Papers

Large pretrained time-series models such as TimesFM are attractive for financial forecasting, but raw directional accuracy is a misleading scoreboard in equity markets. An early LoRA adapter in this project appeared to reach roughly 80% directional accuracy; we show this is not evidence of skill. Over a long horizon in a rising market, a trivial "always-up" rule attains comparably high accuracy without using the input at all. To separate genuine skill from this base-rate artifact, we build a reproducible, frozen-data benchmark with expanding walk-forward folds, a stratified held-out-ticker split, honest baselines (zero-shot TimesFM, always-up, random-walk, persistence, AR(1)), and paired significance tests (McNemar, Diebold-Mariano) under Benjamini-Hochberg FDR control. We apply the identical method to two universes – a tech-heavy NASDAQ-100 and a broad S&P 500 – reporting excess accuracy over the always-up base rate. Three findings replicate. First, when the historical ~80% condition is recreated, the high number is a base rate of ~0.70 that the fine-tuned model scores below. Second, pooled LoRA shows no directional skill over the base rate at any horizon on either universe (negative at the six-month horizon). Third, per-sector specialization is significantly worse than a single pooled adapter (Diebold-Mariano p<0.001 on held-out stocks at h=128). Fine-tuning's only measurable benefit is a statistically significant reduction in point-forecast error relative to zero-shot TimesFM, which nonetheless does not beat naive baselines and confers no tradeable directional edge. The contribution is methodological: a defensible, fully seeded protocol that prevents the base-rate trap, together with the replicated negative result it produces.

Personalization in two-sided marketplaces relies heavily on user-level features, yet for platforms with infrequent, high-consideration purchases, a large fraction of users lack sufficient history for effective recommendation, spanning both paid and organic channels. At Airbnb, a substantial share of search requests comes from logged-out or first-time users, with this challenge especially pronounced on paid-channel landing pages, leaving traditional user-level features unavailable for a large fraction of traffic. Privacy regulations and increasing restrictions on third-party cookies further limit identifier-based tracking for non-essential use cases. This paper introduces Proximity Features, a privacy-compliant feature system that groups users by geographic proximity using geo-IP data and an adaptive clustering algorithm, producing aggregated user-level signals for groups of approximately 1,000 nearby users without requiring a persistent individual identifier at inference time. Privacy is preserved by design: the pipeline operates on consented, aggregated data only within consent-gated privacy controls. The system is deployed in production at Airbnb, serving multiple surfaces including marketing landing pages and destination recommendation, with engagement emails integration under way. Online A/B experiments demonstrate statistically significant lifts in bookings, with the largest gains observed among users with absent or stale history.

Few-shot industrial defect detection remains difficult for standard supervised detectors, which achieve poor performance on boundary-dominated industrial defects. This paper proposes rough path signature-guided geometry augmentation (RPS-GA), a geometry-aware approach in which Canny edge contours are treated as ordered planar paths whose truncated second-order signature responses, especially the antisymmetric Lévy-area term, are aggregated into a spatial map that highlights boundary-related structure through two fusion operators, SIG-AUG and SGAA. The approach is evaluated on NEU-DET and PCB-Defect under a few-shot protocol with 5, 10, 20, or 50 labeled images per class, using an unmodified YOLOv8n detector throughout. Compared with the baseline, RPS-GA delivers large gains when supervision is limited, although the margin shrinks as more labels become available. On NEU-DET, SIG-AUG raises 10-shot mAP@0.5 from 0.341 to 0.583, whereas on PCB-Defect, SGAA improves 10-shot mAP@0.5 from 0.086 to 0.299 and yields usable detection at 5-shot where the baseline fails entirely. These trends are confirmed by multi-seed evaluation across independent random partitions. Overall, the results indicate that second-order path-signature geometry offers a practical way to strengthen few-shot industrial defect detection without meta-learning or detector redesign.

