Science & Mathematics

Clinical expertise is inherently distributed: it resides not in any single dataset or institution but in the accumulated reasoning experiences of thousands of clinicians navigating diverse patient populations, resource constraints, and evolving treatment protocols. Yet current approaches to scaling medical AI, from centralized large language models to federated learning, remain anchored to a data-centric paradigm that treats knowledge as static parameters to be aggregated. In this Perspective, we argue for a fundamental shift from data aggregation to experience aggregation. We propose Decentralized Agentic Systems (DAS), a framework in which distributed AI agents operating across healthcare institutions synthesize collective clinical intelligence by sharing abstracted reasoning trajectories rather than raw data or model parameters. Unlike federated learning, which averages numerical gradients across homogeneous model architectures, DAS exchanges structured records of clinical reasoning, specifically the sequence of observations, hypotheses, tool invocations, and outcome feedback that constitute diagnostic experience. We examine the architectural principles, deployment considerations, governance requirements, and critical open challenges of this framework, laying the groundwork for a collaborative medical AI ecosystem where distributed expertise autonomously adapts to the complexities of longitudinal patient care.

Ferroelectric polarization offers a powerful route to govern the interfacial charge separation in photoelectrodes; however, chemically converted ferroelectric shells lack sub-nanometer thickness control, coherent strain retention, and chemically clean interfaces. Here, we use pulsed laser deposition to fabricate stoichiometric SrTiO3 shells on hydrothermally grown TiO2 nanorod arrays, achieving atomically defined core-shell architectures while preserving the one-dimensional transport geometry. The photocurrent density followed a volcano-shaped dependence on shell thickness and reached 1.9 mA cm-2 at 1.23 V versus RHE under AM 1.5G illumination for the optimized 15-mins sample, exceeding the established hydrothermal benchmark. High-resolution transmission electron microscopy and X-ray photoelectron spectroscopy revealed a tetragonally distorted, oxygen-stoichiometric shell, identifying coherent interfacial strain rather than defect dipoles as the origin of the ferroelectric order. Kelvin probe force microscopy directly imaged a polarization-induced built-in electric field whose magnitude scaled with charge-transfer efficiency and photocurrent density. Together, these complementary structural, chemical, electrostatic, and device-level measurements established a quantitative strain-polarization-built-in field-charge separation relationship. This work transforms ferroelectric photoanode engineering from empirical optimization into an atomically controllable and predictive materials strategy for high-efficiency solar water-splitting devices. This approach is readily extendable to visible-light absorbers and complex ferroelectric heterostructures.

Bayesian variants of Low-Rank Adaptation (LoRA) provide principled uncertainty estimation when fine-tuning large foundation models, but require specifying the prior over the low-rank weight updates. We address this hyperparameter selection problem by introducing a closed-form prior-predictive estimator that operates before fine-tuning. We analyze a second-order expansion of the network and quantify how the low-rank prior drives function-space deviations from the pretrained outputs. We also show that these prior-predictive quantities depend on the prior only through the product of the adaptation rank and the fourth power of the prior scale. Because the training targets share a single draw of the latent weight update, we can leverage the covariance between example residuals to derive a moment-matching estimator. Furthermore, our theoretical analysis establishes that the optimal prior precision scales as $r^{-1/2}$ with the rank $r$, allowing a single estimate to generalize across all ranks. Controlled experiments across regression and classification tasks validate our approximation and show that our estimator recovers the true prior scale. On RoBERTa-large, our approach improves predictive negative log-likelihood (NLL) and calibration over both MAP and marginal-likelihood selection without requiring validation data, while achieving competitive out-of-distribution detection.

Data-enabled predictive control has recently emerged as a promising control paradigm, enabling a transparent optimal-control formulation without the need to learn an explicit dynamics model. This paper presents a number of extensions paving the path towards its use in real-time robotics tasks. Following recent work, we utilize a nonlinear predictor based on Gaussian kernels to represent the relationship between input and output data. We propose a damped Gauss-Newton method that enables real-time control by efficiently exploiting the least-squares structure of the control problem. In addition, we introduce a hybrid buffer management strategy that incorporates informative online data while retaining relevant offline data, allowing the predictor to account for changes in the dynamics. We demonstrate the tracking performance and computational suitability of the proposed method in MuJoCo using a Franka Research 3 arm pushing a block across surface, without prior knowledge of its changing friction values.

