Papers

The SWOT (Surface Water Ocean Topography) mission is currently providing unpreceded high-resolution measurements of Sea Surface Height (SSH), revealing ocean features at finer scales. Nevertheless, the two-dimensional observations of KaRIn altimeter of SWOT suffer from instrumental errors. This noise degradation is altering the high frequencies of SWOT signal, and the small to sub-mesoscale dynamics of interest for some oceanographers. For this reason, Tréboutte et al. (2023) have developed a convolutional neural network (CNN) based on U-Net architecture to separate the noise from the physical signals contained in the SSH. Their approach has demonstrated great potential on simulated SWOT measurements, and Dibarboure et al. (2024) report a positive influence on actual flight data from SWOT. However, degraded denoising performance, and occasional negative side-effects have been observed in atypical conditions (e.g. very high surface waves, internal tides solitons). In this study, we illustrate some of these limitations and we present an improved approach of the CNN-based denoising: we modified the training procedure to obtain a more robust version of the algorithm, to avoid biases and artifacts in the denoised SSH. This paper also presents a more complete validation process with a robust and standardized evaluation benchmark: these metrics could be of interest to assess other SWOT filtering and denoising algorithms.

Is the same data equally useful to every learner? The question is whether the value of a training sample belongs to the sample or arises in its interaction with the learner. We test it with a selection experiment. Under a per-class budget, two kinds of samples compete for the slots: easy samples, which a short training run learns first, and coverage samples, which spread over the class in feature space. Easy samples win at small budgets and coverage samples at large ones, and the budget at which the winner flips, the crossover m0, can be measured. We ask whether m0 is set by the data or by the learner. To separate the two, we freeze the selected subsets and change only the learner that trains on them. On low-resolution ImageNet-100, doubling the width of a ResNet-18 moves the crossover from 57 to 85 samples per class on the same subsets. Enlarging the input grid from 32 to 64 pixels at fixed image information moves it from 85 to 57, and a single stride change reproduces or removes this shift. The grid shift shrinks as width grows. Under the native ImageNet-1k protocol, a 4x range of width moves the boundary by at most a few samples per class. On a Vision Transformer, coverage wins at every measured budget, even with easy subsets that work on the convolutional learner. The budget at which one kind of sample stops being the more useful kind therefore belongs to the learner, and the preferred regime of a selection strategy has to be read against the learner that will use the data.

The quality of a detector design is ultimately determined by the quality of the inference it enables, that is, by the accuracy with which the quantities of interest are reconstructed from the raw detector response. For complex detectors, the inference is performed by machine learning models, and the relation between the design and the attainable inference performance is, in general, non-trivial. In this work, we consider the optimization of the inference performance with respect to the detector design. The conventional approach prescribes retraining the inference model at every candidate design, thus, treating the evaluations as independent tasks and discarding the shared structure of the optimal inference algorithms at different designs. We propose the meta-learned objective estimate (MLOE): instead of solving the inference problem anew at every candidate design, a single meta-inference model, conditioned on the design and trained continually along the optimization path, is shared across all of them. We test MLOE on three families of optimization problems, the last of which comprises two design spaces of the Spectrometer Straw Tracker of the Search for Hidden Particles (SHiP) experiment; under matched budgets of simulation calls, the meta-inference model evaluates a candidate design using fewer simulation calls than the baseline strategies and holds the better rank over the convergence curve in all examined cases.

The thermal behavior of wood is a critical factor in advanced material assembly. However, pixel-level thermal analysis remains fundamentally constrained by the low resolution and noise inherent to infrared thermography. To address this, we introduce an end-to-end computational framework that synthesizes high-resolution thermal responses directly from wood RGB images. We first establish a core physical linkage: because spatial color variation in natural wood is driven by cellular anatomy, optical intensity serves as a reliable geometric proxy for the localized solid volume fraction. By leveraging this theoretical insight, we develop an automated finite-element-method data engine that maps pixel-level optical intensity to a 3D thermodynamic voxel grid, generating high-fidelity synthetic thermal responses. We find that 1) when the thermal conductivity along the thickness direction is uniform or linear, wood RGB images and their corresponding thermal responses exhibit extreme morphological similarities, and the lateral thermal diffusion acts as a low-pass filter that smooths out high-frequency details; 2) when the thermal conductivity along the thickness direction is random, such morphological similarities are destroyed, and wood's 3D structure dominantly governs its thermal response. We further utilize these synthetic thermal responses to supervise a neural surrogate model built upon the DINOv3 foundation model. Our results demonstrate that the neural surrogate model successfully internalizes the governing thermodynamic laws, thereby bypassing computationally expensive simulations and enabling high-resolution thermal inference. This methodology effectively bridges the semantic and thermodynamic domains, unlocking systematic, pixel-level analysis of fine-grained wood thermal responses. Project: https://zekifayes.github.io/seeheat