Extended Dynamic Mode Decomposition (EDMD) approximates Koopman operators from data, but a single global operator is inefficient when different state-space regions exhibit distinct local dynamics. We introduce Cluster-Weighted EDMD (CW-EDMD), which jointly learns a soft phase-space partition and a per-cluster EDMD operator. Its Expectation-Maximization (EM) objective assigns each transition based on both geometric proximity and prediction residuals, so clusters specialize where local Koopman models are accurate rather than where the data are dense. On Lorenz, damped pendulum, and Duffing systems, across 36 configurations and 10 seeds, CW-EDMD improves matched-degree EDMD in one-step and 5s-rollout prediction. Across 288 paired comparisons, there are significant error reductions in 258 cases, increases in 4, and no differences in 26. Median one-step error reductions are 57x, 2.7x, and 12x on pendulum, Duffing, and Lorenz, respectively.

Machine-learned surrogates for the AC power flow (ACPF) problem amortize the cost of repeated solves on a fixed network, but lose one to two orders of magnitude of accuracy when a line outage changes the topology. This degradation is an operator shift. The altered admittance matrix changes the input-to-output map, so identical inputs yield a different output distribution. Existing methods correct this with target-topology data and per-topology gradient steps. We ask whether the correction can instead be made statistical and gradient-free. We propose In-Context Whitening (ICW), which trains an ACPF surrogate in an output space whitened by the base topology's first two moments, and adapts it to an unseen N-1 or N-2 topology by re-estimating that whitening from a few hundred solved cases on the new topology. This adaptation is gradient-free, weight-free, and architecture-agnostic. We prove that among affine whiteners the unique choice that preserves the coordinate-wise semantics of the physical output vector is ZCA whitening, so within efficient invertible corrections, two moments are sufficient. Across the IEEE 30-, 118-, and 300-bus systems under N-1 and N-2 contingencies, ICW reduces overall error by 6$\times$ to 28$\times$ over frozen surrogates (up to 54$\times$ per-quantity under N-2) and cuts worst-bus power-balance mismatch by up to 30$\times$, with consistent gains across three backbones. At deployment scale it matches or beats gradient-based adaptation in accuracy while adapting 21$\times$ to 34$\times$ faster, with a cost that parallelizes on commodity CPU cores rather than requiring one GPU per contingency.

Speculative execution accelerates LLM agents by using a smaller, cheaper model to predict and pre-launch the next step while the environment is idle. However, existing speculators are stateless and discard all information between tasks, preventing prediction quality from improving with experience. We equip the speculator with three online memory systems that learn from past agent trajectories: a contrastive transition table tracking action-sequence statistics, an episodic memory retrieving contextually similar segments, and a confusion tracker suppressing recurring errors. We evaluate this approach on six benchmarks spanning three speculation types: action prediction, observation prediction, and chained prediction. Memory-augmented speculation yields a 19–39% relative accuracy improvement on action prediction and up to a $2.5\times$ increase on observation prediction tasks with repetitive action spaces. These gains grow continuously as memory accumulates and generalize across speculator models of varying cost. All speculation is lossless because it runs during idle time at zero added wall-clock cost, and the actor's trajectory is identical to non-speculative execution.

Brain-computer interfaces (BCIs) decode neural activity into commands, yet most existing systems rely on fixed-window decoding that may result in redundant observation or unreliable predictions due to insufficient evidence. Adaptive temporal decision-making (ATDM) addresses this accuracy-time trade-off by progressively accumulating EEG evidence and deciding when to stop. However, existing EEG encoders are mainly designed for fixed-window decoding and may not provide reliable state representations under variable observation lengths. In addition, current ATDM-oriented encoders are typically tailored to specific EEG paradigms, limiting their applicability across different BCI tasks. To address these limitations, we propose ProtoTrigger, a two-stage prototype learning-based EEG state encoder for ATDM. ProtoTrigger uses prototype matching to extract stable local EEG embeddings and prototype-based attention to aggregate decision-relevant temporal evidence during progressive observation. Offline evaluations across three EEG paradigms demonstrated state-of-the-art accuracy-time trade-offs and strong generalizability across different EEG paradigms. An online human-in-the-loop augmented reality-based BCI experiment further demonstrated its real-time feasibility. These results suggest that ProtoTrigger provides a general EEG state encoding framework for efficient ATDM-based BCI systems.

