Papers

Keeling et al. [Keeling2026Generative] report on a case study recently conducted by Redgrave LLP in which they assert that Relativity's Generative AI Review tool (aiR) achieved 88% recall while Relativity Assisted Learning (RAL), its machine-learning tool, achieved 64% recall. These results are not comparable because the methods employed have different objectives, because of inter-assessor disagreement, and because they represent different recall-precision and cost-benefit trade-offs. Our analysis of the Redgrave data indicates that, when these differences are taken into account, aiR and manual review are comparable. Furthermore, established machine-learning models employing Continuous Active Learning (CAL) show state-of-the-art effectiveness, and superior efficiency, when trained only on a sample of 2,000 first-pass-reviewer responsiveness determinations.

We develop a comprehensive theoretical analysis of Matrix Shampoo for tracking time-varying optima in nonstationary stochastic optimization. We derive finite-horizon high-probability bounds for average tracking error that disentangle initialization, optimizer drift, momentum-induced drift penalty, stochastic noise, and adaptive preconditioning, with explicit dependence on momentum parameter \(\beta_1\) and stabilization parameter \(\varepsilon\). These bounds reveal an explicit drift–noise and stability–geometry tradeoffs: noise-dominated regimes favor larger $\beta_1$, which provides stronger first-moment averaging and greater suppression of gradient noise, whereas drift-dominated regimes favor smaller $\beta_1$, allowing faster adaptation to moving optimizer. Larger $\varepsilon$ reduces the preconditioner-perturbation term while weakening Shampoo's ability to exploit matrix geometry, thereby reducing adaptivity. We establish a matching minimax lower bound under squared path-variation constraint, proving the information-theoretic optimality of Shampoo over the class of causal optimization procedures. Under a constant stepsize, we establish high-probability bounds for the average projected stationarity gap along general nonstationary nonconvex objectives. We validate our theoretical findings on both synthetic and multiple real-world problems.

Online motor imagery-based brain-computer interfaces (MI-BCIs) typically decode continuous EEG using fixed-length sliding windows that may be misaligned with neural events. Existing window optimization methods often evaluate windows independently, overlooking temporal dependencies and potentially producing fragmented or physiologically implausible selections. We propose Temporal Structure Embedded Reinforcement Learning (TSE-RL), a framework that incorporates prior knowledge of MI event structure into a constrained Markov decision process for adaptive window optimization. First, TSE-RL assigns each window an explicit semantic role through Discard, Onset, Sustain, or Offset actions, extending binary retain/discard decisions to event-level alignment. Second, a finite-state machine masks invalid actions to enforce the temporal order of MI phases and promote physiologically plausible window sequences. Third, a Transformer-based history encoder captures dependencies across windows, improving training stability and computational efficiency compared with recurrent architectures. TSE-RL is plug-and-play and can be integrated with diverse MI classifiers. Experiments on multiple benchmark datasets demonstrate consistent improvements in decoding performance in both cross-session and cross-subject settings. Visualization analyses further suggest that the learned window selection policy aligns with neurophysiological principles.

Machine learning engineering (MLE) agents are emerging as a new paradigm for automating end-to-end ML pipeline development. Using large language models for code generation and predictive performance as a machine-gradable feedback signal, these agents iteratively synthesize, execute, and refine pipeline code. Yet current systems are largely evaluated on pre-built, single-table benchmarks that abstract away key data management challenges. This tutorial introduces the foundations of MLE agents and highlights research opportunities for the database community. We cover how agents combine program synthesis with iterative and evolutionary optimization, and discuss open challenges in efficient pipeline execution, data discovery and augmentation over heterogeneous sources, reliability and debugging, as well as fairness and compliance. We connect these challenges to existing data management techniques and conclude with a hands-on session using contemporary MLE agents.