Ambiguity in prescriptive text such as laws, contracts, rules, policies, or recommendations can be both desirable (e.g., if it enables a law to adapt to changing context, such as new technologies) and undesirable (e.g., if it leads to arbitrary discretion or conflict over the law's interpretation). To help navigate these trade-offs, it would be valuable to have methods for quantifying ambiguity that are predictive of such outcomes, but existing measures capture how text is perceived in isolation rather than how it is interpreted in context. In this paper, we propose quantifying ambiguity as the degree of disagreement among a set of stakeholders ("raters") on how alternative prescriptive texts ("policies") apply in each of a set of relevant scenarios ("cases"). Using two policy domains, we present preliminary analysis suggesting that, even when using simulated raters, this approach (i) correlates with real human disagreement, and (ii) is robust to spurious text features. We then apply the measure to existing deliberation datasets to describe the association between policy ambiguity and support for those policies.

Existing manipulation data collection methods typically offer either accurate robot-state acquisition through kinematic-chain-based teleoperation or convenient, portable operation through vision-based approaches. We present DEX-PHD, a portable handheld device that combines these advantages for dexterous manipulation by integrating a dexterous robot hand, continuous per-finger control, and multimodal logging. The system records proprioceptive information, fingertip contact data, and wrist-camera video, while the operator directly controls the robot hand through joint-level mapping. Its untethered design enables data collection across diverse environments. We demonstrate DEX-PHD on contact-rich tasks requiring coordinated multi-finger manipulation and deploy the collected demonstrations through open-loop replay on a physical humanoid robot. The results show that open-loop replay preserves the main temporal patterns of finger motion and task-relevant contact. Therefore, DEX-PHD provides a practical platform for portable, multimodal dexterous manipulation data collection.

Machine learning models are now widely used to predict environmental variable dynamics in domains like hydrology, climate science, agriculture, and ecology. These models are usually compared on fixed benchmark datasets using a single average accuracy score. As models improve, their scores become very close when averaged over large regions and long time periods, so the benchmark can no longer tell them apart, and a single number hides \emph{where} and \emph{when} a model fails. Three failures are common across environmental domains: models get worse as the climate shifts (non-stationarity), models fail at unmonitored sites (spatial extrapolation), and models miss rare but important regimes such as floods and droughts. We present \textbf{EvoBench}, a benchmark framework that shows where a model fails, tracks failures over time, and evolves to focus on model capability gaps. We apply EvoBench to daily streamflow prediction as a case study (\textbf{CAMELS-Evolve}, about 9{,}000 U.S. basins, 1985--2024) and compare an operational physical model (NWM), a conceptual reference (HBV), and several data-driven models. No model wins on every testing scenarios (e.g., in terms of average accuracy, performance at low flows and unmonitored basins). Also we observe that all data-driven models under-predict floods. Additionally, we show that targeted fix based on the reported failures could mitigate weaknesses the benchmark identified in certain scenarios. EvoBench turns model evaluation from a saturated ranking into a moving map of model limitations.

Exploration in physical control environments requires agents to identify actions whose consequences reveal previously unexplored aspects of the underlying dynamics. However, many intrinsic motivation methods rely on immediate prediction errors, state novelty, or local physical changes, which may overlook informative outcomes that emerge across extended temporal horizons. We investigate whether learning structured representations of reachable futures can provide a more effective basis for exploration. To this end, we introduce Hamiltonian Temporal Contrastive Curiosity (HTC$^2$), a framework that combines multi-horizon temporal representation learning with Hamiltonian phase-space structure. A horizon-conditioned contrastive critic learns temporal relationships from future-state pairs sampled within the same episode. The distance between positive-pair representations is then used as an intrinsic reward, encouraging the agent to explore futures that remain insufficiently organized in the learned representation. To incorporate information about physical dynamics, we additionally modulate this reward using bounded multi-step Hamiltonian variation, rather than directly rewarding changes in the Hamiltonian itself. We integrate HTC$^2$ with both Soft Actor-Critic (SAC) and Proximal Policy Optimization (PPO), and evaluate it across six DeepMind Control Suite tasks. Comparisons with ICM, RIDE, E3B, Plan2Explore, and a single-step Hamiltonian curiosity baseline are complemented by state-coverage evaluations, component ablations, and phase-space diagnostics to characterize the contributions of temporal learning and physical structure to exploratory behavior.

