Archives AI News

arXiv:2501.18790v2 Announce Type: replace-cross Abstract: We tackle average-reward infinite-horizon POMDPs with an unknown transition model but a known observation model, a setting that has been previously addressed in two limiting ways: (i) frequentist methods relying on suboptimal stochastic policies having a minimum probability of choosing each action, and (ii) Bayesian approaches employing the optimal policy class but requiring strong assumptions about the consistency of employed estimators. Our work removes these limitations by proving convenient estimation guarantees for the transition model and introducing an optimistic algorithm that leverages the optimal class of deterministic belief-based policies. We introduce modifications to existing estimation techniques providing theoretical guarantees separately for each estimated action transition matrix. Unlike existing estimation methods that are unable to use samples from different policies, we present a novel and simple estimator that overcomes this barrier. This new data-efficient technique, combined with the proposed emph{Action-wise OAS-UCRL} algorithm and a tighter theoretical analysis, leads to the first approach enjoying a regret guarantee of order $mathcal{O}(sqrt{T ,log T})$ when compared against the optimal policy, thus improving over state of the art techniques. Finally, theoretical results are validated through numerical simulations showing the efficacy of our method against baseline methods.

arXiv:2509.05460v1 Announce Type: cross Abstract: Spotify's Home page features a variety of content types, including music, podcasts, and audiobooks. However, historical data is heavily skewed toward music, making it challenging to deliver a balanced and personalized content mix. Moreover, users' preference towards different content types may vary depending on the time of day, the day of week, or even the device they use. We propose a calibration method that leverages contextual bandits to dynamically learn each user's optimal content type distribution based on their context and preferences. Unlike traditional calibration methods that rely on historical averages, our approach boosts engagement by adapting to how users interests in different content types varies across contexts. Both offline and online results demonstrate improved precision and user engagement with the Spotify Home page, in particular with under-represented content types such as podcasts.

arXiv:2502.17077v4 Announce Type: replace-cross Abstract: The label ranking problem is a supervised learning scenario in which the learner predicts a total order of the class labels for a given input instance. Recently, research has increasingly focused on the partial label ranking problem, a generalization of the label ranking problem that allows ties in the predicted orders. So far, most existing learning approaches for the partial label ranking problem rely on approximation algorithms for rank aggregation in the final prediction step. This paper explores several alternative aggregation methods for this critical step, including scoring-based and non-parametric probabilistic-based rank aggregation approaches. To enhance their suitability for the more general partial label ranking problem, the investigated methods are extended to increase the likelihood of producing ties. Experimental evaluations on standard benchmarks demonstrate that scoring-based variants consistently outperform the current state-of-the-art method in handling incomplete information. In contrast, non-parametric probabilistic-based variants fail to achieve competitive performance.

arXiv:2509.05563v1 Announce Type: cross Abstract: High-dimensional compositional data, such as those from human microbiome studies, pose unique statistical challenges due to the simplex constraint and excess zeros. While dimension reduction is indispensable for analyzing such data, conventional approaches often rely on log-ratio transformations that compromise interpretability and distort the data through ad hoc zero replacements. We introduce a novel framework for interpretable dimension reduction of compositional data that avoids extra transformations and zero imputations. Our approach generalizes the concept of amalgamation by softening its operation, mapping high-dimensional compositions directly to a lower-dimensional simplex, which can be visualized in ternary plots. The framework further provides joint visualization of the reduction matrix, enabling intuitive, at-a-glance interpretation. To achieve optimal reduction within our framework, we incorporate sufficient dimension reduction, which defines a new identifiable objective: the central compositional subspace. For estimation, we propose a compositional kernel dimension reduction (CKDR) method. The estimator is provably consistent, exhibits sparsity that reveals underlying amalgamation structures, and comes with an intrinsic predictive model for downstream analyses. Applications to real microbiome datasets demonstrate that our approach provides a powerful graphical exploration tool for uncovering meaningful biological patterns, opening a new pathway for analyzing high-dimensional compositional data.

arXiv:2509.05627v1 Announce Type: cross Abstract: AI audits play a critical role in AI accountability and safety. One branch of the law for which AI audits are particularly salient is anti-discrimination law. Several areas of anti-discrimination law implicate the "less discriminatory alternative" (LDA) requirement, in which a protocol (e.g., model) is defensible if no less discriminatory protocol that achieves comparable performance can be found with a reasonable amount of effort. Notably, the burden of proving an LDA exists typically falls on the claimant (the party alleging discrimination). This creates a significant hurdle in AI cases, as the claimant would seemingly need to train a less discriminatory yet high-performing model, a task requiring resources and expertise beyond most litigants. Moreover, developers often shield information about and access to their model and training data as trade secrets, making it difficult to reproduce a similar model from scratch. In this work, we present a procedure enabling claimants to determine if an LDA exists, even when they have limited compute, data, information, and model access. We focus on the setting in which fairness is given by demographic parity and performance by binary cross-entropy loss. As our main result, we provide a novel closed-form upper bound for the loss-fairness Pareto frontier (PF). We show how the claimant can use it to fit a PF in the "low-resource regime," then extrapolate the PF that applies to the (large) model being contested, all without training a single large model. The expression thus serves as a scaling law for loss-fairness PFs. To use this scaling law, the claimant would require a small subsample of the train/test data. Then, the claimant can fit the context-specific PF by training as few as 7 (small) models. We stress test our main result in simulations, finding that our scaling law holds even when the exact conditions of our theory do not.