Bayesian Networks (BNs) are established tools for letting agents reason under uncertainty in sequential decision-making. They split a joint distribution into local conditional mechanisms arranged on a directed acyclic graph, which makes them modular, interpretable, and ready to perform probabilistic inference. Their classical formulation, however, has limitations in both expressiveness and scale: each conditional is usually stored as a table or a linear-Gaussian form, and inference evaluates the graph one node and one query at a time. In this paper we introduce Neural Bayesian Networks (NBNs), in which each conditional is a differentiable, tensor-valued mechanism, while the graph and its factorization are unchanged. NBNs retain the modularity and local mechanisms of BNs, can represent any learnable conditional distribution, and run natively on GPUs. We also release NBN, a PyTorch-based library for learning and inference in NBNs, together with a public benchmark. Our evaluation shows gains in both expressiveness and scale: NBNs fit distributions that classical BN libraries struggle to represent, and on large networks and large batches of queries they run two orders of magnitude faster thanks to GPU parallelization. NBN is available at https://github.com/Giovannibriglia/NeuralBayesianNetworks.

Emotions are expressed on a continuum and are naturally described in the valence-arousal (VA) space. Existing VA-based image editors are typically training-based and domain-specific, limiting them to VA regions seen during training and making them brittle across datasets and domains. Meanwhile, despite producing high-quality edits, state-of-the-art foundation image editors are not continuous affective controllers: non-linear emotion representations in text encoders makes prompt interpolation poorly calibrated, so accurate and continuous VA control remains unreliable. To fill this gap, we introduce Barycentric Guidance, a training-free framework that turns pretrained foundation image editors into continuous affective controllers over VA, requiring no finetuning, extra data, or model changes. Our method maps any target VA point to simplex-based guidance weights, enabling VA controllability in three aspects: (i) accurate targeting of VA states, (ii) broad coverage across the VA plane, and (iii) continuous control. Across three face datasets, we improve VA target accuracy by 25-29%, expand VA-plane coverage by 28-50%, and increase expression diversity by 29-49%, over the strongest baseline. Unlike prior domain-specific methods, our training-free approach exploits the foundation model's inherent versatility, enabling generalization beyond human faces to animals, artworks, and complex scenes within a single framework. Code is available at https://github.com/harveymannering/BarycentricGuidance.

Inferring tumor-growth laws from sparse, noisy observations requires more than accurate curve fitting. We study an AI-assisted workflow involving symbolic regression, mathematical admissibility checks, coefficient refitting, and large language model (LLM) ranking. Synthetic Gompertz data allow trajectory fit, forecasting, and recovery of the governing growth law to be assessed separately. A dense noiseless control recovers the generating family, but sparse noisy sampling frequently leads to alternative forms. Among retained candidates in the tested configurations, physics-informed neural network refitting often increases growth-law error, whereas classical nonlinear least squares usually reduces it. A separate exploratory analysis of archived LLM rankings shows no consistent improvement over selection by training error and identifies errors in mathematical reasoning. Plausible saturation, accurate trajectory fits, and convincing explanations can therefore coexist with incorrect growth laws. Mathematical validity and law recovery require evaluation alongside predictive performance.

We introduce FORGE (Fast On-chip Rotor Generation for Experts), a procedural weight generation system for Sparse Mixture-of-Experts (MoE) inference. Rather than storing and loading every expert weight tensor from HBM, we show how a compact accelerator inside each Streaming Multiprocessor can regenerate them on-the-fly from $O(\log^2 d)$ rotor parameters per projection. Per replaced layer, gigabytes of stored expert weights are approximated with under a megabyte of seeds. This addresses the dual memory bottleneck of autoregressive Sparse MoE decode, where per-expert FFN operations are HBM-bandwidth-bound at small routed-token counts and the aggregate expert footprint forces multi-node Expert Parallelism (EP) with all-to-all traffic. We realize the approach as the Rotor Generation Unit (RGU), a fixed-function datapath that generates FFN weight tiles without competing with Tensor Core execution. With 30% of the MoE FFN layers replaced across five Sparse MoE models, we estimate 1.3-2.8x higher system throughput at matched GPU count and 1.10-1.15x lower per-user decode latency at short context. In exchange, downstream accuracy on three model families falls below that of the same model fine-tuned without rotor replacement, and the replacement ratio controls this tradeoff. FORGE also expands KV cache capacity by 1.2-2.6x. It is agnostic to the bit width at which weights are stored, so it can compose with quantization. The RGU is fully implemented/verified in SystemVerilog at ~0.9% SM area overhead (0.032 mm^2/SM, 7nm).