In this paper, we propose GERIS, a game-theoretic framework for instance selection in the data augmentation phase of license plate recognition systems. During augmentation, synthetic license plate images are generated and transformed using stochastic noise to simulate real-world conditions. However, certain noise configurations lead to highly distorted, unreadable images that degrade model performance by introducing instance-dependent label noise. GERIS formulates a non-cooperative game in which each noise vector competes for inclusion in the training set based on its similarity to labeled data and its contribution to model reliability. By identifying and pruning low-quality instances, GERIS improves the overall quality of the augmented dataset. Unlike traditional black-box learning methods, GERIS offers a transparent, theoretically grounded mechanism for data filtering. Experimental results demonstrate that GERIS outperforms existing instance selection methods in terms of classification accuracy and robustness.

With the rapid progress of AI, the number of AI-generated images has increased significantly in recent years. However, the increasing variety of image generation models makes detection more difficult. In this work, we use Fast Fourier Transform (FFT) representation with EfficientNet-B0 for AI-generated image detection. EfficientNet-B0 provides a lightweight architecture that can be useful for resource-limited applications. Most frequency-domain detectors use a standard encoder to extract features from the FFT spectrum and directly pass them to a classifier. We explored a different approach by investigating ECA, U-Net, and Attention U-Net as alternatives to this direct encoder-to-classifier approach. ECA applies channel attention, while U-Net and Attention U-Net use decoder-based architectures to recover and refine spatial information in the extracted frequency features. We used a balanced subset of the MS COCOAI dataset that includes AI-generated images from five different models. Three runs were carried out for each experiment, and the average values were recorded. Experimental results indicate that EfficientNet-B0 obtained an accuracy of 84.64%, which is 4.50 percentage points higher than the ResNet-50 baseline reported in the dataset paper. EfficientNet-B0 with U-Net provided a small improvement, achieving an accuracy of 84.85%, while ECA did not increase the overall performance. EfficientNet-B0 with Attention U-Net achieved the best overall performance, with an accuracy of 85.51% and an ROC-AUC of 92.99%. This represents an improvement of 0.87 percentage points in accuracy compared to the EfficientNet-B0 baseline and 5.37 percentage points over the ResNet-50 baseline reported in the dataset paper.

We analyze a simple stochastic inertial Krasnosel'skii–Mann (iKM) method for finding a fixed point of a nonexpansive operator in a real Hilbert space. Our method is obtained simply by adding two inertial extrapolations to stochastic KM [Bravo and Cominetti, 2024], and it retains one call to a possibly biased stochastic oracle per update and achieves sharp rates in both the stochastic and deterministic regimes. Specifically, with our proposed parameter schedule, we prove the following last-iterate fixed-point residual bound: \[ {O}\!\left(\frac{1}{K} +\frac{\sigma\log K}{\sqrt K} +\frac{B_K\log K}{K}\right), \] where $K$ is the horizon, $\sigma$ is the noise level and $B_K$ is the accumulated root-mean-square bias. When $B_K=O(\sqrt K)$, this yields $\widetilde O(\epsilon^{-2})$ sample complexity that matches, up to a logarithmic factor, the stochastic-oracle lower bound given under the unbiased subclass of our model [Foster et al., 2019, Theorem 2]. It also improves the best-known $O(\epsilon^{-4})$ random-iterate guarantee for stochastic KM [Bravo and Cominetti, 2024, Corollary 5.4]. To our knowledge, this is the first single-loop method for general nonexpansive fixed-point problems to attain this near-optimal sample complexity without variance reduction or batching. When the oracle is exact, the same method attains the worst-case-optimal $O(K^{-1})$ last-iterate residual rate [Park and Ryu, 2022, Theorem 4.6], improving the $O(K^{-1/2})$ rate of classical KM [Cominetti et al., 2014; Bravo and Cominetti, 2018].

