Traditional statistical inference techniques, such as likelihood ratio tests, have seen renewed interest in recent years, driven in part by the growing emphasis on methodologies based on e-values and e-processes, rather than conventional p-values. Unlike p-values, e-values possess several properties that make them particularly appealing for modern data science applications. In particular, e-value-based methods have played an instrumental role in advancing multiple and safe testing (Grünwald et al., 2020; Vovk and Wang, 2021; Shafer, 2021; Wang and Ramdas, 2022), anytime-valid inference (Waudby-Smith and Ramdas, 2024), and asymptotic confidence sequences (Waudby-Smith et al., 2024). This list is far from exhaustive, and we refer to Ramdas et al. (2023) for a broader overview of recent developments. This manuscript revisits the work of Wasserman et al. (2020), who introduced universal inference, a general hypothesis testing framework based on split likelihood ratio statistics, which is also an e-value. This framework provides simple procedures for many complex composite testing problems that previously lacked actionable solutions, such as testing logconcavity (Dunn et al., 2024) and causal inference under unknown causal structures (Strieder et al., 2021), among others. Specifically, universal inference combines the classical idea of sample splitting (Cox, 1975) and Markov's inequality to establish finite-sample validity. The procedure follows three steps.

Treatment effects are key quantities of interest in applied domains such as medicine and social sciences, as they determine the impact of interventions like novel treatments or policies on outcomes of interest. To achieve this goal, researchers often rely on randomized trials since randomizing the treatment assignment guarantees unbiased treatment effect estimates under mild assumptions. However, methods relying on randomized data face several issues, such as small sample sizes, sample populations that do not reflect those seen in the real world, and ethical or financial constraints. As a result, there is growing interest in using observational data to estimate treatment effects. A fundamental challenge in using observational data is the selection of a valid adjustment set, i.e. a set of covariates that can be used to identify and estimate the treatment effect. Although criteria for identifying valid adjustment sets are well-established, they rely on the knowledge of the underlying causal graph. When the graph is not known, practitioners often adjust for all available covariates [5]. Yet, this approach runs the risk of including bad controls--covariates that open backdoor paths between the treatment (T) and the outcome (Y), thereby introducing bias into the treatment effect estimate.

Methods to quantify uncertainty in predictions from arbitrary models are in demand in high-stakes domains like medicine and finance. Conformal prediction has emerged as a popular method for producing a set of predictions with specified average coverage, in place of a single prediction and confidence value. However, the value of conformal prediction sets to assist human decisions remains elusive due to the murky relationship between coverage guarantees and decision makers' goals and strategies. How should we think about conformal prediction sets as a form of decision support? We outline a decision theoretic framework for evaluating predictive uncertainty as informative signals, then contrast what can be said within this framework about idealized use of calibrated probabilities versus conformal prediction sets. Informed by prior empirical results and theories of human decisions under uncertainty, we formalize a set of possible strategies by which a decision maker might use a prediction set. We identify ways in which conformal prediction sets and posthoc predictive uncertainty quantification more broadly are in tension with common goals and needs in human-AI decision making. We give recommendations for future research in predictive uncertainty quantification to support human decision makers.

Small area estimation models are essential for estimating population characteristics in regions with limited sample sizes, thereby supporting policy decisions, demographic studies, and resource allocation, among other use cases. The spatial Fay-Herriot model is one such approach that incorporates spatial dependence to improve estimation by borrowing strength from neighboring regions. However, this approach often requires substantial computational resources, limiting its scalability for high-dimensional datasets, especially when considering multiple (multivariate) responses. This paper proposes two methods that integrate the multivariate spatial Fay-Herriot model with spatial random effects, learned through variational autoencoders, to efficiently leverage spatial structure. Importantly, after training the variational autoencoder to represent spatial dependence for a given set of geographies, it may be used again in future modeling efforts, without the need for retraining. Additionally, the use of the variational autoencoder to represent spatial dependence results in extreme improvements in computational efficiency, even for massive datasets. We demonstrate the effectiveness of our approach using 5-year period estimates from the American Community Survey over all census tracts in California.

Noise in data appears to be inevitable in most real-world machine learning applications and would cause severe overfitting problems. Not only can data features contain noise, but labels are also prone to be noisy due to human input. In this paper, rather than noisy label learning in multiclass classifications, we instead focus on the less explored area of noisy label learning for multilabel classifications. Specifically, we investigate the post-correction of predictions generated from classifiers learned with noisy labels. The reasons are two-fold. Firstly, this approach can directly work with the trained models to save computational resources. Secondly, it could be applied on top of other noisy label correction techniques to achieve further improvements. To handle this problem, we appeal to deep generative approaches that are possible for uncertainty estimation. Our model posits that label noise arises from a stochastic shift in the latent variable, providing a more robust and beneficial means for noisy learning. We develop both unsupervised and semi-supervised learning methods for our model. The extensive empirical study presents solid evidence to that our approach is able to consistently improve the independent models and performs better than a number of existing methods across various noisy label settings. Moreover, a comprehensive empirical analysis of the proposed method is carried out to validate its robustness, including sensitivity analysis and an ablation study, among other elements.

