In this paper, we propose a probabilistic reduced-dimensional vector autoregressive (PredVAR) model to extract low-dimensional dynamics from high-dimensional noisy data. The model utilizes an oblique projection to partition the measurement space into a subspace that accommodates the reduced-dimensional dynamics and a complementary static subspace. An optimal oblique decomposition is derived for the best predictability regarding prediction error covariance. Building on this, we develop an iterative PredVAR algorithm using maximum likelihood and the expectation-maximization (EM) framework. This algorithm alternately updates the estimates of the latent dynamics and optimal oblique projection, yielding dynamic latent variables with rank-ordered predictability and an explicit latent VAR model that is consistent with the outer projection model. The superior performance and efficiency of the proposed approach are demonstrated using data sets from a synthesized Lorenz system and an industrial process from Eastman Chemical.

Closed-source language models such as GPT-4 have achieved remarkable performance. Many recent studies focus on enhancing the capabilities of smaller models through knowledge distillation from closed-source language models. However, due to the incapability to directly access the weights, hidden states, and output distributions of these closed-source models, the distillation can only be performed by fine-tuning smaller models with data samples generated by closed-source language models, which constrains the effectiveness of knowledge distillation. In this paper, we propose to estimate the output distributions of closed-source language models within a Bayesian estimation framework, involving both prior and posterior estimation. The prior estimation aims to derive a prior distribution by utilizing the corpus generated by closed-source language models, while the posterior estimation employs a proxy model to update the prior distribution and derive a posterior distribution. By leveraging the estimated output distribution of closed-source language models, traditional knowledge distillation can be executed. Experimental results demonstrate that our method surpasses the performance of current models directly fine-tuned on data generated by closed-source language models.

In the quest to advance human-centric natural language generation (NLG) systems, ensuring alignment between NLG models and human preferences is crucial. For this alignment, current popular methods leverage a reinforcement learning (RL) approach with a reward model trained on feedback from humans. However, inherent disagreements due to the subjective nature of human preferences pose a significant challenge for training the reward model, resulting in a deterioration of the NLG performance. To tackle this issue, previous approaches typically rely on majority voting or averaging to consolidate multiple inconsistent preferences into a merged one. Although straightforward to understand and execute, such methods suffer from an inability to capture the nuanced degrees of disaggregation among humans and may only represent a specialized subset of individuals, thereby lacking the ability to quantitatively disclose the universality of human preferences. To address this challenge, this paper proposes a novel approach, which employs a Bayesian framework to account for the distribution of disagreements among human preferences as training a preference model, and names it as d-PM. Besides, considering the RL strategy's inefficient and complex training process over the training efficiency, we further propose utilizing the contrastive learning strategy to train the NLG model with the preference scores derived from the d-PM model. Extensive experiments on two human-centric NLG tasks, i.e., emotional support conversation and integrity "Rule-of-Thumb" generation, show that our method consistently exceeds previous SOTA models in both automatic and human evaluations.

Recently, machine learning (ML) has become a popular approach to support self-adaptation. ML has been used to deal with several problems in self-adaptation, such as maintaining an up-to-date runtime model under uncertainty and scalable decision-making. Yet, exploiting ML comes with inherent challenges. In this paper, we focus on a particularly important challenge for learning-based self-adaptive systems: drift in adaptation spaces. With adaptation space we refer to the set of adaptation options a self-adaptive system can select from at a given time to adapt based on the estimated quality properties of the adaptation options. Drift of adaptation spaces originates from uncertainties, affecting the quality properties of the adaptation options. Such drift may imply that eventually no adaptation option can satisfy the initial set of the adaptation goals, deteriorating the quality of the system, or adaptation options may emerge that allow enhancing the adaptation goals. In ML, such shift corresponds to novel class appearance, a type of concept drift in target data that common ML techniques have problems dealing with. To tackle this problem, we present a novel approach to self-adaptation that enhances learning-based self-adaptive systems with a lifelong ML layer. We refer to this approach as lifelong self-adaptation. The lifelong ML layer tracks the system and its environment, associates this knowledge with the current tasks, identifies new tasks based on differences, and updates the learning models of the self-adaptive system accordingly. A human stakeholder may be involved to support the learning process and adjust the learning and goal models. We present a general architecture for lifelong self-adaptation and apply it to the case of drift of adaptation spaces that affects the decision-making in self-adaptation. We validate the approach for a series of scenarios using the DeltaIoT exemplar.

The Ultra-Reliable Low-Latency Communications (URLLC) paradigm in sixth-generation (6G) systems heavily relies on precise channel modeling, especially when dealing with rare and extreme events within wireless communication channels. This paper explores a novel methodology integrating Extreme Value Theory (EVT) and Generative Adversarial Networks (GANs) to achieve the precise channel modeling in real-time. The proposed approach harnesses EVT by employing the Generalized Pareto Distribution (GPD) to model the distribution of extreme events. Subsequently, Generative Adversarial Networks (GANs) are employed to estimate the parameters of the GPD. In contrast to conventional GAN configurations that focus on estimating the overall distribution, the proposed approach involves the incorporation of an additional block within the GAN structure. This specific augmentation is designed with the explicit purpose of directly estimating the parameters of the Generalized Pareto Distribution (GPD). Through extensive simulations across different sample sizes, the proposed GAN based approach consistently demonstrates superior adaptability, surpassing Maximum Likelihood Estimation (MLE), particularly in scenarios with limited sample sizes.

