AI agents that act as statistical analysts evolve through model upgrades, prompt edits, and new tools, but each change risks inflating Type I error. This paper introduces continual validation: before any update is deployed, it is frozen and tested on newly simulated null data using exact binomial confidence bounds with a decaying per-update risk budget. The result is a lifetime guarantee that no invalid version passes, which matters because classical fixed-procedure guarantees break under adaptive, open-ended evolution.

Pairwise comparison data power modern preference learning, but disagreement in such data can mean persistent heterogeneity, random noise, or hidden non-binary responses. This paper develops a nonparametric latent semantic model with three querywise quantities—aggregate preference $\mu_q$, persistent disagreement $\tau_q$, and non-binary mass $\alpha_q$—and proves sharp identifiability frontiers. Anonymous binary feedback point-identifies only $\mu_q$, while same-annotator repeats identify $\tau_q$ via covariance and abstention identifies $\alpha_q$ directly; only their combination recovers the full semantic triple. The results imply that semantics-aware decisions like denoising versus personalization cannot be justified from anonymous forced-choice data alone.

SAM3RL replaces the hand-crafted FIFO memory rule in SAM-based visual trackers with a lightweight controller learned by reinforcement learning. It frames memory-bank eviction as a Markov Decision Process and trains a 3.2 M-parameter MLP with PPO on bounding-box LaSOT sequences to maximize tracking quality. The resulting policy sets a new state of the art on 10 of 12 benchmarks for both SAM 2 and SAM 3 backbones, while remaining fully online and running without extra feature extraction.

Offline RL promises to outperform the data but is notoriously unstable, whereas imitation learning is stable yet bounded by data quality. This paper introduces Advantage Guidance (AG), a policy extraction method that retains a simple supervised-learning objective—training a flow policy conditioned on advantage estimates—while correcting the failure mode where naïve conditioning is ignored by the network. At test time, AG uses a derived contrastive diffusion guidance to target the KL-regularized optimal policy $\pi_{\mathrm{KL}}^*(a \mid s) \propto \hat{\pi}(a \mid s)\exp(A(s,a)/\beta)$. On 26 long-horizon sparse-reward manipulation tasks, the method outperforms strong actor-critic, supervised-learning, and diffusion-guidance baselines.

Population health stratification traditionally groups patients by clinical complexity, yet these descriptive frameworks offer little guidance on how to prioritize limited care resources. PROMETHEUS reformulates this task as a multi-treatment causal ranking problem: instead of estimating precisely calibrated treatment effects for every patient, it learns a global priority ordering over patient–profile opportunities by combining cross-fitted doubly robust pseudo-outcomes with a Siamese contrastive neural network. This matters because it separates clinical need from actionability, allowing health systems to identify which eligible patients should receive more intensive care first when capacity is constrained.

ActMOF addresses the challenge of optimizing metal-organic framework (MOF) syntheses when experimental data are sparse and outcomes noisy, by introducing an experimentally grounded benchmark for comparing active-learning policies. Starting from only 95 measured microwave syntheses of an aluminum-carboxylate MOF, the authors emulate a landscape of over six million encoded conditions and evaluate 39 policies—Gaussian processes, tree ensembles, and large language models—before transferring the best strategy to six wet-lab syntheses. The demonstration that all six campaigns yielded crystalline products within a 12-experiment budget makes the work relevant as both a reproducible testbed and a proof of concept for transferable, AI-guided reticular chemistry.

Looped Mixture-of-Experts models reuse parameters across depth, yet prior work treats each loop visit identically. This paper shows that loop visits learn strongly asymmetric residual scales—the first visit updates representations far less than the second—and proposes LatentLoop, which compresses the first pass into a latent sequence and reallocates saved FLOPs to model width. Extensive scaling experiments up to $7$B parameters show consistent validation and downstream gains, framing asymmetric compute allocation as a principled way to improve looped MoE efficiency.

Goal-conditioned RL in complex mazes suffers from a geometric mismatch: Euclidean embedding distances model straight-line proximity, yet walls can make actual route distances highly tree-like and deceptive. This paper introduces Hyperbolic CRL, a drop-in replacement of the critic's Euclidean distance with a hyperbolic geodesic distance on the hyperboloid. Its central finding is that this geometric alignment allows a depth-64 hyperbolic network to outperform a 1024-layer Euclidean baseline on hard humanoid maze tasks while training more than $12\times$ faster and using roughly $13\times$ fewer parameters.

Foundation models for dynamical systems can roll out unseen PDEs and ODEs, but it is unclear whether they genuinely learn in-context forecasting strategies or merely memorize. This paper proposes Test-Time Forecasting (TTF), a framework in which looped transformers use recurrent depth to perform gradient-descent iterations on a latent forecasting objective at inference time. A lightweight TTF model trained only for one-step prediction can generalize to harder, out-of-distribution system dynamics during autoregressive rollout, suggesting that recurrent depth offers a principled, parameter-efficient path to in-context learning for continuous dynamics.

Predicting which cellular region a mobile user will visit next lets operators cache content and allocate resources proactively, yet a predictor that never abstains commits to choices even when its evidence is weak.
This work applies MetaModelSelect, a lightweight post-hoc metamodel that scores the trustworthiness of a frozen RNN's top-1 next-location decision using only its output logits: a learned embedding of the predicted region, the four largest softmax probabilities, and a global concentration statistic.
By filtering predictions through this learned correctness score rather than retraining the backbone, the paper shows how to trade coverage for selective risk in a compact, deployable wrapper.