The recently envisioned goal-oriented communications paradigm requires machine learning inference to be performed directly on wirelessly transferred data, by splitting neural network computations between the transmitter, the Receiver (RX), and the wireless channel. Such approaches aim to offload large portions of the computation onto the channel itself, which is often controlled by reconfigurable MetaSurfaces (MSs). Typical MS-parametrized systems, however, enable only limited computation offloading in the wave domain, usually corresponding to a single neural network layer, due to adverse channel fading effects and the almost linear operation of the channel and conventional MS models. In this paper, we demonstrate that an eXtremely Large (XL) Multiple-Input Multiple-Output (MIMO) system equipped with Stacked Intelligent MSs (SIM) at the RX side, together with a single antenna and its accompanying radio-frequency chain, may act as a two-layer neural network operating exclusively over the air. The first (uncontrollable) network layer corresponds to the transformations offered by wireless fading, while the second (controllable) layer is implemented via SIM operations. In particular, only the first, front-end MS layer needs to exhibit nonlinear response to serve as an activation function. Leveraging the Extreme Learning Machine (ELM) framework, the proposed XL MIMO system architecture is shown to be a universal approximator, and the optimal linear-SIM-based combining is derived in closed form, allowing for fast training even and partial signal observability. Numerical evaluations across diverse datasets showcase that the proposed XL-MIMO-ELM framework achieves classification accuracy comparable to idealized digital models, thereby, proving the viability of low complexity, wave-domain neural computations.

Many decision-making tasks are inherently multi-objective. Interactive evolutionary multi-objective optimization algorithms (iEMOAs) are particularly well-suited to help the decision-maker find their most preferred solution. However, many of these algorithms employ black-box preference models. For high-stakes decision-making tasks, like healthcare and defense, a high level of transparency is needed, for instance to support accountability. We propose an interpretable iEMOA which uses active learning to efficiently learn a tree-based preference model directly from pairwise comparisons. The survival operator of NSGA-II and NSGA-III is extended with the learned preference model to guide the algorithm towards the region of interest. An extensive experimental study compares the proposed method to a state-of-the-art black-box model and a state-of-the-art interpretable model. Results indicate that our proposed model outperforms the alternatives in terms of finding the most preferred solution at the cost of a slightly worse average final population. Finally, a case study on a real-life engineering problem illustrates how inspecting the learned model reveals which objectives drive the decision-maker's preferences.

Safety fine-tuning can make undesirable model behaviors sufficiently rare that measuring their frequency by direct sampling becomes prohibitively expensive. We introduce Amplify, Locate, Estimate (ALE), a method for estimating such rare failure probabilities efficiently. ALE uses the change in token log-probabilities induced by the safety fine-tuning to define an amplification direction, along which we increase the failure rate for cheap measurement. We develop a score-function estimator that uses the selected amplified batch to recover the local slope of the log failure rate from the token probabilities of the sampled failures, enabling extrapolation back to rare target rates. We also use the amplified batch to obtain an importance-sampling estimate. Across 41 evaluation targets, ALE's estimate-to-reference ratios have a median close to one, with 37 ALE estimates within a factor of three of reference values obtained from independent direct measurements. Ablation tests show that with a budget of up to $\sim$1000 labeled rollouts, in addition to a cheap unlabeled scan, ALE estimates failure probabilities for Qwen3-8B fine-tunes down to about $10^{-5}$, where direct measurement can require hundreds of thousands of rollouts, as the required sampling budget scales inversely with failure probability at fixed relative precision.