Continuous symmetry discovery is typically formulated as a separate estimation problem for each function or dataset. We instead investigate whether the map from a function to its symmetry structure can be learned across a family of functions from finite observations. Using the infinitesimal differential characterization of continuous symmetries, we formulate this map as an inverse symmetry operator, analogous to an inverse PDE problem. For global Lie-group symmetries, we establish a representation-theoretic population characterization: the symmetry Lie algebra is the nullspace of a positive-semidefinite operator constructed from representation coefficients and the infinitesimal group action. We also show that the inverse symmetry operator is equivariant under coordinate changes through the adjoint action. This characterization separates the geometric symmetry map from finite-sample estimation, motivating a hybrid approach that amortizes statistically difficult components while retaining analytic structure. On SO(5) benchmarks, amortized equivariant estimation of harmonic coefficients improves symmetry recovery from sparse observations, whereas ordinary least squares suffices for SL(2,R). We extend this viewpoint to local symmetries, where symmetry directions form smoothly varying tangent distributions rather than a fixed Lie algebra. Under the Lie-bracket consistency condition, Frobenius' theorem guarantees local integral submanifolds along which observables remain unchanged. We recover this geometry using a hybrid estimator that learns observation weights while preserving analytic differential and geometric recovery. Across global, non-compact, local, and physically motivated synthetic benchmarks, our results support learning the statistically difficult finite-sample components while preserving the known geometric structure of symmetry recovery.

Autoregressive electronic health record (EHR) foundation models predict the next event token in patient trajectories: sequences of structured medical events. A typical pretraining strategy is to concatenate all trajectories into a global stream and sample token windows from it. This gives patients with longer sequences more training exposure and allows tokens to attend across patient boundaries. We argue that these choices create a misalignment between pretraining and downstream clinical tasks. To explore this, we investigate how the composition of pretraining sequences affects clinical prediction performance. Using the global-stream approach as a baseline, we propose alternative sequence composition methods that control how patient trajectories contribute to pretraining sequences. Across four MIMIC-IV v2.2 tasks, our best method improves macro AUROC, AUPRC, F1, and balanced accuracy, with relative gains of 9.3% in F1 and 2.7% in AUPRC under a fixed training recipe. We find that these gains generalize across four additional evaluation settings, including the INSPIRE and eICU-CRD datasets, with improvements in macro AUROC, AUPRC, and F1 score. Our results show that simply changing how EHR data are presented to the model during pretraining can improve downstream performance, without changing the model or collecting additional data.

Simplicity bias, which can arise from sequential feature learning during training, has been proposed as a general explanation for the success of deep learning. An important driver of this phenomenon is saddle-to-saddle behavior, in which the gradient flow trajectory remains close to a sequence of saddle manifolds while learning one feature at a time. However, Muon, a state-of-the-art optimizer based on spectral descent, does not exhibit such dynamics in the full-batch setting and instead learns features in parallel. How is it then that Muon is so successful in deep learning settings? We show that stochastic gradient noise fundamentally changes this behavior: small-batch noise restores sequential feature learning, providing a mechanism for simplicity bias that is absent from full-batch Muon. We explain this transition through stochastic smoothing of the spectral descent operator underlying Muon. When gradients are small relative to the noise scale, the expected spectral update becomes approximately proportional to the actual gradient, which restores gradient-flow-like dynamics. We formalize this mechanism for deep linear models by deriving the effective continuous dynamics of noisy spectral descent. Experiments on low-rank matrix reconstruction, vision transformers, and large language models (LLMs) support these findings.