Neurological prognostication after cardiac arrest commonly relies on electroencephalography (EEG). However, EEG demands high clinical resources. Bedside electrocardiography (ECG) is standard and low-cost. Yet, its value for predicting neurological outcomes remains underexplored. In this study, we propose NeuroECG, an ECGFounder-based deep representation framework for EEG-free auxiliary prognostication. NeuroECG adapts a pretrained ECG foundation model via task-specific fine-tuning. We implement a gradual unfreezing strategy on single-channel bedside monitoring ECG. Multiple ECG segments per patient are encoded into segment-level deep features. These embeddings are aggregated via quantile pooling (q = 0.24) and compressed using principal component analysis (PCA). Experiments on 412 ECG-available patients from the multicenter I-CARE database show that the adapted ECGFounder backbone achieves the best performance among ECG-only backbone baselines, with a test AUROC of 0.7333. We further combine the learned deep ECG representation with static clinical covariates. The proposed NeuroECG model achieves a test AUROC of 0.8077 and an AUPRC of 0.8970. These results support deep bedside ECG representations as a useful source of auxiliary prognostic information. Their integration with static clinical covariates improves prediction in an EEG-free setting. The source code is available at https://github.com/goddream66/NeuroECG

Reinforcement Learning (RL) has been extensively used in autonomous and self-* systems, but RL policies, especially deep RL ones relying on neural networks, lack transparency and are difficult to understand. This can lead to diminished user trust, and makes for a more challenging verification of systems. To address this challenge, this paper introduces Explanations using Alternative Realities for Reinforcement Learning (EARL), a Python library to produce counterfactual explanations in RL settings. This library allows the user to produce explanations by exploring What-if scenarios to clarify agent behavior by comparing possible outcomes. Counterfactual explanations have been shown to be intuitive and user-friendly in psychology research, but have only recently been explored in RL, with existing implementations usually limited to toy examples and benchmarks. EARL supports counterfactual explanation generation in realistic RL-based self-adaptive systems. To demonstrate its applicability, we demonstrate its use in a simulation of CitiBikes, a self-adaptive bike-sharing system, and we provide evaluations showing how it performs in real applications.

We present a structured review of commonly used Explainable machine learning (XML) methodologies, including global and local interpretability tools such as SHapley Additive exPlanations (SHAP), Local Interpretable Model-Agnostic Explanations (LIME), Partial Dependence Plots (PDP), and Individual Conditional Expectation (ICE) plots. For each method, we explain the underlying mechanism at a high level, visualize representative outputs, and provide structured guidance on interpretation, appropriate use, and limitations, illustrated using the publicly available Heart Disease dataset. XML techniques provided intuitive visual and quantitative insights into how predictors influence model predictions. Global methods characterized population-level feature effects, whereas local methods revealed patient-level contributions useful for individualized interpretation. Our worked examples demonstrate how XML outputs can identify nonlinear relationships, detect interaction effects, and reveal heterogeneity in predicted risk across patients, addressing key challenges in translating ML predictions into interpretable outputs for clinical research. XML tools offer valuable interpretability for ML models and support more transparent and accountable ML applications in clinical research. By providing a methodologically grounded overview alongside practical implementation examples and structured guidance on each method's strengths and limitations, this primer helps bridge the gap between advanced ML methodology and clinical applicability. Thoughtful adoption of XML approaches may facilitate better understanding, communication, and critical evaluation of ML predictions in healthcare research, ultimately supporting evidence-based clinical decision-making.

We introduce a vocabulary for automated research systems built from one or more agents to make their design choices easier to describe and compare. The vocabulary specifies 1) who the agents are, 2) what operations are available in the system, 3) who may invoke them, 4) how agents communicate, 5) what information is visible within and across runs, 6) how the next action is chosen, 7) how a run begins, and 8) how outputs are evaluated. A trajectory records one run from the input task to the returned artifact. Because agents, operations, and initialization may be stochastic, repeated runs on the same task induce a distribution over trajectories rather than a single behavior. Our vocabulary turns structural design questions, such as when agents should communicate, gain or lose a capability, or carry information across runs, into testable choices. It also makes the evaluator a component of the system, since reported gains depend on how closely the proxy score matches true quality. That separation also splits the vague complaint that these systems lack taste into two failures with different solutions. Generative taste is the rate at which a system proposes novel trajectories before any score is observed, and evaluative taste is the gap between the proxy score and the quality it should match. We instantiate the vocabulary on recent autoresearch systems to illustrate that it covers designs that differ widely in structure.

