Search for: All records

The advent of diverse frequency bands in 5G networks has promoted measurement studies focused on 5G signal propagation, aiming to understand its pathloss, coverage, and channel quality characteristics. Nonetheless, conducting a thorough 5G measurement campaign is markedly laborious given the large number of 5G measurement samples that must be collected. To alleviate this burden, the present contribution leverages principled active learning (AL) methods to prudently select only a few, yet most informative locations to collect 5G measurements. The core idea is to rely on a Gaussian Process (GP) model to efficiently extrapolate 5G measurements throughout the coverage area. Specifically, an ensemble (E) of GP models is adopted that not only provides a rich learning function space, but also quantifies uncertainty, and can offer accurate predictions. Building on this EGP model, a suite of acquisition functions (AFs) are advocated to query new locations on-the-fly. To account for realistic 5G measurement campaigns, the proposed AFs are augmented with a novel distance-based AL rule that selects informative samples, while penalizing queries at long distances. Numerical tests on 5G data generated by the Sionna simulator and on real urban and suburban datasets, showcase the merits of the novel EGP-AL approaches. 

Relying on prior knowledge accumulated from related tasks, meta-learning offers a powerful approach to learning a novel task from a limited number of training data. Recent approaches use a family of prior probability density functions or recurrent neural network models, whose parameters can be optimized by utilizing labeled data from the observed tasks. While these approaches have appealing empirical performance, expressiveness of their prior is relatively low, which limits generalization and interpretation of meta-learning. Aiming at expressive yet meaningful priors, this contribution puts forth a novel prior representation model that leverages the notion of algorithm unrolling. The key idea is to unroll the proximal gradient descent steps, where learnable piecewise linear functions are developed to approximate the desired proximal operators within tight theoretical error bounds established for both smooth and non-smooth proximal functions. The resultant multi-block neural network not only broadens the scope of learnable priors, but also enhances interpretability from an optimization viewpoint. Numerical tests conducted on few-shot learning datasets demonstrate markedly improved performance with flexible, visualizable, and understandable priors. 

Using a subset of observed network links, high-order link prediction (HOLP) infers missing hyperedges, that is links connecting three or more nodes. HOLP emerges in several applications, but existing approaches have not dealt with the associated predictor’s performance. To overcome this limitation, the present contribution develops a Bayesian approach and the relevant predictive distributions that quantify model uncertainty. Gaussian processes model the dependence of each node to the remaining nodes. These nonparametric models yield predictive distributions, which are fused across nodes by means of a pseudo-likelihood based criterion. Performance is quantified by proper measures of dispersion, which are associated with the predictive distributions. Tests on benchmark datasets demonstrate the benefits of the novel approach. 

Utilizing task-invariant prior knowledge extracted from related tasks, meta-learning is a principled framework that empowers learning a new task especially when data records are limited. A fundamental challenge in meta-learning is how to quickly "adapt" the extracted prior in order to train a task-specific model within a few optimization steps. Existing approaches deal with this challenge using a preconditioner that enhances convergence of the per-task training process. Though effective in representing locally a quadratic training loss, these simple linear preconditioners can hardly capture complex loss geometries. The present contribution addresses this limitation by learning a nonlinear mirror map, which induces a versatile distance metric to enable capturing and optimizing a wide range of loss geometries, hence facilitating the per-task training. Numerical tests on few-shot learning datasets demonstrate the superior expressiveness and convergence of the advocated approach. 

Labeled data can be expensive to acquire in several application domains, including medical imaging, robotics, computer vision and wireless networks to list a few. To efficiently train machine learning models under such high labeling costs, active learning (AL) judiciously selects the most informative data instances to label on-the-fly. This active sampling process can benefit from a statistical function model, that is typically captured by a Gaussian process (GP) with well-documented merits especially in the regression task. While most GP-based AL approaches rely on a single kernel function, the present contribution advocates an ensemble of GP (EGP) models with weights adapted to the labeled data collected incrementally. Building on this novel EGP model, a suite of acquisition functions emerges based on the uncertainty and disagreement rules. An adaptively weighted ensemble of EGP-based acquisition functions is advocated to further robustify performance. Extensive tests on synthetic and real datasets in the regression task showcase the merits of the proposed EGP-based approaches with respect to the single GP-based AL alternatives. 

Recent years have witnessed the emergence of mobile edge computing (MEC), on the premise of a costeffective enhancement in the computational ability of hardwareconstrained wireless devices (WDs) comprising the Internet of Things (IoT). In a general multi-server multi-user MEC system, each WD has a computational task to execute and has to select binary (off)loading decisions, along with the analog-amplitude resource allocation variables in an online manner, with the goal of minimizing the overall energy-delay cost (EDC) with dynamic system states. While past works typically rely on the explicit expression of the EDC function, the present contribution considers a practical setting, where in lieu of system state information, the EDC function is not available in analytical form, and instead only the function values at queried points are revealed. Towards tackling such a challenging online combinatorial problem with only bandit information, novel Bayesian optimization (BO) based approaches are put forth by leveraging the multi-armed bandit (MAB) framework. Per time slot, the discrete offloading decisions are first obtained via the MAB method, and the analog resource allocation variables are subsequently optimized using the BO selection rule. By exploiting both temporal and contextual information, two novel BO approaches, termed time-varying BO and contextual time-varying BO, are developed. Numerical tests validate the merits of the proposed BO approaches compared with contemporary benchmarks under different MEC network sizes. 

Partial correlations (PCs) and the related inverse covariance matrix adopted by graphical lasso, are widely applicable tools for learning graph connectivity given nodal observations. The resultant estimators however, can be sensitive to outliers. Robust approaches developed so far to cope with outliers do not (explicitly) account for nonlinear interactions possibly present among nodal processes. This can hurt the identification of graph connectivity, merely due to model mismatch. To overcome this limitation, a novel formulation of robust PC is introduced based on nonlinear kernel functions. The proposed scheme leverages robust ridge regression techniques, spectral Fourier feature based kernel approximants, and robust association measures. Numerical tests on synthetic and real data illustrate the potential of the novel approach. 

Conic programming has well-documented merits in a gamut of signal processing and machine learning tasks. This contribution revisits a recently developed first-order conic descent (CD) solver, and advances it in three aspects: intuition, theory, and algorithmic implementation. It is found that CD can afford an intuitive geometric derivation that originates from the dual problem. This opens the door to novel algorithmic designs, with a momentum variant of CD, momentum conic descent (MOCO) exemplified. Diving deeper into the dual behavior CD and MOCO reveals: i) an analytically justified stopping criterion; and, ii) the potential to design preconditioners to speed up dual convergence. Lastly, to scale semidefinite programming (SDP) especially for low-rank solutions, a memory efficient MOCO variant is developed and numerically validated.