Reliable agentic tool use requires training data that pair diverse multi-step workflows with executable evidence that the requested outcomes were achieved. Scaling either trajectory coverage or verification alone is insufficient as new environments introduce distinct states, rules, and dependencies that must also be reflected in the success criteria. To address this challenge, we introduce Davinci-Tool, a 30B tool-use model, and Davinci-ToolData, an execution-grounded dataset designed to scale environment coverage and outcome verification together. Davinci-ToolData preserves restorable states, execution traces, outcomes, and verifiers as a unified substrate for supervision. Two complementary pipelines populate this substrate: Execution-Grounded Task Synthesis (EGTS) derives verified demonstrations from independently authored systems, and Executable Environment Synthesis (EES) synthesizes auditable environments, state-grounded tasks, and executable verifiers across single and coordinated domains. These pipelines yield 30,395 verified trajectories spanning 2,260 invoked tool names. We first train on these demonstrations with supervised fine-tuning, then reuse EES states and verifiers to compare alternative actions and construct outcome-grounded preferences for Direct Preference Optimization. Davinci-Tool improves Qwen3-Coder-30B-A3B-Instruct from 52.4 to 88.2 on $\tau^2$-bench, from 30.4 to 62.8 on BFCL v4 multi-turn, and from 16.9 to 56.2 on MCP-Atlas, outperforming Qwen3-Coder-480B-A35B-Instruct on all three benchmarks with only 1/16 of its total parameters. Code is available at https://github.com/MassLab-SII/Davinci-Tool.

Reinforcement learning (RL) post-training with generative reward models (GenRM) has become a central recipe for improving large language models (LLMs), and it is also increasingly resource-intensive: policy rollout, GenRM-based reward evaluation, and policy training all compete for GPU resources. However, existing colocated and disaggregated systems rely on static resource allocation for these roles. Such allocation cannot adapt as each role's workload changes during RL and therefore leaves a role's resources underutilized. We present PivotRL, an RL post-training system that dynamically shifts resources among different roles according to their runtime demands. PivotRL introduces a resource-shiftable runtime built on a Global Parameter Storage, together with a unified mechanism that efficiently shifts role-specific context. A shift-aware scheduler enables sequence migration upon a shift, while a planner considers the aggregate throughput benefit and the shift overhead to determine when and how to shift. We evaluate PivotRL on a 128-GPU cluster using four model groups spanning dense and mixture-of-experts models. PivotRL enables resource shifting within seconds and improves end-to-end throughput by 1.81-2.05 times over colocated deployment and 1.50-1.95 times over disaggregated deployment, while preserving RL convergence.

Organizations pairing a human and a machine on a task decide, often by default, which executes each instance and what happens when both do. We account for what this attains under the squared rule, with performance distributed over instances to capture generalization difficulty. A gross stake lies between the skills' mean and best per instance. It splits at the best single skill into a systematic part, known from separate assessments and secured by commitment, and a residual, the expected value of perfect information on which skill each instance favors, known only from paired observation and split further by classes. Skills set the stake; workflows decide the take, via two channels. Routing directs work, weighting or selecting responses; with binary truth it is capped by the best skill per instance. Reading responses as joint evidence is not routing; what they jointly reveal caps it. Synergy lets skills interact before responding; bounded by no skill either way, its lift over independent responses must clear a break-even and can fall below one. As algorithms, arrangements form a ladder stating what each reads, holds and records, and a procedure whose checks keep earlier gains when class means are known and lose little on finite records given a margin. The theory extends to many skills, each worth what it adds where it beats all others held, and to competing skills, where at parity a single failure's allowed cost decides. In a public chest-radiograph study with 227 radiologists, they are near parity with the model and the stake is mostly residual; assisted reading, no better on average originally, returned less than averaging both responses; recalibrating the model alone beats any deployable arrangement of raw responses. Organizations can learn from their records what is at stake, which arrangement collects it and what evidence each step needs; pricing costs and errors is future work.