Traditional student models only use responses to standalone questions, ignoring the rich, fine-grained evidence embedded in tutoring dialogues. This paper proposes URF-QD, a framework that rewrites tutor-posed mathematical subtasks as self-contained questions and models them jointly with original items on a shared item response theory scale. The key novelty is a relative-order regularization that uses tutor pedagogical moves to constrain consecutive subtask difficulties, enabling dynamic tracking of student ability across both questions and dialogues. The authors show consistent improvements on two real-world math datasets, though the approach leans heavily on LLM annotations and modest sample sizes.

Batch black-box optimization requires selecting many parallel queries across only a few expensive rounds, yet conventional Bayesian optimization suffers from costly joint acquisition maximization and per-task surrogate refitting. This paper proposes Flow Matching for Thompson-Sampling Refinement (FMTR), an amortized approach that offline-learns a continuous, history-conditioned proposal distribution over query locations from which variable-size batches can be directly sampled at test time with fixed parameters. By marrying in-context adaptation to a generative flow model, FMTR aims to eliminate online acquisition optimization while preserving sample efficiency.

This paper studies whether learning an epsilon-optimal average-reward policy from a single continuing trajectory must pay an extra navigation cost compared with generative access. It develops an adaptive algorithm that learns a low-cost sampling allocation via inverse-count exploration and converts nonuniform samples into an all-state policy certificate using a transient-corrected quantile bound. For broad model classes and sufficiently high accuracy, it proves that the worst-case sample complexity matches the generative lower bound up to logarithmic factors, removing the extra diameter factor that plagued prior continuing-access bounds.

Vision-language models are moving onto battery-powered edge devices, yet their total-system energy footprint on NPU-equipped single-board computers has remained largely unmeasured. This paper presents a hardware-instrumented study of Qwen2-VL-2B-Instruct inference on a Raspberry Pi 5 with a Hailo-10H NPU and a Jetson Orin Nano Super, separating fixed request energy from per-token decode costs. It shows that NPU offloading cuts Pi energy per token by roughly two-thirds, that text generation dominates vision encoding at typical response lengths, and that the Pi-NPU and Orin-GPU stacks have comparable per-token energy with a crossover near 40 output tokens. The findings are practically useful for engineers choosing deployment stacks under power and thermal constraints.

Molecular communication simulators are computationally expensive, so machine learning models are often trained to estimate performance metrics like round-trip time, but protocol evaluation proceeds sequentially from simple to complex settings and causes catastrophic forgetting when models are fine-tuned. This paper proposes Affine Memory Replay (AMR), a continual learning method that stores past simulation samples and synthesizes additional replay samples through affine interpolation and extrapolation to regularize a small neural network. AMR aims to retain knowledge from earlier experimental settings while adapting to new ones, offering a practical way to amortize simulation costs in molecular communication research.

This paper tackles offline parameter identification for a non-ideal DC-DC buck converter using Physics-Informed Neural Networks (PINNs). The core idea is a staged training schedule that progressively unfreezes physical parameters—inductance, capacitance, and parasitic resistances—to improve convergence and interpretability. The work matters because accurate component-level estimates enable condition monitoring and predictive maintenance without intrusive frequency-response tests.

Training-free inference-time scaling aims to improve LLM reasoning without updating model weights or relying on external reward models. Existing reward-free methods sharpen the full sequence distribution, which can suppress diverse reasoning paths that support the same correct answer, and suffer from path collapse in Sequential Monte Carlo samplers. This paper proposes Decoupled Marginal Sharpening (DMS), a two-stage framework that separately controls reasoning-path concentration and answer-marginal consensus via independent exponents, paired with a novel Fully Adapted Annealed SMC sampler. The authors report the highest mean pass@1 across three base models and four benchmarks, including mathematics, science, and code tasks.

This critical review reframes interactive segmentation, active learning, and machine teaching through a shared information-theoretic lens. It separates target information gain from predictive distortion reduction and human effort, exposing why predicting a teacher's clicks does not by itself improve reusable learning. By formalizing corrective interactions as nested memory records, the paper shows that current-mask refinement and reusable-model training are distinct objectives requiring different evaluation assumptions.

CktHGVAE tackles analog/mixed-signal CMOS circuit automation by learning a joint latent manifold of design parameters and transistor-level operating states. The core idea is a symmetric heterogeneous graph VAE that treats SPICE-derived state variables and design variables equally as inputs and outputs, enabling closed-loop design verification without a separate surrogate model. This is significant because prior graph learning approaches for analog circuits largely rely on macromodel abstractions or topology generation, leaving fine-grained operating-state modeling underexplored.

This paper asks how shallow transformers learn on unbounded-length inputs and whether layer-wise training can be decoupled. It proves that a one-step attention-then-output schedule fails to learn the occurrence language, while a three-phase output-attention-output schedule succeeds in both population and finite-sample settings. The finding is significant because it formally explains why attention weights often lag behind the output layer during training, challenging the feedforward intuition that feature learning must come first.