Trajectory datasets used in ADAS evaluation are heavily biased toward routine driving; genuine vehicle-to-vehicle conflict events are rare, and the rarer the event, the higher the cost when an ADAS system fails to handle it. Existing generative approaches address this imbalance by conditioning on scene-level properties - spatial goals, agent structure, or natural-language adversarial objectives - but none can accept a target Time-to-Collision (TTC) value as input and be held to producing it within a measurable error. This paper introduces SevDiff, a severity-conditioned denoising diffusion probabilistic model (DDPM) that accepts a requested minimum TTC value as a scalar conditioning signal and generates paired vehicle interaction trajectories whose realized conflict severity matches the request, evaluated through a hit-rate metric. Trained on 468 interaction windows extracted from the UTE SQM-W-1 expressway weaving-section dataset (1,041 vehicles, 822,691 observations after smoothing), SevDiff achieves 100% hit-rate within +/-0.5 s for TTC targets of 0.5-1.5 s and 97-99% at 2.0-2.5 s, with graceful degradation to 39% at TTC = 5.0 s. Generated kinematic features are physically plausible, with a maximum out-of-range rate of 4.7% across 12 features and no negative speed or gap values in more than 96.5% of samples. The hit-rate degradation pattern is physically interpretable as the strength of the conditioning signal relative to the training prior, making it a precision characterization of the generator rather than a pass/fail result.

Higher-order optimizers such as Muon and SOAP offer faster convergence than AdamW, but their computational cost and numerical stability challenges have limited adoption at scale. In this work, we adapt and enhance preconditioned gradient methods to overcome the practical challenges of large-scale LLM pretraining. We first identify instabilities in SOAP at large batch sizes and propose algorithmic modifications including per-step QR orthogonalization and improved preconditioning strategies that eliminate loss spikes and enable stable training in these regimes. We then present a unified empirical study of SOAP, Muon, and AdamW using update-RMS matching to ensure fair learning rate transfer across optimizers. As part of this analysis, we empirically evaluate the orthogonalization quality of Muon. Our experiments on multi-billion-parameter models trained on trillions of tokens reveal that SOAP and Muon consistently outperform AdamW at the scales we tested. Notably, at batch sizes of up to 100M tokens for next-token prediction, these optimizers maintain training stability and quality while AdamW degrades. To enable efficient training at large scale, we introduce a layer-wise distributed optimizer compatible with Megatron-LM. Our implementation balances memory and hides communication while avoiding approximations to the optimizer computations, thus retaining their convergence benefits. Additionally, we identify and build specific system-level improvements to further accelerate our layer-wise implementation. To support the research community, we release a codebase that contains emerging algorithms for optimization: https://github.com/NVIDIA-NeMo/Emerging-Optimizers

Safe Advanced Air Mobility operations require aircraft to maintain separation when surveillance information is noisy, delayed, incomplete, or temporarily unavailable. This study develops a Deep Q-Network-based Multi-Agent Reinforcement Learning framework for decentralized conflict resolution among heterogeneous small unmanned aerial vehicles and electric vertical takeoff and landing aircraft operating within a structured three-dimensional corridor. Separate policies are trained for the two aircraft categories using local observations and a 14-action space that includes maintaining course, turning, vertical maneuvering, landing, and speed control. The simulation incorporates aircraft-specific dynamics, energy use, corridor constraints, observation noise, communication delay, information dropout, wind disturbance, actuator uncertainty, and model uncertainty. The trained policies are evaluated across 90 combinations of traffic density and minimum separation thresholds. Loss-of-separation frequency and duration generally increase with traffic density and separation requirements, although most events are resolved within 1s. Under safe conditions, agents maintain their motion approximately 79% of the time. During conflicts, turning accounts for 33% of actions, followed by maintaining motion at 29%, speed control at 25%, and vertical maneuvers at 13%. Six Pareto-optimal configurations reveal trade-offs between safety and corridor capacity. The framework supports the simulation-based evaluation of safer AAM conflict-resolution strategies under degraded surveillance conditions.