This paper presents GpFlash, a Gaussian process library that is focused on fast training and posterior evaluations, minimal memory usage, and scalability to high-dimensional problems. GpFlash was compared to the popular GP libraries scikit-learn, GPy, GPflow, GPyTorch, and SMT on a MacBook Pro and the supercomputer Trillium. GpFlash was found to use less memory and to be faster for training and posterior evaluations relative to the other GP libraries for all cases considered. For example, when using $\nx = 10^4$ training points in a 10-dimensional space GpFlash required 9.6 to 35 times less memory and was 2.6 to 54 times faster than the other GP libraries for the evaluation of the marginal log likelihood and its gradient. The memory reduction was achieved by only using one matrix of size $\nx \times \nx$ to store different symmetric matrices in its lower and upper triangular entries. The computational speedup was the result of designing algorithms to access memory efficiently, using BLAS and LAPACK methods extensively, and factoring kernel equations so that expensive exponential calculations could be reused. GpFlash was more than 440 times faster than all of the other GP libraries considered when evaluating the posterior mean together with its gradient for the case with $10^4$ training points and 10 dimensions. GpFlash is available at https://github.com/marchildon/gpflash.

Diffusion and flow-matching models may struggle to accurately represent distributional tails, especially for heavy-tailed data. Existing approaches often use heavy-tailed noise or modify the generative dynamics, requiring dedicated retraining and more difficult estimation. We take a simpler approach: rather than relearning the full distribution, we correct only the model's behavior in the extreme regime. Our method builds on the radial-angular decomposition of multivariate heavy tails, where radius and direction become asymptotically independent. This allows us to preserve the directional structure learned by the pretrained model and correct only its radial tail. We estimate the tail onset from reference and generated samples, then rescale generated extremes so that their radial survival matches the target distribution. The resulting method is a lightweight post-training correction applied to a frozen diffusion or flow-matching model, requiring neither retraining nor changes to its sampling dynamics. Experiments on challenging synthetic heavy-tailed distributions and real weather and wildfire datasets show that this simple correction substantially improves extreme-tail coverage using only a small number of fitting samples.

Standardizing a rare feature can help stochastic gradient descent learn its effect within a finite budget, but it also changes the noise cost of learning. We characterize the lowest rare-subgroup risks attainable by standardized and unscaled training when each chooses a nonnegative constant learning rate and a deterministic stopping time under the same observation cap. For two independent binary features, we derive both optimized risks as prevalence vanishes while the expected number of rare observations remains finite. Uniform control over rates and stopping times reduces each limiting oracle to a scalar minimization. With positive noise, the preferred scaling changes exactly once as we move signal to the rare feature while the limiting risk at initialization stays fixed and positive. The formulas also explain how noise and budget affect the comparison. Separate finite-prevalence bounds favor standardization on a rare-dominant signal family and unscaled training on a common-dominant family. The results concern final iterates with population preprocessing and fresh observations. Exact-moment searches assess the approximation at finite prevalence. In a controlled study, both orderings persist under observation reuse, estimated preprocessing, and intercept fitting when tuning choices are fixed before evaluation. Selecting a predictor after observing training can reverse an ordering, even when selection uses exact population risk.

Real-world systems often serve many objectives but can deploy only a few options, whose performance must be learned from noisy feedback. Modeling each option as an arm in a multi-objective bandit, we introduce few-for-many best-arm identification, which seeks a small set of arms minimizing the worst weighted degradation from the objective-wise optima, rather than the best arm for every objective. We consider both known objective groups and latent groups specified only by a representative budget, and propose GFM-Elim and AFM-Elim with instance-dependent sample-complexity guarantees. For Gaussian rewards, we further derive a lower bound for the latent-group setting and show that it is asymptotically attainable with high probability by an oracle-based Track-and-Stop procedure. Our results show that sharing observations across objectives and ignoring irrelevant objective-wise distinctions can substantially reduce the sampling cost, as confirmed by experiments on synthetic and real-world tasks.

