arXiv subjects

Search papers

Search indexed arXiv papers on artificial intelligence, large language models, computer vision and robotics. Read source abstracts and follow links to arXiv.

At least 1,747 records · Page 97Linked to original sources

arXiv2026

GNN-VIEW: Visual Inspection and Explanation of Graph-based Earth-System Models

Machine learning Earth-system models are becoming increasingly important, and so is their development. Graph-based models, one of the most prominent approaches, allow developers to constrain the learning process before training via graph connectivity. Yet, developing these models remains challenging: the graph is built algorithmically with little intuition about the final result beyond static plotting, which requires extra work and is often not detailed enough. Even after successful training, how the model arrives at forecasts remains obscure. We present GNN-VIEW, an interactive browser-based tool that supports the development of such systems. GNN-VIEW exposes the graph's information pathways, allowing model developers to debug the graph structure before training. After training, GNN-VIEW uses attribution-based explainability to reveal which input features are used by the trained model for a given forecast. This helps developers understand why a model fails. We showcase the usefulness of GNN-VIEW in the construction stage of a graph for an ocean forecasting model, and in explaining poorly predicted precipitation of a trained weather forecasting model during Hurricane Laura.

cs.LG↗

arXiv2026

Not One Belief: Answer Payloads, Persistence, and Dissociated Confidence Signals in Conversational Question Answering

Conversational QA systems should fix errors when users push back and signal which answers to trust. We split each follow-up into its speech act (re-ask, agree, or challenge) and its payload (an answer string, correct or not), and measure four confidence signals on the same transcripts: verbalized confidence, length-normalized likelihood, a frozen hidden-state probe, and P(True). Answers change when the user supplies an answer string and almost never under the speech act alone: in every retained full-size cell, the correct answer converts at least four times as many wrong answers as re-asking or agreement. The four signals disagree. On Qwen2.5-7B the frozen probe rises with answer persistence rather than truth; this holds on held-out questions, with Llama-3-8B (cleaned answers) as support, but not on four other models. Among 81 Qwen2.5-7B errors that survive one re-ask, re-asking, a fresh chat, or abstain repair almost none, while a gold answer repairs 75-95%. In two preregistered replications, every repair rate without an answer string stays below 0.15; a larger post-hoc rerun excludes 0.15 except for the empty elicit. With BM25 TriviaQA evidence, sentences that contain the answer repair 67-76% of persistent errors versus at most 12% when they do not. An end-to-end contain-gate cuts harm on TriviaQA, but a deployable gate that does not know the gold answer repairs few errors.

cs.CL↗

arXiv2026

Towards a Unified Framework of Optimal Input-Side Dimensional Configuration for Deep CTR Prediction

Deep click-through rate (CTR) prediction has achieved substantial success through increasingly expressive feature interaction architectures. However, predictive performance also depends on how input-side information is configured before interaction modeling. Existing studies have shown that selecting informative samples, features, or embedding units can improve prediction accuracy. Nevertheless, they primarily optimize these selection mechanisms independently along a single input-side dimension, overlooking their shared basic form and the inherent dependencies across input levels, thereby limiting the effectiveness of input-side dimensional configuration. To address this issue, we propose OptIDC, a unified framework for Optimal Input-side Dimensional Configuration in deep CTR prediction. OptIDC jointly learns sample, feature, and embedding-unit level configurations within a unified framework by integrating their corresponding selection mechanisms. To explicitly coordinate selection decisions across different levels, OptIDC further introduces a selection-weighted Optimal Transport mechanism that aligns coarse- and fine-grained selections via transport over finite measures, thereby alleviating cross-level selection conflicts while preserving the semantics of each input level. Moreover, OptIDC is model-agnostic and can be seamlessly integrated into existing CTR models. Extensive experiments on four real-world benchmark datasets demonstrate that OptIDC consistently achieves higher AUC than existing input-side selection methods, validating the effectiveness of unified input-side dimensional configuration.