We show that a single climate realization can be decomposed into forced and internal components by treating external forcing as a dynamical driver within a linear stochastic system, an idea grounded in pullback attractor theory. In doing so, we address a central methodological challenge in climate science with direct implications for climate projection and the detection and attribution of the forced response, disentangling the forced climate response from internal variability in a single observed record. Statistical methods range from approaches trained on large ensembles to techniques operating on single realizations. The latter often rely on linear frameworks such as linear inverse models (LIMs) and linear regression. LIMs ignore forcing predictors, whereas linear regression omits climate system dynamics. Here we introduce PullbackDMDc, a method grounded in non-autonomous dynamical systems theory and dynamic mode decomposition with control (DMDc), incorporating pullback attractor estimation to decompose a single climate realization into spatial modes and their associated forced and internal components, yielding a physically interpretable picture of the underlying dynamics. We illustrate the utility of PullbackDMDc for Earth System Model (ESM) evaluation by applying it to near-surface air temperature and sea-level pressure from reanalysis and four ESM large ensembles. PullbackDMDc estimates the forced response with skill matching or exceeding established baselines and identifies optimal forcing predictors against model-based ground truth. Its internal variability components reveal that ESMs qualitatively capture interannual and decadal modes while exhibiting systematic differences relative to each other and to observations. Skillful forced response estimation and a novel decomposition position PullbackDMDc as a practical tool for single-realization climate analysis and ESM evaluation.

Edge intelligence systems, the intersection of edge computing and artificial intelligence (AI), are pushing the frontier of AI applications. However, the complexity of deep learning models and heterogeneity of edge devices make the design of edge intelligence systems a challenging task. Hardware-agnostic methods face some limitations when implementing edge systems. Thus, hardware-aware methods are attracting more attention recently. In this paper, we present our recent endeavors in hardware-aware design and optimization for edge intelligence. We delve into techniques such as model compression and neural architecture search to achieve efficient and effective system designs. We also discuss some challenges in hardware-aware paradigm.

Continuous EEG monitoring for epilepsy is constrained by the limited power and memory budgets of wearable and implantable devices. Deep neural networks can detect seizures with high accuracy, but their computational cost and model size make them difficult to deploy on such platforms. In this work we use a single 1D CNN seizure detector on the CHB-MIT scalp EEG dataset as a common baseline, and then investigate three brain-inspired efficiency strategies: (i) conversion of the CNN into a spiking neural network (SNN) via parameter transfer, (ii) EEG channel pruning combined with 2:4 structured weight sparsity, and (iii) INT8 quantization using FX- and ONNX-based workflows, including quantization-aware training and operator fusion. The quantized CNN variants reduce stored model size from 1.63 MB to 0.44 MB, lower estimated energy per inference by up to 64%, and achieve as much as 2.8 times speedup in CPU latency while preserving, and in one case slightly improving, AUC. The pruned CNN halves the number of input channels and non-zero weights with only a modest accuracy drop, and the SNN conversion provides a spiking implementation with sparse temporal activity. Together, these experiments characterize three complementary efficiency directions for seizure detection.

Noisy-label learning in small-scale medical image classification is challenging and hinders the superiority of deep neural networks. Recent studies suggest that quantum neural networks (QNNs) have shown potential in limited-data regimes, yet their use for noisy-label learning remains under-explored. A key obstacle is QNNs' intrinsic "natural smoothness", which may regularize training but also obscure high-confidence samples needed for noise-transition estimation. We propose Supermartingale-based Label Transition (SLT), an anchor-free loss correction framework for robust QNN-based medical image classification under noisy labels. SLT models entropy reduction in predictive distributions as a supermartingale and uses its monotonic behavior to identify stable transition-matrix refinement steps. This enables dynamic transition updates while reducing noise-driven oscillations during QNN training. We further provide a convergence analysis showing that the proposed transition-refinement process reaches a steady state. Experiments on multiple public small-scale medical image datasets demonstrate that SLT consistently improves QNN-based classification and stably outperforms classic noise-label learning baselines under synthetic and real-world label noise.

The notion of circuit refers to a cyclic oriented influence between the elements of a dynamical system. There are two classes of circuit: positive and negative. R. Thomas conjectured that a necessary condition of multi stationarity is the existence of positive circuits. In this paper we use dynamical system tools and planar analysis to find conditions for which the conjecture holds for planar systems.