This personalized approach not only enhances the efficacy of treatments but also minimizes the risk of adverse effects (Agyeman and Ofori-Asenso, 2015; Kumari et al., 2023). However, the success of precision medicine heavily relies on the interpretation of complex, multidimensional data sets, where statistical analysis plays a pivotal role (Alyass et al., 2015). The integration of statistical methodologies in precision medicine is not just a mere addition but a fundamental necessity. Advanced statistical techniques enable the extraction of meaningful insights from large-scale genomics, proteomic, and metabolomic data, which are the cornerstone of precision medicine (Wafi and Mirnezami, 2018; Pinu et al., 2019). These methodologies include, but are not limited to, predictive modeling, machine learning algorithms, and complex data visualization techniques, all of which contribute to more accurate diagnosis, prognosis, and treatment planning (Bellazzi and Zupan, 2008; Davatzikos et al., 2018; Richter and Khoshgoftaar, 2018). The heterogeneity of data sources in precision medicine, ranging from electronic health records (EHRs) to high-throughput sequencing data, presents unique challenges in data integration and interpretation (Martinez-Garcia and Hernández-Lemus, 2022). Statistical analysis serves as a bridge, merging these diverse data types into coherent, interpretable information that can guide clinical decision-making. However, the field is not without its challenges. Issues such as overfitting, handling of highdimensional data, and maintaining the balance between model complexity and interpretability are ongoing areas of research (Bolón-Canedo et al., 2015; Xu et al., 2019; Bommert, 2020; Pes, 2020; Hou and Behdinan, 2022).

The flexibility of choosing the ad action as a function of the consumer state is critical for modern-day marketing campaigns. We study the problem of identifying the optimal sequential personalized interventions that maximize the adoption probability for a new product. We model consumer behavior by a conversion funnel that captures the state of each consumer (e.g., interaction history with the firm) and allows the consumer behavior to vary as a function of both her state and firm's sequential interventions. We show our model captures consumer behavior with very high accuracy (out-of-sample AUC of over 0.95) in a real-world email marketing dataset. However, it results in a very large-scale learning problem, where the firm must learn the state-specific effects of various interventions from consumer interactions. We propose a novel attribution-based decision-making algorithm for this problem that we call model-free approximate Bayesian learning. Our algorithm inherits the interpretability and scalability of Thompson sampling for bandits and maintains an approximate belief over the value of each state-specific intervention. The belief is updated as the algorithm interacts with the consumers. Despite being an approximation to the Bayes update, we prove the asymptotic optimality of our algorithm and analyze its convergence rate. We show that our algorithm significantly outperforms traditional approaches on extensive simulations calibrated to a real-world email marketing dataset.

Autonomous agents rely on sensor data to construct representations of their environment, essential for predicting future events and planning their own actions. However, sensor measurements suffer from limited range, occlusions, and sensor noise. These challenges become more evident in dynamic environments, where efficiently inferring the state of the environment based on sensor readings from different times is still an open problem. This work focuses on inferring the state of the dynamic part of the environment, i.e., where dynamic objects might be, based on previous observations and constraints on their dynamics. We formalize the problem and introduce Transitional Grid Maps (TGMs), an efficient analytical solution. TGMs are based on a set of novel assumptions that hold in many practical scenarios. They significantly reduce the complexity of the problem, enabling continuous prediction and updating of the entire dynamic map based on the known static map (see Fig.1), differentiating them from other alternatives. We compare our approach with a state-of-the-art particle filter, obtaining more prudent predictions in occluded scenarios and on-par results on unoccluded tracking.

Positive-Unlabelled (PU) learning is a growing field of machine learning that aims to learn classifiers from data consisting of labelled positive and unlabelled instances, which can be in reality positive or negative, but whose label is unknown. An extensive number of methods have been proposed to address PU learning over the last two decades, so many so that selecting an optimal method for a given PU learning task presents a challenge. Our previous work has addressed this by proposing GA-Auto-PU, the first Automated Machine Learning (Auto-ML) system for PU learning. In this work, we propose two new Auto-ML systems for PU learning: BO-Auto-PU, based on a Bayesian Optimisation approach, and EBO-Auto-PU, based on a novel evolutionary/Bayesian optimisation approach. We also present an extensive evaluation of the three Auto-ML systems, comparing them to each other and to well-established PU learning methods across 60 datasets (20 real-world datasets, each with 3 versions in terms of PU learning characteristics).

Exemplar-free class-incremental learning (CIL) poses several challenges since it prohibits the rehearsal of data from previous tasks and thus suffers from catastrophic forgetting. Recent approaches to incrementally learning the classifier by freezing the feature extractor after the first task have gained much attention. In this paper, we explore prototypical networks for CIL, which generate new class prototypes using the frozen feature extractor and classify the features based on the Euclidean distance to the prototypes. In an analysis of the feature distributions of classes, we show that classification based on Euclidean metrics is successful for jointly trained features. However, when learning from non-stationary data, we observe that the Euclidean metric is suboptimal and that feature distributions are heterogeneous. To address this challenge, we revisit the anisotropic Mahalanobis distance for CIL. In addition, we empirically show that modeling the feature covariance relations is better than previous attempts at sampling features from normal distributions and training a linear classifier.