The goal of this note is to give a detailed proof, to the best of our understanding, of the recent presentation by Harrison and Leeman (arXiv:2609.17650v01 and arXiv:2609.17650v02) of the proof by Astra on the lower bound for differentially private continual counting. We believe a more natural and easy proof is possible and hope that this note will help in that effort. Prior to the initial preprint by Harrison and Leeman (arXiv:2609.17650v01), Bairaktari and Larsen (arXiv:2607.00876) gave an elegant proof to show a lower bound of $\Omega(\log^{3/2}(n))$ for both pure and approximate-DP continual counting, and in personal communication had informed us that they have a proof of optimal $\Omega(\log^{2}(n))$ for pure-differential private continual counting as well. They have subsequently published their $\Omega(\log^{2}(n))$ bound, which is now a joint work of Bairaktari, Dahl, and Larsen (arXiv:2607.00876v3). Their new result is an elegant extension of their technique for approximate-differential privacy. Although the two proofs are technically different, the Astra argument uses related tree geometry introduced in Bairaktari and Larsen.

Modern conformal forecasting systems often maintain several adaptive pipelines that differ in base forecasters, conformity scores, calibration windows, and update rules. Comparing them is difficult because coverage is a hard constraint, whereas efficiency should be optimized only among feasible pipelines. We formulate this problem as sequential inference for a stochastic constrained argmin. At each time, the target is the set of minimum-cost pipelines satisfying multiple prefix-average conditional miscoverage constraints. We introduce Coverage-Constrained Sequential Model Confidence Sets (CC-SMCS), which separate certifiably feasible, possibly feasible, and possibly constrained-optimal pipelines. Using simultaneous martingale confidence sequences, CC-SMCS projects a rectangular confidence region onto the constrained argmin and admits an exact closed-form rule. With probability at least $1-\delta$, it contains every constrained-optimal pipeline simultaneously over all times. This finite-sample guarantee requires no stationarity or mixing assumptions and remains valid under data-dependent stopping. We also establish an impossibility result for safe certification at the coverage boundary and extend the construction to delayed multi-horizon feedback and outcome-dependent efficiency objectives.

Before acting upon an unobservable physical system, an autonomous agent must determine which latent distinctions govern downstream tasks, how many active interventions are necessary to certify them, and when to abstain to prevent catastrophic errors. Classical observability treats state reconstruction as an unconditioned binary predicate, failing when passive observations cannot break latent degeneracies without perturbation, full microscopic inversion is prohibitively costly, and distinguishing task-irrelevant degrees of freedom wastes interaction budgets. We formalize task-conditioned active observability complexity: the minimum worst-case expected interaction cost required to identify task-relevant states under certified error and safe abstention guarantees. We prove that task-predictive equivalence induces the unique minimal sufficient quotient $\mathcal{H}/\!\sim_\tau$, leaving active observability complexity strictly invariant while eliminating superfluous distinctions. In deterministic regimes, this complexity is characterized by an optimal adaptive distinguishing tree and Bellman recursion; in noisy regimes, it obeys a stopped-transcript relative-entropy lower bound and adaptive martingale certificates that compose without independence assumptions. We instantiate a prospective certified observer with staged recovery: a nominal verifier defers candidate compilation, triggering active probing only upon evidence, while a history-measurable score shell prunes hypotheses without sacrificing risk bounds. Stress audits across high-dimensional physical systems and thousands of operational trials demonstrate certified state recovery with zero false acceptances and substantial reductions in sensor reads and model steps.

Action-chunked visuomotor policies predict overlapping trajectories, so every executed action is covered by several predictions. Temporal ensembling smooths execution by combining these predictions with an exponentially weighted mean. One corrupted prediction can move the aggregate without bound: its breakdown point is 0. We use adversarial corruption to stress this deployed aggregator and to compare two kinds of guarantee. A metric guarantee bounds the response to a perturbation of a given size. A combinatorial guarantee instead bounds the damage when at most q of the M candidates covering a timestep are corrupted, whatever their size. Encoder adversarial fine-tuning recovers 44% of the loss under the published patch attack, but only 7.3% after the attacker's step size is increased. By contrast, the coordinate-wise median of the same candidate set keeps its recovered fraction flat as attack optimisation increases. Median temporal ensembling costs one line and requires no retraining. Across 25 (configuration, corruption-level) combinations it is never worse than the mean and is significantly better in 15. It also transfers to a second policy class, and it recovers performance under a failure with no attacker in the loop at all: camera frames that arrive blank. Its effect on clean data is configuration-dependent, from -0.04 to +0.07. We also give the boundary: corruption that shifts every covering prediction by the same amount is invisible to this whole family of statistics, and no equivariant aggregator can remove it.