Orthogonal-split trees perform well, but evidence suggests oblique splits can enhance their performance. This paper explores optimizing high-dimensional $s$-sparse oblique splits from $\{(\vec{w}, \vec{w}^{\top}\boldsymbol{X}_{i}) : i\in \{1,\dots, n\}, \vec{w} \in \mathbb{R}^p, \| \vec{w} \|_{2} = 1, \| \vec{w} \|_{0} \leq s \}$ for growing oblique trees, where $ s $ is a user-defined sparsity parameter. We establish a connection between SID convergence and $s_0$-sparse oblique splits with $s_0\ge 1$, showing that the SID function class expands as $s_0$ increases, enabling the capture of more complex data-generating functions such as the $s_0$-dimensional XOR function. Thus, $s_0$ represents the unknown potential complexity of the underlying data-generating function. Learning these complex functions requires an $s$-sparse oblique tree with $s \geq s_0$ and greater computational resources. This highlights a trade-off between statistical accuracy, governed by the SID function class size depending on $s_0$, and computational cost. In contrast, previous studies have explored the problem of SID convergence using orthogonal splits with $ s_0 = s = 1 $, where runtime was less critical. Additionally, we introduce a practical framework for oblique trees that integrates optimized oblique splits alongside orthogonal splits into random forests. The proposed approach is assessed through simulations and real-data experiments, comparing its performance against various oblique tree models.

While the 2024 Nobel Prize in Physics ignites a worldwide discussion on the origins of neural networks and their foundational links to physics, modern machine learning research predominantly focuses on computational and algorithmic advancements, overlooking a picture of physics. Here we introduce the concept of entropy into neural networks by reconceptualizing them as hypothetical physical systems where each parameter is a non-interacting 'particle' within a one-dimensional space. By employing a Wang-Landau algorithms, we construct the neural networks' (with up to 1 million parameters) entropy landscapes as functions of training loss and test accuracy (or loss) across four distinct machine learning tasks, including arithmetic question, real-world tabular data, image recognition, and language modeling. Our results reveal the existence of \textit{entropy advantage}, where the high-entropy states generally outperform the states reached via classical training optimizer like stochastic gradient descent. We also find this advantage is more pronounced in narrower networks, indicating a need of different training optimizers tailored to different sizes of neural networks.

Neural operators have emerged as powerful surrogates for modeling complex physical problems. However, they suffer from spectral bias making them oblivious to high-frequency modes, which are present in multiscale physical systems. Therefore, they tend to produce over-smoothed solutions, which is particularly problematic in modeling turbulence and for systems with intricate patterns and sharp gradients such as multi-phase flow systems. In this work, we introduce a new approach named high-frequency scaling (HFS) to mitigate spectral bias in convolutional-based neural operators. By integrating HFS with proper variants of UNet neural operators, we demonstrate a higher prediction accuracy by mitigating spectral bias in single and two-phase flow problems. Unlike Fourier-based techniques, HFS is directly applied to the latent space, thus eliminating the computational cost associated with the Fourier transform. Additionally, we investigate alternative spectral bias mitigation through diffusion models conditioned on neural operators. While the diffusion model integrated with the standard neural operator may still suffer from significant errors, these errors are substantially reduced when the diffusion model is integrated with a HFS-enhanced neural operator.

The rapid advancement of artificial intelligence (AI) technologies has led to an increasing deployment of AI models on edge and terminal devices, driven by the proliferation of the Internet of Things (IoT) and the need for real-time data processing. This survey comprehensively explores the current state, technical challenges, and future trends of on-device AI models. We define on-device AI models as those designed to perform local data processing and inference, emphasizing their characteristics such as real-time performance, resource constraints, and enhanced data privacy. The survey is structured around key themes, including the fundamental concepts of AI models, application scenarios across various domains, and the technical challenges faced in edge environments. We also discuss optimization and implementation strategies, such as data preprocessing, model compression, and hardware acceleration, which are essential for effective deployment. Furthermore, we examine the impact of emerging technologies, including edge computing and foundation models, on the evolution of on-device AI models. By providing a structured overview of the challenges, solutions, and future directions, this survey aims to facilitate further research and application of on-device AI, ultimately contributing to the advancement of intelligent systems in everyday life.

The clinical translation of nanoparticle-based treatments remains limited due to the unpredictability of (nanoparticle) NP pharmacokinetics$\unicode{x2014}$how they distribute, accumulate, and clear from the body. Predicting these behaviours is challenging due to complex biological interactions and the difficulty of obtaining high-quality experimental datasets. Existing AI-driven approaches rely heavily on data-driven learning but fail to integrate crucial knowledge about NP properties and biodistribution mechanisms. We introduce a multi-view deep learning framework that enhances pharmacokinetic predictions by incorporating prior knowledge of key NP properties such as size and charge into a cross-attention mechanism, enabling context-aware feature selection and improving generalization despite small datasets. To further enhance prediction robustness, we employ an ensemble learning approach, combining deep learning with XGBoost (XGB) and Random Forest (RF), which significantly outperforms existing AI models. Our interpretability analysis reveals key physicochemical properties driving NP biodistribution, providing biologically meaningful insights into possible mechanisms governing NP behaviour in vivo rather than a black-box model. Furthermore, by bridging machine learning with physiologically based pharmacokinetic (PBPK) modelling, this work lays the foundation for data-efficient AI-driven drug discovery and precision nanomedicine.