Bayesian Belief Networks (BBNs) are powerful tools for decision-making under uncertainty. However, building their structures and estimating parameters are difficult. Currently, researchers must choose between relying on expert judgement or using large datasets to learn the structure and parameters of the network. We propose a new methodology using Large Language Models to bridge the gap between expert opinion and data-driven learning. This approach uses a panel of AI agents to estimate probabilities based on specific personas and context. We then apply a trimmed-mean rule to remove noise from these responses. We develop a six step BBN framework and illustrate it to model customer intention to consult a doctor in an alternative healthcare system. The model reveals that while self efficacy appears to be a major factor, its actual causal impact is small. In contrast, subjective norms have a much stronger effect in modelling customers' intention. The most effective strategy is to improve both confidence and community norms simultaneously.

Large Language Models (LLMs) are rapidly moving from research settings into the wild, deployed on enterprise infrastructure, personal devices, and edge platforms. While cloud deployments offer scalable compute, concerns over data sovereignty, compliance, latency, and third-party dependence are driving organizations toward edge and on-premise LLMs. This shift introduces new security and privacy challenges: limited compute and memory force aggressive optimizations, including quantization, pruning, model partitioning, and parameter-efficient adaptation, each of which can introduce vulnerabilities and reshape the threat landscape. We describe this tension as the Security-Efficiency Paradox, mechanisms that improve efficiency may weaken robustness, expose new attack surfaces, or increase privacy risks. We examine how compression can degrade safety alignment, how partitioned inference enables reconstruction attacks, and how continuous local adaptation may cause privacy leakage and model drift. To analyze these risks, we introduce a deployment-centric taxonomy organized around three architectural constraints: the Memory Wall, the Quadratic Wall, and the Compute Wall. We derive a unified constraint model that quantifies when unsafe optimizations become unavoidable, linking each wall to specific attack surfaces. Building on this model, we propose the Secure Operational Efficiency Score (SOES), a holistic metric balancing task accuracy, jailbreak resistance, and privacy against energy, memory, and latency, enabling practitioners to configure edge LLMs under real-world hardware limits. We further present a practical decision procedure and targeted mitigations for each optimization-induced vulnerability. Together, these contributions provide a co-designed framework for jointly evaluating security, privacy, and efficiency, laying a foundation for securing edge-native intelligent systems.

We present nsfaguard, a guardrail framework for securing agentic AI systems against operational threats, such as prompt injection, sensitive information extraction, malicious code requests, dangerous tool misuse, and resource exhaustion. We first introduce the NSFA taxonomy, which organizes 185 risk variants into a CIA-triad-grounded hierarchy and is cross-validated against three well-established OWASP guidelines. Based on this taxonomy, we construct a benchmark suite spanning 133 languages, comprising over 93K purpose-built samples targeting both user queries and agent responses, along with 3,435 cross-source samples adapted from five public agent-security datasets. To detect these operational threats in practice, we develop a dual-mode approach combining SFT-based generative reasoning for interpretable offline auditing with discriminative classification heads on the frozen backbone, enabling real-time detection at approximately 50,ms. We release four models with 0.8B, 2B, 4B, and 9B parameters, all achieving $\geq$94% F1 on purpose-built benchmarks and surpassing the strongest competing guardrails by 6 to 12 absolute points. On cross-source evaluation, the 9B model attains 91.29% F1 with a more balanced precision–recall trade-off. Moreover, ablation experiments show that classification heads can equip a guardrail with risk detection capabilities beyond its original scope and achieve state-of-the-art performance. These results demonstrate the extensibility of the approach and its generality as a plug-in enhancement.

Accurately assessing personality from text is challenging because traits are latent, context-dependent, and often subtly expressed across long narratives. Large language models (LLMs) offer new opportunities by processing extensive textual contexts, but pretraining of these models can induce latent "personality-like" biases, making single-model inferences inconsistent. We propose a fine-tuned multi-agent framework for detecting OCEAN personality traits, in which sub-agents are conditioned to adopt high, low, or neutral perspectives for each trait through masked language modeling (MLM) and psychometric supervision. A judge LLM aggregates and compares sub-agent outputs to generate final trait predictions, capturing multiple complementary perspectives while mitigating individual model biases. We evaluate the framework on life narrative dataset through quantitative and qualitative experiments, including baselines, ablations, and inference quality analyses. Our approach offers a scalable and interpretable method for text-based personality inference, highlighting the benefits of multi-agent reasoning grounded in psychometric supervision.