Capability and efficiency are two key dimensions of reasoning in large language models (LLMs). Capability refers to the ability to solve a given problem correctly, whereas efficiency refers to the ability to do so with limited resources. When LLMs use Chain-of-Thought (CoT) reasoning to solve problems of controlled hardness, both the number of problems solved correctly and the number of tokens required to reach a correct answer depend on problem hardness and model size. However, how these factors jointly shape capability and efficiency remains poorly understood. Here, we use hierarchical Bayesian models to evaluate the capability and efficiency of LLMs from the DeepSeek-R1-Distill model family across four classes of arithmetic and algorithmic reasoning problems. At a fixed model size, the probability of correctly solving an instance decays approximately exponentially with instance size, our proxy for problem hardness. The decay scale grows sublinearly with model size, indicating that larger models are more capable, but that capability gains diminish with scale. Output length grows as a power law with instance size, which serves as a proxy for difficulty. However, the parameters of this power law do not vary systematically with model size, suggesting that larger models do not become more efficient. Together, these findings reveal potential limitations of naive scaling as a strategy for developing more capable AI systems: capability improves with diminishing returns, while efficiency shows little to no improvement.

To make claims about representations that generalize beyond a particular trained model, we need to specify when two representations should count as equivalent. The Linear Representation Hypothesis is often discussed without making this equivalence explicit. Different notions of equivalence preserve different structures, so metrics, probes, and interventions that appear to study the same representation may in fact correspond to different hypotheses. We therefore argue that the Linear Representation Hypothesis is not one hypothesis but a family of claims distinguished by representation equivalence. We formalize this idea using group actions, specifying the representation object, the procedure that produces it, and the property ultimately asserted, while accounting for equivalences imposed by the model architecture. This framework clarifies how assumptions can change across metrics, reading points, and analysis stages, and we use it to audit common representation quantities and recent interpretability analyses.

Large reasoning models can produce correct yet unnecessarily long reasoning traces. Existing methods improve reasoning efficiency with trajectory-level objectives or local token- and step-level signals, but rarely model inter-step semantic dependencies. This limits their ability to distinguish redundant steps from those that support later deductions, making it harder to shorten reasoning without sacrificing accuracy. We introduce RECAP (REdundancy-aware Credit Assignment via Propagation), which addresses this limitation by assigning credit where it is due based on both a step's downstream role in the reasoning structure and its contribution to solving the problem correctly. We define structural responsibility to capture the step's downstream role by measuring how strongly later reasoning depends on it, using credit propagated backward from the final-answer node through an outcome-independent, LLM-annotated semantic dependency graph. However, a step can have high structural responsibility yet steer the reasoning away from the correct solution. RECAP therefore introduces step efficacy to measure answer-directed progress through changes in gold-answer log-likelihood as each step is added. Together, these signals reshape rollout-level GRPO advantages into step-specific updates. RECAP requires neither a separately trained process reward model nor preconstructed concise trajectories. Across two 7B models and four mathematical reasoning benchmarks, RECAP improves the accuracy-efficiency trade-off. On Qwen2.5-Math-7B, it improves pass@1 by 2.0-3.7 percentage points while reducing reasoning tokens by 8%-31% relative to GRPO across all four benchmarks. Analysis suggests these savings reflect fewer reasoning operations and less dead-end reasoning, rather than more compact expression.

With recent advancements in large language models (LLMs) and LLM-based agents, these agents are becoming increasingly autonomous and gaining broader access to act on users' behalf on the internet. However, the vulnerability of automated agents deployed on social media platforms (e.g., for managing a user's personal account) remains underexplored. Existing studies on agent poisoning typically assume that the adversary can expose poisoned content to the agent. Although such an attack is direct and effective, it is more easily detected and mitigated. In the context of social media platforms, this leaves open whether the recommendation system itself would surface such content to the agent in a more subtle manner. Through theoretical analysis, we show that the like-score mechanism used in OASIS can be exploited, and we characterize the conditions under which a multi-stage chain of poisoned posts can steer the agent's feed. Based on these insights, we further develop an algorithm that crafts realistic poisoned posts. Experiments support our theoretical findings and demonstrate the effectiveness of the proposed algorithm. Notably, by exploiting the like-score feedback loop, the attack causes the recommendation system to select poisoned posts even when their user-post similarity falls below the retrieval threshold.

