard projection can be effectively executed on an adiabatic quantum annealer, harnessing recent advancements in quantum computation. We also introduce a projected version of this update rule and prove that if a fixed point exists in the binary variable space, the modified updates will converge to it. Last but not least, our algorithm is implemented layer-wise, making it suitable to train larger networks on resource-limited quantum hardware. Through extensive evaluations, we show that QP-SBGD outperforms or is on par with competitive and well-established baselines such as BinaryConnect, signSGD and ProxQuant when optimising the Rosenbrock function, training BNNs as well as binary graph neural networks.

High Frequency, High Accuracy Pointing onboard Nanosats using Neuromorphic Event Sensing and Piezoelectric Actuation

As satellites become smaller, the ability to maintain stable pointing decreases as external forces acting on the satellite come into play. At the same time, reaction wheels used in the attitude determination and control system (ADCS) introduce high frequency jitter which can disrupt pointing stability. For space domain awareness (SDA) tasks that track objects tens of thousands of kilometres away, the pointing accuracy offered by current nanosats, typically in the range of 10 to 100 arcseconds, is not sufficient. In this work, we develop a novel payload that utilises a neuromorphic event sensor (for high frequency and highly accurate relative attitude estimation) paired in a closed loop with a piezoelectric stage (for active attitude corrections) to provide highly stable sensor-specific pointing. Event sensors are especially suited for space applications due to their desirable characteristics of low power consumption, asynchronous operation, and high dynamic range. We use the event sensor to first estimate a reference background star field from which instantaneous relative attitude is estimated at high frequency. The piezoelectric stage works in a closed control loop with the event sensor to perform attitude corrections based on the discrepancy between the current and desired attitude. Results in a controlled setting show that we can achieve a pointing accuracy in the range of 1-5 arcseconds using our novel payload at an operating frequency of up to 50Hz using a prototype built from commercial-off-the-shelf components. Further details can be found at https://ylatif.github.io/ultrafinestabilisation

Domain Adaptation for Satellite-Borne Hyperspectral Cloud Detection

The advent of satellite-borne machine learning hardware accelerators has enabled the on-board processing of payload data using machine learning techniques such as convolutional neural networks (CNN). A notable example is using a CNN to detect the presence of clouds in hyperspectral data captured on Earth observation (EO) missions, whereby only clear sky data is downlinked to conserve bandwidth. However, prior to deployment, new missions that employ new sensors will not have enough representative datasets to train a CNN model, while a model trained solely on data from previous missions will underperform when deployed to process the data on the new missions. This underperformance stems from the domain gap, i.e., differences in the underlying distributions of the data generated by the different sensors in previous and future missions. In this paper, we address the domain gap problem in the context of on-board hyperspectral cloud detection. Our main contributions lie in formulating new domain adaptation tasks that are motivated by a concrete EO mission, developing a novel algorithm for bandwidth-efficient supervised domain adaptation, and demonstrating test-time adaptation algorithms on space deployable neural network accelerators. Our contributions enable minimal data transmission to be invoked (e.g., only 1% of the weights in ResNet50) to achieve domain adaptation, thereby allowing more sophisticated CNN models to be deployed and updated on satellites without being hampered by domain gap and bandwidth limitations.

Direct initial orbit determination

Initial orbit determination (IOD) is an important early step in the processing chain that makes sense of and reconciles the multiple optical observations of a resident space object. IOD methods generally operate on line-of-sight (LOS) vectors extracted from images of the object, hence the LOS vectors can be seen as discrete point samples of the raw optical measurements. Typically, the number of LOS vectors used by an IOD method is much smaller than the available measurements (\ie, the set of pixel intensity values), hence current IOD methods arguably under-utilize the rich information present in the data. In this paper, we propose a \emph{direct} IOD method called D-IOD that fits the orbital parameters directly on the observed streak images, without requiring LOS extraction. Since it does not utilize LOS vectors, D-IOD avoids potential inaccuracies or errors due to an imperfect LOS extraction step. Two innovations underpin our novel orbit-fitting paradigm: first, we introduce a novel non-linear least-squares objective function that computes the loss between the candidate-orbit-generated streak images and the observed streak images. Second, the objective function is minimized with a gradient descent approach that is embedded in our proposed optimization strategies designed for streak images. We demonstrate the effectiveness of D-IOD on a variety of simulated scenarios and challenging real streak images.