We study deletion in a context memory that fits a support-vector boundary around stored keys and uses the resulting nonnegative coefficients to weight their values. An exactly zero coefficient lets us remove a key without changing the current normalized readout. In a three-key construction, however, a later admission makes the discarded key active in a solve over the full history. For a positive-weight key, a decremental update targets a refit on the remaining keys at the original coefficient cap, under its stated assumptions. Earlier standalone comparisons show close typical gate-score agreement, while similar keys with different values can yield larger readout differences. The journal extension tests chained deletions and admissions, checking coefficient feasibility and optimality conditions in both returned states and fresh references. The initial audit completes 66 of 80 trajectories and exposes numerical failures. A separately checked wrapper completes all 5,120 scheduled operations at unchanged acceptance thresholds, with eight initialization rescues and 84 update rescues. Its largest readout difference is 0.5273% of retained-value range. These finite results support explicit acceptance and rescue rules for the evaluated memory edits. They do not extend the current-state removal guarantee to arbitrary future admissions.

Decentralized gradient descent (DGD) is widely used for solving distributed optimization problems over networks of agents. While its convergence properties are well understood, less is known about the communication and computation resources required to attain a prescribed accuracy. In this paper, we study DGD from a resource-aware perspective and characterize the communication-computation budget required to attain a target error level. We develop a bottleneck-centric framework in which different factors dominate the optimization dynamics at different error scales. Specifically, we identify operating regimes governed by initialization, objective heterogeneity and network connectivity, gradient noise, and communication noise. To capture these effects, we introduce two fundamental quantities: the gradient-Diversity-to-Network-connectivity Ratio (DNR) and the Gradient-to-Communication-noise Ratio (GCR). We show that these quantities determine the sequence of bottlenecks encountered during optimization and the corresponding budget-optimal operating strategy. Using a multi-stage analysis, we derive optimal stepsize selections and explicit budget-complexity bounds that quantify the budget resources required to attain a prescribed accuracy. The resulting expressions reveal how the overall budget decomposes into contributions associated with successive bottlenecks and provide insight into the fundamental tradeoffs among objective heterogeneity, network connectivity, gradient noise, and communication noise.

In rectified-flow-based generative models, the neural network can be trained to predict two different targets, such as the instantaneous velocity or the data endpoint, to perform denoising. Although prior work shows that these parameterizations lead to different empirical behaviors, the mechanisms underlying their respective advantages remain to be underexplored, and how to combine them effectively is still unclear. In this work, we analyze how learning errors from different parameterizations affect the generation performance. We show that predicting the data endpoint has a clear training signal that stabilizes training, whereas predicting the velocity maintains stable sampling dynamics near the data manifold. Motivated by these insights, we propose Self-Consistent Flow (SC-Flow), a new method that unifies the benefits of both parameterizations. By employing a lightweight consistency loss, SC-Flow jointly trains a single network to predict both the local velocity and the data endpoint, and the consistency between the two predictions improves the model's performance. The method requires no major architectural changes and adds minimal computational overhead. Extensive experiments on image generation tasks demonstrate that SC-Flow substantially stabilizes optimization and improves the straightness of generation paths, leading to significant gains in generation quality over standard rectified-flow baselines.