We consider the problem of learning risk scores to prioritize individuals for scarce resources or interventions, from historical observational data affected by unobserved confounding. Decisions about who receives scarce resources are often guided by risk scores based on recorded characteristics, such as responses to a survey. These risk scores are increasingly being learned directly from observational data: historical records of individuals' characteristics, allocation decisions, and outcomes. Standard methods such as inverse propensity weighting (IPW), which corrects for the bias introduced by the historical allocation policy, can be used to learn accurate risk scores if the historical decision process is fully explained by the recorded characteristics. In practice, however, historical decisions often depend on unrecorded information, causing learned risk scores to systematically under-prioritize exactly the individuals whose unrecorded circumstances drove past prioritization. We propose a method for learning risk scores that are robust to this kind of unobserved confounding, building on IPW. Since propensity weights cannot be reliably estimated under unobserved confounding, we instead treat them as belonging to an uncertainty set determined by the observable data and domain-informed estimates of the degree of confounding, combining sensitivity analysis from causal inference with Wasserstein distributionally robust optimization. The resulting robust risk score learning problem admits a sample-based approximation that we reformulate as an exponential cone program compatible with off-the-shelf solvers. We demonstrate the effectiveness of our approach on semi-synthetic data derived from datasets in the UCI Machine Learning Repository. Our method improves calibration by up to 29.2% over traditional benchmarks and up to 11.1% over the state of the art, without compromising other metrics.

Heterogeneity across Collaborative Perception (CP) agents is a major challenge for emerging CP frameworks due to domain gaps from differing sensors, architectures, and training data. Prior works mitigate this challenge by aligning features in a unified space via model retraining or per-agent-type interpreters. These strategies (a) require access to neighbor configurations, (b) do not fully address real-time CP deployment, and (c) generalize poorly to unseen agents joining at run time. To overcome these challenges, we present PEARL, a Prompt-Embedding framework for Anonymous and Real-time Lightweight heterogeneous CP. PEARL supports multiple CP interpreters and selects one for a new-joining agent in real time using two lightweight, multi-scale interpreters trained in parallel: a sparse-detection (LWSD) interpreter that aligns salient regions for cooperative detection, and a dense, domain-invariant (LWDDI) interpreter that produces agent-invariant features for fast interpreter selection. Both interpreters use low-rank visual prompts to reduce computation, storage, and model complexity. Extensive experiments on simulated (OPV2V, V2XSet) and real (DAIR-V2X) datasets show that PEARL generalizes across simulated and real-world cooperative driving scenarios. Its real-time model-selection strategy yields an 8.2% Average Precision (AP) gain over a random-selection baseline while running in 1.67 ms on average. Although primarily designed for real-time CP, PEARL also outperforms state-of-the-art heterogeneous CP frameworks under traditional offline training by 5.6% AP on average while reducing communication cost by up to 34.7 times. Equally important, PEARL does not require sharing agents' configurations or model settings, thereby protecting information that may be proprietary or private. These results establish PEARL as a scalable and practical framework for heterogeneous collaborative perception.

Vision-language models (VLMs) are deployed as zero-shot judges of image aesthetics, and panels of several models are recommended, on thin evidence, as the way to make such judges reliable. On two human-rated datasets, EVA and PARA, we find that a panel of holistic judges never significantly beats its best member, whether the verdicts are averaged or fused by a learned combiner. What a panel is worth depends on what it is fed. We therefore have each model score each image on the five dimensions of a frozen, human-written rubric and fuse those scores, alongside each model's verdict, across model families with an out-of-fold combiner. The dimension scores measure what their labels claim: with the overall human score partialled out, a dimension prompt carries more attribute-specific information than the holistic prompt in 28 of 30 model-attribute cells. Fused, they beat the best single VLM in all ten three-family panels on EVA (against that best single model, +0.07 Spearman rho for the strongest trio and +0.10 for the pre-declared one, and +0.06 and +0.07 when averaged over twenty fold partitions; against the panel mean, the primary test gives +0.118 on its EVA design set), and on PARA they reach parity under Spearman rho and a small, non-significant loss under Kendall tau-b, where one model already captures 85% of the human noise ceiling. It is not a feature-count artefact: giving the same combiner an equal number of pure holistic columns, split from the same repetitions, does not reproduce it. The gain costs a few hundred labels, which do not transfer between datasets, and 4.8x the API calls on EVA; we report it with paired bootstraps and Kendall tau-b, alongside a failed pre-registration and the configurations that lost.

Robotic embodiment encompasses the sensing, kinematics, dynamics, geometry, actuation, and control through which an agent physically interacts with the world. These properties vary across robots and change over time. We argue that general embodied intelligence requires learning that accumulates across these differences. Prevailing methods that engineer correspondences to bridge embodiment differences offer immediate practical gains, but their assumptions limit the scope of transfer in the long run. Instead, a more general approach should discover representations that support transfer to a larger range of embodiments as experience grows. We propose embodiment diversity as a promising axis of scaling, and identify broad learned priors as a complementary ingredient. We call for evaluations that better characterize embodiment gaps and transfer performance. More broadly, cross-embodiment learning connects the practical challenge of learning from heterogeneous robot experience with a broader scientific pursuit inspired by nature - physical intelligence that adapts and co-evolves with its embodiments to gain agency over its behavior and physical forms.