When does a shared representation become insufficient for jointly supporting multiple tasks? This question lies at the heart of multi-task learning models. Although task interference in shared representations has been widely observed, its onset, measurable form, and capacity boundary remain poorly understood. In this paper, we revisit this problem through the lens of Bell's theorem. We show that if four binary task outcomes can be jointly represented by a common shared representation, their correlations must satisfy the classical CHSH bound $|S|\leq2$. We introduce NCnet, a controlled multi-task model for studying how representational capacity affects task compatibility. Experiments reveal a clear capacity-dependent transition. Once nontrivial task fitting emerges, constrained capacity can yield $|S|>2$. As capacity further increases, these exceedances progressively disappear and $S$ returns to the bounded region. This systematic pattern provides evidence that limited representational capacity is an important source of the observed incompatibility. We observe similar transitions across CNNs, PLMs, and LLMs, suggesting that the phenomenon extends beyond the controlled NCnet setting. We refer to this capacity-dependent transition as the Bell Boundary, providing a measurable diagnostic for representation bottlenecks, task incompatibility, and capacity allocation in multi-task models.

Deepfake speech detectors are typically evaluated as binary classifiers on fixed benchmarks, but deployed systems face evolving attacks, codecs, speakers, and recording conditions, with labels unavailable during operation. This makes it difficult to know when a detector operates in familiar conditions and can be relied on. We propose an unsupervised drift monitor that operates on the internal embeddings of a deployed detector. The monitor compares each input against bona fide and spoof references using regularized Mahalanobis distances. These distances define trusted, ambiguous, and novel regions of the detector embedding space, allowing unfamiliar inputs to be routed for review. We evaluate the monitor on four modern detectors on three datasets. The monitor raises alerts for dataset mismatches and substantially reduces the security-critical case of accepting deepfakes at the cost of reduced automatic coverage. The results also show that drift affects both classes: bona fide recordings can move away from the training reference, increasing alert rates even in in-domain settings. Our findings suggest that deployed deepfake speech detectors should be treated not only as a binary classification, but also as a reliability problem.

A release decision about a fine-tuned model is an operational commitment about the next rerun of its training recipe. Yet standard benchmark reporting gives only the mean accuracy of past seeds. These target fundamentally different questions: release decisions depend on catastrophic tail risk, whereas standard evaluation reports only central tendency. Because fine-tuning induces bimodal mode collapse, high average accuracy frequently conceals severe failure rates. Auditing $2\,124$ fine-tuned models across eleven language models, we show that shipping models based on mean accuracy violates a $10\%$ deployment failure budget $48.0\%$ of the time. Common heuristics like $\bar b - 2\sigma$ prove equally dangerous: by underestimating variance on small samples, they become most permissive when evidence is weakest. We introduce recipe cards, an evaluation framework that directly certifies rerun tail failure rates using exact finite-sample binomial bounds. Recipe cards reduce held-out budget violations to $0.6\%$, enforce a mathematical sample floor ($N \ge 22$) that prevents certification on thin evidence, and explicitly refuse unverified claims under domain and task shifts.

Architecture attribute predictors reduce the cost of neural architecture search by estimating accuracy or latency from a small set of evaluated candidates. The key difficulty is that neural architectures are operation-labeled DAGs: edge direction defines computational flow, but existing predictors often treat direction as a local structural cue. We propose DTA-GT, a Direction- and Topology-Aware Graph Transformer based on encoder-wide directional consistency, which preserves directed computational semantics across node initialization, global spectral encoding, pairwise interaction, and feature update. DTA-GT realizes this principle with direction- and topology-aware node initialization, Magnetic Laplacian spectral encoding, direction-dual structural attention, and a Direction-Sensitive MoE-FFN. Across accuracy and latency prediction on NAS-Bench-101 and NAS-Bench-201, DTA-GT consistently outperforms representative sequence-, GNN-, Transformer-, and hybrid predictors under limited-budget settings. Ablations and Magnetic spectral controls show that the improvements come from coordinated direction-aware stages and edge-orientation-sensitive spectral information, while mechanistic diagnostics confirm that the modules exhibit the intended directional behavior. These results support encoder-wide directional consistency as an effective design principle for architecture DAG representation.