Sparse autoencoders (SAEs) are the standard for decomposing superposed neural representations into interpretable features, and evaluation relies predominantly on correlational recovery metrics – cosine similarity between ground-truth directions and decoder atoms. We show this conflates two distinct claims: decoder-geometry alignment and encoder-activation behavior. We reproduce the superposition phase diagram of Elhage et al. (2022), identifying a convergence artifact at high sparsity and an under-described diffuse sharing regime at extreme overcompleteness. We reproduce the TopK-versus-L1 comparison of Gao et al. (2024), with direct evidence of L1 shrinkage. Our central result is causal: subjecting every recovered feature to ablation and steering, we find up to 77% of features passing a recovery bar (cosine >= 0.90) in a degraded SAE – and 9% in a well-trained one – are causally inert: the matched atom never fires when the feature is present, including matches at cosine ~1.000. We package the method as sae-causal-audit, a model-agnostic instrument with a deterministic pipeline. Re-auditing refines the finding: inertness decomposes by cause into structural inertness (antipodal-pair geometry, present in good SAEs) and competitive inertness (a TopK pathology of degraded SAEs), and by direction into read- and write-inertness, which five antipodal pairs dissociate completely – unmonitorable yet steerable through the same atom, with steering specificities of 143-310 attached to zero ablation effects. We document why byte-exact reproducibility is unavailable by construction, and propose reporting it as a stack of claims with explicit scopes. Applying the instrument to a production SAE reproduces the pattern at small scale (14% inert) and surfaces an atom-collision signal: a handful of atoms recur as the nearest match for dozens of unrelated concepts, replicated across three batches.

True causal relationships are rarely known, and inferring causal graphs from data is hard. A fundamental challenge is how to assess whether a given causal graph is good in the absence of a ground truth. We propose falsifying candidate causal graphs based on whether they can explain the propagation of an outlier event. Our approach leverages a key principle: weak outliers rarely cause strong ones. While this principle has previously been used in root cause analysis to identify root causes without prior knowledge of the graph, we turn it on its head and use it to falsify candidate causal graphs whose implied outlier propagation is inconsistent with the data. To this end, we present the first statistical tests for the hypothesis that a candidate graph is the true causal graph, and show they have false positive control, power guarantees against incorrect causal graphs, and can operate with a single outlier sample.

Accurately modeling the dynamics of planetary parachute and entry vehicle systems is critical for Entry, Descent, and Landing events such as vehicle separation and sensor activation. These dynamics are difficult to capture with traditional system-identification methods as parachute motion is highly nonlinear, the governing equations are not fully known, and relevant test data are scarce and expensive to acquire. In this work, we sidestep these challenges by leveraging a physics-aware generative modeling approach that learns parachute dynamics directly from data. The proposed method, Symplectic Parachute Generative Adversarial Network (SPar-GAN), adapts a Hamiltonian generative architecture to the parachute setting by conditioning on canopy design and freestream velocity, while enforcing conservation of energy through symplectic integration. We apply SPar-GAN to subscale parachute tests conducted at the National Full-Scale Aerodynamics Complex and show that it reproduces qualitatively accurate pitch-yaw dynamics of different parachute configurations while recovering a compact two-degree-of-freedom phase-space consistent with canopy axisymmetry. These results suggest that physics-constrained generative models can characterize parachute dynamics across operating conditions and may help reduce the volume of physical testing required to assess performance.

We present an agentic approach to autonomous neural operator discovery based on an AI scientific community, which consists of a swarm of virtual laboratories that interact under a citation-based economy of influence. Highly-cited labs found new labs that follow their research direction and replace non-performing labs. Each virtual lab contains three agents: an LLM planner that proposes an architecture, a numerical worker that trains and measures it, and an LLM reviewer that participates in cross-lab peer review. All labs share a common vocabulary consisting of DeepONet (branch-trunk), Fourier, Transformer (attention), wavelet, and residual convolutional neural operator building blocks. We evaluate the neural operator AI scientific community on five problems, namely piecewise regression, the linear advection and Burgers 1D PDEs, and the Navier-Stokes and Darcy flow 2D PDEs, while repeating the simulation three times for each problem. The results show that the neural operator AI scientific community is capable of discovering high-accuracy, low-parameter-count neural operator architectures. All 9,623 LLM calls are logged and audited, which reveals that the virtual lab LLM planners choose to hybridize in 99.8% of their logged decisions, consistently returning multi-family hybrids. Moreover, we conducted an ablation study by replacing the LLM agents in each lab by rule-based alternatives, which caused the scientific community to collapse to non-hybridized single-family stacks in several cases, showing that LLM agency is needed to preserve diversity. The results suggest a no-free-lunch theorem for neural operators: there is no universal winner. The code, configurations, and the complete LLM transcripts are released at https://github.com/luislootx/AI-SC.