Many node-classification GNNs apply a linear classifier to a nonnegative mixture of local messages. An error can reflect either poor mixture weights or a reachable logit set poorly positioned for the classifier. We separate these causes with an exact-mass linear program and two learned post-hoc repairs. Every reweighted prediction has an equivalent centered logit translation, but only translations in a message-induced displacement set are realizable by reweighting. Across eight datasets, eight GNN backbones, and ten splits, mean accuracy rises from 62.6% for the frozen models to 63.8% with reweighting and 65.3% with set-conditioned translation. A parameter-matched node-only translator reaches 64.6%, showing that translation explains most of the gain while the message set supplies a smaller additional benefit. Although oracle reweighting can correct many errors, label-free reweighting captures little of this potential: local evidence is often present but hard to select, and relaxing the evidence constraint is more effective than learning within it.

As serving capacity demand surpasses that of training, serving efficiency becomes increasingly important. Prefill-decode (P/D) disaggregation improves serving efficiency through specialization and isolation of the two phases. These benefits rest on a static partitioning. Phase demand, however, is not static. We observe that in a large LLM fleet the ratio of uncached input to output tokens has peak-to-mean ratios up to 4.7x at minute timescales, and that in a public agentic trace the hourly ratio spans a median 24.5x within a single day, while reassigning a replica takes tens of minutes. Agentic traffic sharpens the mismatch. Sizing each pool at its ninety-fifth percentile leaves up to 17% of cluster capacity unused; sizing below it converts the same imbalance into queueing and unrealized throughput. We present Crossflow, which makes this boundary elastic without changing node roles. Each decode node publishes a short-lived, revocable lease that bounds local-prefill compute, KV capacity, transfer work, and projected output. Across public and internal traces, Crossflow improves token throughput by 16.2-17.4% on geometric mean over static P/D, and by up to 43.4% at high load, while reducing mean TTFT at every evaluated point.

This study introduces a dual-matrix computational architecture to mathematically quantify the morphological and theological divergence of 196 Hindu and Vajrayana Buddhist esoteric deities. Physical morphology is evaluated via a discrete Gower distance matrix enhanced by a novel "Cardinality Weighting" algorithm, while theological function is mapped via dense vector embeddings generated from Large Language Model (LLM) semantic expansions, explicitly utilized as a synthetic proxy to mitigate circular reasoning. The multi-modal topological projections provide algorithmic validation of "iconographic camouflage", demonstrating how distinct visual forms structurally obscure shared cross-tradition functions. Furthermore, I computationally model the "Atin Effect" - serving simultaneously as a psychological observation of sequential cognitive bias and a machine learning benchmark - demonstrating how high-cardinality esoteric anchors (e.g., a veena or a severed head) override systemic theological disparities to mathematically cluster orthodox and Tantric entities. Cross-tradition spatial analysis establishes that the highest esoteric manifestations, such as the Hindu Chinnamasta and the Buddhist Chinnamunda, share a near-identical mathematical coordinate across both visual ($D_G = 0.288$) and semantic ($D_C = 0.068$) boundaries, indicating a 1:1 esoteric transfer. By open-sourcing this architecture, I provide a scalable, unsupervised machine learning tool for Digital Humanities scholars and comparative theologians to rigorously map latent structural continuities across qualitative cultural corpora.

Graph neural networks dominate recent work on microservice root-cause analysis, yet recent results question whether the graph contributes. Those results compare whole pipelines, so when a flat model wins one cannot tell whether structure is useless or redundant. We run the comparison they imply on RCAEval: three learned arms with identical features, optimiser, validation split, early-stopping rule and scoring head, in which a single term separates the graph arms. Across two RCAEval benchmarks, two topology sources and four regimes we find no reliable graph-specific effect: in-distribution the graph model leads the conventional flat model by 0.003 Avg@5 (p = 0.844, n = 6 disjoint folds). Auditing the pipeline surfaced two benchmark properties that condition any result on it. RCAEval injects faults into only five services per system while exposing 12 to 70 in telemetry, and the headline metric is Avg@5: a ranker reading no telemetry at all places the true culprit in the top five on 99.7 percent of held-out incidents, scoring Avg@5 0.488. That prior, not the uniform-random 0.137, is the honest in-distribution floor, and it collapses to 0.192 across systems. The second property is a non-uniform column schema that silently zeroes telemetry for most RE1 cases. We reproduce a published baseline, BARO, RCAEval's own reference implementation; on the one system with a clean schema it reaches similar aggregate accuracy to our heuristic, within 0.004, under a different scoring rule. The audit motivated a new model. PSC-GRCA separates a candidate score into a system prior, telemetry evidence and a centred graph residual, and reaches mean Avg@5 0.915 against 0.864 for the flat baseline, while its ablations locate most of the gain in the prior term rather than the graph. We close with a twelve-item checklist for graph-versus-flat ablation studies, distilled from sixty-two recorded defects.