In simulation-based inference (SBI), neural posterior estimators trained on a bounded range of sample sizes can become inaccurate on larger datasets because their parameterizations do not specify how posterior location, uncertainty and shape must change as observations accumulate. We propose center-estimated NPE (CE-NPE), which combines a statistical center with an explicit contraction rate and learns a local center correction and residual densities in posterior-scale coordinates. The center is a statistical estimate, such as a closed-form statistic or a simulator-based moment match, which does not require an evaluable likelihood. Our analysis explains how contraction amplifies center error and decomposes Gaussian posterior error into shape, mean and covariance contributions. On six regular benchmarks, CE-NPE gives the best reference-posterior agreement of all tested amortized estimators at every out-of-range size, up to five times the largest training size, without retraining; it also fits better within the training range on most of them. Matched controls and nonregular examples characterize the scope of these gains and identify the conditions on which they depend: an appropriate contraction rate, an accurate center, and a stable residual representation.

Matching one distribution to many targets can require repeated evaluations of every target. CAL-$\alpha$-MODM aggregates a nonlinear predictor into one barycenter, calibrates its residual derivatives, and checks steps against the original $f$-divergence objective. Under smoothness assumptions, calibration controls parameter-step model error without requiring targets to be close to the model. Exact-oracle stationarity and local convergence guarantees are established. At common matching, canonical geometry cancels the leading quadratic iteration error, giving cubic convergence with second-order calibration and sufficiently accurate inner solves. High-precision experiments support this prediction. Sensitivity-based initialization also accelerates preference continuation along regular stationary branches. On 300-target Jensen–Shannon problems, CAL-p2 achieves median paired speedups of $3.15\times$ over L-BFGS and $1.10\times$ over second-order Direct-TR, across 23/25 and 24/25 objective-comparable stationary pairs. The savings arise when cheaper predictor queries offset calibration overhead.

Progressive depth growth builds large networks by starting with shallow models and inserting layers during training. Recent advances in large language models have renewed interest in this longstanding approach, demonstrating benefits for training efficiency and reasoning alongside changes in internal computations. However, how growing affects representation quality and computational patterns in self-supervised vision transformers remains unclear. In this work, we systematically investigate progressive depth growth in two popular self-supervised families, namely masked autoencoders (MAE) and DINOv3, and analyze how growing affects downstream performance, intermediate representations, and internal computations. Across global and dense tasks, we show that growing improves downstream performance while reducing training FLOPs by 13.9–24.8%, with ImageNet linear-probe accuracy gains of up to 3.36 percentage points. Growing also improves downstream performance at intermediate layers and reshapes internal computational patterns, introducing recurring alternations between global and localized attention in MAE and earlier local-to-global attention progression in self-distilled DINOv3. Together, these findings establish progressive depth growth as an effective and efficient pretraining strategy for improving downstream performance across global and dense visual tasks with less computation, and advance our understanding of how growing shapes visual computation.

Global reanalyses are too coarse to resolve the convective systems that produce most West African monsoon rainfall. In this work, we use U-Net based architectures to downscale daily ERA5 fields (0.25{\deg}) to IMERG V07 satellite precipitation (0.1{\deg}) over West Africa for the June to October seasons of 2000 to 2024, using geopotential, temperature, humidity and wind at three pressure levels together with elevation and land fraction. Adding ERA5 precipitation as input slightly improves performance for this particular challenging region. We find that all the deterministic networks investigated for this work are too dry and lose almost all observed variance below a few hundred kilometers, which motivated our use of a GAN-based alternative. We also investigated permutation importance for the U-Net that ranks 500 hPa specific humidity first, consistent with the sensitivity of deep convection to mid-tropospheric moisture.

Calibration has been extensively studied for classifiers, but remains largely unexplored for crowdsourcing aggregation methods, which are increasingly used for making probabilistic decisions. To the best of our knowledge, we provide the first theoretical and empirical study of calibration in label aggregation. To our great surprise, we find that widely used aggregation methods can be severely miscalibrated, and in opposite directions: Majority Vote is underconfident while model-based aggregators are overconfident. We prove that the Bayesian aggregator is perfectly calibrated under its data-generating model. We also show that exact calibration of weighted majority votes requires highly restrictive conditions, which helps explain why Majority Vote is generally miscalibrated in realistic crowds. Empirically, we study calibration across $39$ human crowdsourcing datasets and LLM-based crowds. In both cases, we observe systematic miscalibration for aggregation models such as Majority Vote, Dawid–Skene, GLAD, and CrowdFM. Finally, we propose a post-hoc calibration procedure that aggregates the aggregators: for any aggregation function $\phi$, we introduce a transformation that first combines its predictive distribution with those of complementary aggregators, before applying temperature scaling. This simple and interpretable transformation makes temperature scaling more effective than when applied directly to $\phi$. It improves calibration, especially in the high-redundancy regime where there are many annotations per item, while generally preserving or even improving accuracy.