Determining when athletes reach peak performance is a central question in sports science, with implications for talent identification, training design, and cross-discipline comparisons. Statistical inference is challenging because peak age is not directly observed: competition records are noisy and selective measurements of underlying performance capacity and are often limited to incomplete observation windows. We develop a hierarchical Bayesian joint modeling framework that links a longitudinal model for athlete-specific performance trajectories to a regression submodel for latent peak age. The longitudinal component models observed competition performances as noisy measurements of athlete-specific performance trajectories over age. The peak-age component then treats the maximizer of each trajectory as a latent response in a parametric accelerated failure time (AFT) model, allowing athlete-level covariates to explain earlier or later peak age. The hierarchical structure borrows information across athletes, while joint posterior inference propagates uncertainty from trajectory reconstruction to both individual peak-age estimates and population-level covariate effects. Simulation studies show that, as observation windows become less informative, the joint model preserves substantially smaller regression bias and reflects the information loss through wider uncertainty, whereas pseudo-peak and two-stage alternatives exhibit larger bias. An application to World Athletics ranking data illustrates how the joint model estimates discipline-specific peak-age patterns and covariate effects under incomplete and irregular observation windows. Beyond sports, the proposed framework applies broadly to problems that couple longitudinal trajectories with function-derived latent features.

Automated fake news detection is vital on social media, where fake news is easily rewritten to fool detectors. However, existing approaches typically rely on superficial writing cues or single-style text rewrites, resulting in poor robustness under diverse cosmetic modifications. We construct MANTIS-Bench, a multi-axis presentation-attack benchmark that pairs original news with rewritten views across writing style, emotional expression, and dissemination framing. We further propose MANTIS, an instance-adaptive adversarial framework enabling dynamic worst-case training tailored to each news article with zero inference-time overhead. This reframes fake news defense from heuristic format smoothing into targeted, fact-invariant credibility assessment. Experiments on MANTIS-Bench show that MANTIS reduces the Union Attack Success Rate (UASR) by 3.55 percentage points relative to SheepDog on GossipCop and consistently outperforms prior single-axis defenses on Macro-F1 while maintaining competitive clean accuracy, demonstrating the effectiveness of multi-axis and instance-adaptive worst-case training.

As AI coding assistants generate more code, the need to review it for security flaws grows. Existing approaches leave a gap: static analyzers identify program sites but offer little explanation, while language-model explanations can be difficult to verify. Small open-weight code models could make security review available without sending proprietary code to a hosted service. We study whether fine-tuning helps these models turn security evidence into explanations that developers and tools can check. We present secSemScan, which pairs a static analyzer with a locally deployable model that writes a schema-constrained security monologue naming program sites, assigning a CWE, and giving a machine-parseable verdict. We also introduce OGSM (Oracle-Grounded Security Monologue), a dataset of 1,993 C/C\texttt{++} functions covering 22 CWEs in the taint and lifetime families. Each record pairs a function with analyzer-derived or fix-diff-derived evidence and a monologue. We fine-tune three open-weight code models in two stages, first on the output format and then on the monologue content. Because evaluation prompts supply the verdict and evidence, our results measure how well models use that information, not whether they detect vulnerabilities independently. On 355 test records, fine-tuning raised Qwen2.5-Coder-7B's verdict accuracy from 4.8\% to 96.6\%; DeepSeek-Coder-6.7B reached 78.9\%, and StarCoder2-7B reached 43.7\%. Qwen2.5-Coder-7B cited 87.5\% of Tier~1 oracle sites, and every line it cited belonged to the oracle evidence. The results suggest that fine-tuning helps small models produce more structured, evidence-linked security reasoning.

We evaluate ErouVe, a vehicular eco-routing protocol in which roadside units (RSUs) collect per-segment travel-time and CO2 reports from passing vehicles and answer route requests with the lowest-cost alternative, across a range of road architectures and traffic scenarios. We conduct the evaluation with a working implementation of the protocol inside Veins, adapted to drive SUMO's own router rather than a hand-rolled shortest-path search, on a network offering a short, signal-controlled route and a longer, signal-free bypass, under two traffic scenarios. This evaluation reveals clear strengths and weaknesses of the protocol. Under light, uncongested traffic, ErouVe's fleet-wide effect is light. Under traffic dense enough to stress the signalized route's throughput and produce a real, growing queue, the same protocol succeeds a medium fleet-wide CO2 reduction, with diverted vehicles individually saving even more fuels. We use these findings to propose three targeted corrective methods; a knowledge-gated cost-acceptance rule, a diversion-locking rule, and a distance-normalized CO2 cost function, that respectively address a cold-start knowledge asymmetry, decision flapping under noisy crowd-sourced data, and a per-second CO2 metric that rewards idling. Each of these weaknesses are generic pitfalls for any crowd-sourced routing signal and all proposed methods measurably restore protocol effectiveness.