Large language model (LLM)-based agents use Electronic Design Automation (EDA) tools to generate and revise register-transfer-level (RTL) designs under synthesis and verification feedback. Recent methods learn from this feedback by distilling reusable skills from execution traces or by training on rewards derived from EDA-tools. Both methods are typically evaluated on the tasks that produced the experience. Repeated access to benchmark feedback on the same task can reward task-specific revision rather than creating reusable knowledge that transfers. We introduce ChipMEM, a verification-grounded memory layer for EDA agents. It combines cross-task procedural memory with within-trajectory statistical guidance. Its procedural component distills and stores a skill only after it passes synthesis, simulation, or formal checks, rather than relying on model self-assessments. A Bayesian component maintains hierarchical Beta estimates over tool-call outcomes and ranks recovery strategies that succeeded under comparable errors. A common adapter applies the same memory interface to RTL optimization and testbench-generation agents while preserving each domain's tools and acceptance criteria. We measure performance on training tasks and evaluate whether learned skills transfer to unseen tasks. On RTLRewriter-Bench, under matched model and tool settings, ChipMEM produces equivalence-passing outputs on 39/54 scored designs versus 35/54 without memory; on the 49-design short suite, mean area improvement is 8.69% versus 5.66%. On held-out CVDP tasks, ChipMEM with a frozen procedural library achieves 20/20 accepted outcomes versus 18/20 without memory in a single evaluation per setting.

Large language models have enabled rapid deployment of language classification and analysis, allowing users to describe complex constructs in natural language, such as pedagogical moves or expressions of learning and engagement, and label them in conversational datasets such as classroom or tutoring transcripts. However, because LLM reasoning is opaque, we have no verifiable, mechanistic insight into why a model chose a label for an utterance, and output differences over time or across models challenge replication. We introduce the EduBehaviors framework to generate interpretable, reliable, and scalable construct annotations in educational dialogue data that retain human agency while leveraging the strengths of generative AI, encoders, and machine learning models. We evaluate the framework on the TalkMoves dataset, predicting the Teacher TalkMoves labels. Our best configuration results in a macro-F1 of 0.764 and 0.756 Cohen's kappa, demonstrating improvements over direct prompting approaches and showing competitiveness with fine-tuned classifiers while gaining the benefit of greater transparency and comparability. To facilitate adoption and operationalization of the EduBehaviors framework in educational dialogue data by other researchers, we also release the EduBehaviors Toolkit with two complementary open-source tools based on the TalkMoves application.

Multi-GPU servers have become the standard building block of modern data centers, providing aggregated capacity through high-bandwidth interconnects. At the same time, workloads such as LLM inference exhibit highly dynamic memory demands, which can cause one GPU to exhaust its local memory while others remain underutilized. This mismatch motivates a model of elastic resource sharing across GPUs. We present EMA, a memory sharing system that allows GPUs within a server to borrow and reclaim memory from each other, forming an elastic pool of capacity. EMA ensures performance transparency for both borrowers and lenders. For borrowers, prefetching hides remote access costs so that applications experience remote and local memory as indistinguishable in performance. For lenders, borrowed resources remain reclaimable on demand, guaranteeing that performance never falls below that of static partitioning. While our design focuses on memory, the same principle naturally extends to other GPU resources. Our evaluation shows that EMA improves individual user throughput by up to 52%, achieves 96% of the throughput of a system provisioned with 2X capacity, and maintains latency similar to the static local baseline.