Vision foundation models (VFMs) provide transferable visual representations, but their inference cost limits deployment in many settings. Fine-tuning specializes a VFM without shrinking it; distillation shrinks it, but through expensive teacher-guided training. Structured pruning offers a more direct route by extracting a compact subnetwork from the VFM. However, as compression gets aggressive, existing methods fail because they treat subnetwork selection and recovery as separate axes. Across layers, capacity must follow task sensitivity rather than spread uniformly. Within a layer, a removal cannot be judged by importance alone; its impact also depends on how well the lost contribution can be recovered from what remains. Selection and recovery must therefore be decided jointly. We introduce Carve VFM, which resolves this interdependence. Given a VFM, a calibration set, and a parameter budget, it (i) searches for task-specific capacity allocations across layers; (ii) guides the search via an adaptive recovery-aware fitness criterion; and (iii) reconstructs removed behavior through iterative closed-form low-rank corrections merged into the surviving weights. The result is a dense encoder produced in minutes on a single GPU without backbone retraining. Across twelve benchmarks and four models from DINOv3 and MetaCLIP-2, Carve VFM achieves the highest reported mean accuracy in seven of eight backbone-budget configurations among the evaluated dense structured-pruning methods. Further fine-tuning surpasses the strongest evaluated edge model (88.1% vs. 85.8%), and carved models cut compute and memory while delivering 1.3-2.0x higher GPU throughput than the uncompressed one. These results show that the compact model you need is inside the VFM; it only has to be carved out.

Symbolic regression (SR) aims to discover interpretable mathematical expressions from data and plays a key role in scientific discovery. However, existing methods face a common bottleneck: the enormous search space and severe combinatorial explosion make it difficult to achieve a favorable trade-off between accuracy and algorithmic efficiency. We argue that, in addition to designing increasingly complex solver-specific SR strategies, an equally promising direction is to equip them with a shareable auxiliary strategy that can provide reliable search guidance. To this end, we propose DASS, a solver-adaptable Dynamic Auxiliary Search Strategy that improves SR by constructing and refining a high-potential auxiliary search subspace. DASS represents the auxiliary search space as a functional subspace spanned by a set of high-potential basis function terms. It estimates term utility through multi-environment evaluation, filters unstable guidance, and updates term potential scores with a Gibbs posterior, thereby calibrating the subspace to provide more reliable guidance for downstream solvers. The high-quality solutions produced by the solver are used to update the set of basis terms. This forms a virtuous closed-loop optimization process: the subspace guides the downstream solver, and the solver feedback reshapes the subspace for subsequent exploration. DASS can be easily integrated with various types of SR solvers. Extensive experiments on LLM-SRBench show that DASS significantly improves the accuracy of original solvers with a manageable increase in runtime.

This study addresses the prediction of audio popularity on an online sharing platform by leveraging both metadata and acoustic feature representations. Two distinct categories of audio content are considered: music and sound effects. Popularity is formulated as a supervised classification task based on two target variables: the number of downloads and the average user rating. We conduct a comparative evaluation of multiple machine learning models, including Random Forest, XGBoost, and a Neural Network model. Experiments are performed across different feature sets, including metadata, platform-provided descriptors, and eGeMAPS acoustic features. Results indicate that metadata consistently provide the most informative representation for predicting popularity. Nevertheless, acoustic features alone achieve performance significantly above chance level, demonstrating that intrinsic properties of the audio signal contribute to popularity prediction.