Semantic Segmentation on 3D Point Clouds with High Density Variations

LiDAR scanning for surveying applications acquire measurements over wide areas and long distances, which produces large-scale 3D point clouds with significant local density variations. While existing 3D semantic segmentation models conduct downsampling and upsampling to build robustness against varying point densities, they are less effective under the large local density variations characteristic of point clouds from surveying applications. To alleviate this weakness, we propose a novel architecture called HDVNet that contains a nested set of encoder-decoder pathways, each handling a specific point density range. Limiting the interconnections between the feature maps enables HDVNet to gauge the reliability of each feature based on the density of a point, e.g., downweighting high density features not existing in low density objects. By effectively handling input density variations, HDVNet outperforms state-of-the-art models in segmentation accuracy on real point clouds with inconsistent density, using just over half the weights.

Federated Neural Radiance Fields

The ability of neural radiance fields or NeRFs to conduct accurate 3D modelling has motivated application of the technique to scene representation. Previous approaches have mainly followed a centralised learning paradigm, which assumes that all training images are available on one compute node for training. In this paper, we consider training NeRFs in a federated manner, whereby multiple compute nodes, each having acquired a distinct set of observations of the overall scene, learn a common NeRF in parallel. This supports the scenario of cooperatively modelling a scene using multiple agents. Our contribution is the first federated learning algorithm for NeRF, which splits the training effort across multiple compute nodes and obviates the need to pool the images at a central node. A technique based on low-rank decomposition of NeRF layers is introduced to reduce bandwidth consumption to transmit the model parameters for aggregation. Transferring compressed models instead of the raw data also contributes to the privacy of the data collecting agents.

Training Multilayer Perceptrons by Sampling with Quantum Annealers

A successful application of quantum annealing to machine learning is training restricted Boltzmann machines (RBM). However, many neural networks for vision applications are feedforward structures, such as multilayer perceptrons (MLP). Backpropagation is currently the most effective technique to train MLPs for supervised learning. This paper aims to be forward-looking by exploring the training of MLPs using quantum annealers. We exploit an equivalence between MLPs and energy-based models (EBM), which are a variation of RBMs with a maximum conditional likelihood objective. This leads to a strategy to train MLPs with quantum annealers as a sampling engine. We prove our setup for MLPs with sigmoid activation functions and one hidden layer, and demonstrated training of binary image classifiers on small subsets of the MNIST and Fashion-MNIST datasets using the D-Wave quantum annealer. Although problem sizes that are feasible on current annealers are limited, we obtained comprehensive results on feasible instances that validate our ideas. Our work establishes the potential of quantum computing for training MLPs.

Assessing Domain Gap for Continual Domain Adaptation in Object Detection

To ensure reliable object detection in autonomous systems, the detector must be able to adapt to changes in appearance caused by environmental factors such as time of day, weather, and seasons. Continually adapting the detector to incorporate these changes is a promising solution, but it can be computationally costly. Our proposed approach is to selectively adapt the detector only when necessary, using new data that does not have the same distribution as the current training data. To this end, we investigate three popular metrics for domain gap evaluation and find that there is a correlation between the domain gap and detection accuracy. Therefore, we apply the domain gap as a criterion to decide when to adapt the detector. Our experiments show that our approach has the potential to improve the efficiency of the detector's operation in real-world scenarios, where environmental conditions change in a cyclical manner, without sacrificing the overall performance of the detector. Our code is publicly available at https://github.com/dadung/DGE-CDA.