Polymer property prediction lacks open, standardized benchmarks that enable rigorous comparison of machine-learning methods, with existing resources covering only a narrow fraction of polymer architectures, such as homopolymers. We introduce Polymer Benchmark 2026 (PolyBench26), an open dataset comprising nearly 250,000 polymer-property datapoints across eight physical properties, including data from experimental measurements, density functional theory, and molecular dynamics. The benchmark supports four evaluation tasks across homopolymers and alternating, random, and block copolymers: in-distribution property prediction, dataset-size scaling, repeat-unit complexity, and transfer to held-out polymer architectures. We compare language model, graph-based, and descriptor-based approaches and find graph-based models provide the lowest errors in property prediction, retain their advantage across the evaluated training-set sizes, and remain robust to increasing repeat-unit complexity. PolyBench26 provides a reproducible foundation for developing models for the increasingly complex polymer design space. The PolyBench26 benchmark is available open-source at https://github.com/rlearsch/PolymerBenchmark2026.

Group Relative Policy Optimization (GRPO) and related policy-gradient methods for training language model agents collapse an entire multi-turn rollout into a single scalar trajectory reward before it enters the policy update. When the task composes distinct skills, especially under sparse and delayed environmental feedback, this collapsing is lossy: the optimizer must implicitly infer which competency drove the outcome and how that should change behavior. We argue the right primitive is not a better scalar but a decomposition: trajectory reward should be split along subtasks before it enters the policy update. We introduce Reinforcement Learning with Decomposed Subtasks (RLDS), whose core is Subtask-Decomposed Advantage Estimation (SDAE): a replacement for the scalar GRPO advantage that splits trajectory reward into per-subtask shares on a fixed taxonomy, computes a group-relative advantage per subtask, and distributes per-token credit by weighting each subtask's advantage by its importance, concentrating it around the step where a reflection marks that subtask's execution as consequential. We evaluate on four agentic benchmarks: FrozenLake (sparse grid navigation), HotpotQA (multi-hop QA, one retrieval tool), ScienceWorld (long-horizon embodied science), and DeepResearch (long-form research, four tools, composite rubric reward). Heterogeneity diagnostics emitted during training show where decomposition pays off - gains scale with subtask heterogeneity, largest on the high-heterogeneity tasks ScienceWorld (+11.5 points, paired-bootstrap 95% CI [+9.8, +13.3]) and FrozenLake (+9.8 points, [+7.0, +12.8]), and within noise on HotpotQA and DeepResearch, where the diagnostics predicted little to recover. ScienceWorld is also more compute-efficient under RLDS than scalar GRPO (-10.9% wall-clock per step), as long rollouts amortize the fixed reflect-and-grade overhead.

We study prediction in new environments when the training data contains rare, large distribution shifts. Anchor regression penalizes the average of the squared mean residual across environments. It protects against shifts in an ellipsoid determined by the second moment of the training shifts. Covering rare shifts may therefore require a large penalty $\gamma \gg 1$, expanding the ellipsoid in every direction and reducing accuracy on common environments. We propose CVaR anchor regression, which replaces the average of the squared mean residuals with a tail average. We prove an exact worst-case risk guarantee under a linear structural model that allows for heteroscedastic noise. CVaR anchor regression protects against shifts in a union of ellipsoids determined by second moments of the shifts in the worst $\alpha$-fraction of environments. For $\alpha$ small enough, this is the scaled convex hull of the training shifts and their negatives. It therefore covers all training shifts when $\gamma \geq 1$. Unlike CVaR or GroupDRO applied directly to prediction risks, CVaR anchor regression does not give environments more weight solely because their noise levels are high. Examples show how the method can improve protection against rare shifts while retaining accuracy on common environments. We illustrate the method on the New York City taxi dataset.

Reward fine-tuning aims to update a pre-trained flow-based generative model to improve the downstream reward of its generated samples. Existing methods typically formulate this problem as sampling from a reward-tilted distribution, the solution to a KL-regularized reward-maximization problem. Here, we introduce an optimal transport regularizer built directly from the pre-trained drift. Unlike KL reward tilting, the resulting objective transports individual samples toward higher reward rather than reweighting the base distribution. We show that the resulting problem is equivalent to a deterministic optimal control problem on the flow. Given a pre-trained flow map, this equivalence yields a simulation-free reinforcement learning algorithm for fine-tuning generative flows. We call the resulting framework Wasserstein-Tilted Flow Maps (WTF), the first end-to-end fine-tuning recipe native to flow maps. The output is a fine-tuned flow map that retains strong reward-aligned performance at few-step inference budgets without post-hoc distillation. Experiments on ImageNet-256 and text-to-image show that WTF achieves higher reward with comparable or higher diversity than baselines, while requiring up to $280\times$ less training compute. More broadly, we argue that accelerated samplers such as flow maps are essential infrastructure for efficient post-training, and that the dominant KL-regularized formulation is only one of many choices worth revisiting.