arXiv:2508.07556v2 Announce Type: replace-cross Abstract: Machine learning (ML) systems are increasingly deployed in high-stakes domains where reliability is paramount. This thesis investigates how uncertainty estimation can enhance the safety and trustworthiness of ML, focusing on selective prediction -- where models abstain when confidence is low. We first show that a model's training trajectory contains rich uncertainty signals that can be exploited without altering its architecture or loss. By ensembling predictions from intermediate checkpoints, we propose a lightweight, post-hoc abstention method that works across tasks, avoids the cost of deep ensembles, and achieves state-of-the-art selective prediction performance. Crucially, this approach is fully compatible with differential privacy (DP), allowing us to study how privacy noise affects uncertainty quality. We find that while many methods degrade under DP, our trajectory-based approach remains robust, and we introduce a framework for isolating the privacy-uncertainty trade-off. Next, we then develop a finite-sample decomposition of the selective classification gap -- the deviation from the oracle accuracy-coverage curve -- identifying five interpretable error sources and clarifying which interventions can close the gap. This explains why calibration alone cannot fix ranking errors, motivating methods that improve uncertainty ordering. Finally, we show that uncertainty signals can be adversarially manipulated to hide errors or deny service while maintaining high accuracy, and we design defenses combining calibration audits with verifiable inference. Together, these contributions advance reliable ML by improving, evaluating, and safeguarding uncertainty estimation, enabling models that not only make accurate predictions -- but also know when to say "I do not know".

arXiv:2509.05671v1 Announce Type: cross Abstract: Human Activity Recognition (HAR) using multimodal sensor data remains challenging due to noisy or incomplete measurements, scarcity of labeled examples, and privacy concerns. Traditional centralized deep learning approaches are often constrained by infrastructure availability, network latency, and data sharing restrictions. While federated learning (FL) addresses privacy by training models locally and sharing only model parameters, it still has to tackle issues arising from the use of heterogeneous multimodal data and differential privacy requirements. In this article, a Graph-based Multimodal Federated Learning framework, GraMFedDHAR, is proposed for HAR tasks. Diverse sensor streams such as a pressure mat, depth camera, and multiple accelerometers are modeled as modality-specific graphs, processed through residual Graph Convolutional Neural Networks (GCNs), and fused via attention-based weighting rather than simple concatenation. The fused embeddings enable robust activity classification, while differential privacy safeguards data during federated aggregation. Experimental results show that the proposed MultiModalGCN model outperforms the baseline MultiModalFFN, with up to 2 percent higher accuracy in non-DP settings in both centralized and federated paradigms. More importantly, significant improvements are observed under differential privacy constraints: MultiModalGCN consistently surpasses MultiModalFFN, with performance gaps ranging from 7 to 13 percent depending on the privacy budget and setting. These results highlight the robustness of graph-based modeling in multimodal learning, where GNNs prove more resilient to the performance degradation introduced by DP noise.

arXiv:2509.05724v1 Announce Type: new Abstract: Recent advances in neural density estimation have enabled powerful simulation-based inference (SBI) methods that can flexibly approximate Bayesian inference for intractable stochastic models. Although these methods have demonstrated reliable posterior estimation when the simulator accurately represents the underlying data generative process (GDP), recent work has shown that they perform poorly in the presence of model misspecification. This poses a significant problem for their use on real-world problems, due to simulators always misrepresenting the true DGP to a certain degree. In this paper, we introduce robust variational neural posterior estimation (RVNP), a method which addresses the problem of misspecification in amortised SBI by bridging the simulation-to-reality gap using variational inference and error modelling. We test RVNP on multiple benchmark tasks, including using real data from astronomy, and show that it can recover robust posterior inference in a data-driven manner without adopting tunable hyperparameters or priors governing the misspecification.

arXiv:2509.05766v1 Announce Type: cross Abstract: Anomaly detection underpins critical applications from network security and intrusion detection to fraud prevention, where recognizing aberrant patterns rapidly is indispensable. Progress in this area is routinely impeded by two obstacles: extreme class imbalance and the curse of dimensionality. To combat the former, we previously introduced Precision-Recall Curve (PRC) classification trees and their ensemble extension, the PRC Random Forest (PRC-RF). Building on that foundation, we now propose a hybrid framework that integrates PRC-RF with autoencoders, unsupervised machine learning methods that learn compact latent representations, to confront both challenges simultaneously. Extensive experiments across diverse benchmark datasets demonstrate that the resulting Autoencoder-PRC-RF model achieves superior accuracy, scalability, and interpretability relative to prior methods, affirming its potential for high-stakes anomaly-detection tasks.

arXiv:2408.09718v2 Announce Type: replace Abstract: Confirmation bias, the tendency to interpret information in a way that aligns with one's preconceptions, can profoundly impact scientific research, leading to conclusions that reflect the researcher's hypotheses even when the observational data do not support them. This issue is especially critical in scientific fields involving highly noisy observations, such as cryo-electron microscopy. This study investigates confirmation bias in Gaussian mixture models. We consider the following experiment: A team of scientists assumes they are analyzing data drawn from a Gaussian mixture model with known signals (hypotheses) as centroids. However, in reality, the observations consist entirely of noise without any informative structure. The researchers use a single iteration of the K-means or expectation-maximization algorithms, two popular algorithms to estimate the centroids. Despite the observations being pure noise, we show that these algorithms yield biased estimates that resemble the initial hypotheses, contradicting the unbiased expectation that averaging these noise observations would converge to zero. Namely, the algorithms generate estimates that mirror the postulated model, although the hypotheses (the presumed centroids of the Gaussian mixture) are not evident in the observations. Specifically, among other results, we prove a positive correlation between the estimates produced by the algorithms and the corresponding hypotheses. We also derive explicit closed-form expressions of the estimates for a finite and infinite number of hypotheses. This study underscores the risks of confirmation bias in low signal-to-noise environments, provides insights into potential pitfalls in scientific methodologies, and highlights the importance of prudent data interpretation.