cs.IR↗

arXiv2026

Reflective Computing

Interactive technology is increasingly built for speed, and generative AI now answers before a question is fully formed. Yet the capacity to pause, interpret and decide for oneself is central to wellbeing, learning and meaningful use. HCI has named paradigms for sensing emotion, changing behavior, calming attention and tracking the self. However, none of them takes human reflection, and the slowing down it requires, as its organizing aim. This article defines that paradigm, reflective computing, and derives it from what thirty years of designs lack. We analyzed 134 interactive systems from 17 ACM venues (1996-2026) in two phases. First, an abductive gap analysis read every system against reflection theory, and seven recurring gaps defined seven dimensions, namely tempo agency, invitational form, interpretive openness, reflective depth, temporal span, protected idleness and ethical non-extraction. Second, two independent coders scored every system with a codebook (Krippendorff's alpha 0.67 to 0.84 for five dimensions; ICC 0.82 for the overall score). The system sets the pace in 51 designs, 74 support only momentary use, only 12 scaffold reflection beyond description, and 56 papers are silent on the inner states they collect. No system expresses all seven dimensions strongly, and only 20 meet the definition. Slow and experience centered design comes closest, whereas persuasive technology and human-AI interaction score lowest. Factor and latent class analyses reveal two cultures of reflection, one working through form and the other through scaffolded thought, which only one system combines. Systems that let people set the pace score markedly higher, even with paradigm and form controlled (odds ratio 13.5). Hence, reflective computing is proposed as a paradigm whose aim is a human capacity rather than a system function, and which treats slowness as valuable only when people choose it.

cs.HC↗

arXiv2026

MiRL: Error-Targeted Reinforcement Learning for Correcting Terminal Agent Misbehaviors

Error-targeted learning gives a legible way to improve a model: identify specific weaknesses of the model, train the model to fix these weaknesses, and check whether the weaknesses are fixed. We study how to apply error-targeted learning to terminal agents, which solve long-horizon tasks specified in natural language by interacting with a command shell. Existing error-targeted methods are designed for single-turn tasks, and existing analyses of agent failures are used to evaluate and debug agents rather than to train them. We propose Misbehavior-driven RL (MiRL), which first profiles the errors of a policy to find generalizable misbehaviors: kinds of wrong steps that recur across many environments. MiRL then trains the policy with RL, starting rollouts at the steps where these misbehaviors occur and rewarding a rollout only if it passes the tests and avoids the misbehavior. On Qwen3.5-4B, eight misbehaviors account for 52.1% of wrong steps. After training with MiRL, the policy repeats the measured misbehaviors 43% to 57% less often on average on unseen prefixes. MiRL outperforms TMax-4B on Terminal-Bench Lite (51.8% vs. 47.4%) and matches it on Terminal-Bench 2.1 (21.3% vs. 20.6%), while using 35% of its rollouts. These results show that error-targeted learning can be applied to realistic, long-horizon terminal tasks.

cs.LG↗

arXiv2026

From Self-Confidence to Statistical Accuracy: Scaling Human-in-the-Loop Evaluation with Direct and LLM-Assisted Auditing

For large language models (LLMs) and multimodal large language models (MLLMs), model-reported self-confidence is often miscalibrated, yet it may still be useful for efficient evaluation. At the same time, LLM-as-a-judge offers a scalable alternative to direct human auditing, but its practical benefit depends on calibration quality and inference cost. This paper develops a unified statistical framework for both evaluation routes under alternative sampling strategies. For direct auditing, confidence-stratified sampling is used to recover population accuracy from partially verified outputs without requiring calibrated confidence. Finite-sample uncertainty is characterized, and the analysis shows how highly uneven confidence strata affect audit efficiency and label allocation. For judge-assisted evaluation, the judge is modeled through sensitivity and false-positive rate, yielding an error-corrected estimator with calibration uncertainty. The two routes are then compared through human-label and total-cost efficiency boundaries. Experiments on MMLU-Pro and MMMU-Pro show that miscalibrated self-confidence can still support statistically valid stratified auditing. Under highly imbalanced confidence distributions, proportional allocation substantially reduces estimation error relative to equal-per-stratum allocation. For judge-assisted evaluation, only a small subset of target--verifier pairs satisfies the idealized human-label efficiency condition, and automated inference costs can further limit potential savings. These findings clarify the statistical and economic conditions under which confidence-guided auditing and automated verification are useful.

cs.CL↗

arXiv2026

Bell-pair distillation under imperfect operations

We recently proposed an entanglement distillation protocol based on local operations and classical communication that probabilistically and universally distills a maximally entangled Bell pair from multiple copies of an unknown two-qubit state [Phys. Rev. Lett. 135, 260202 (2025)]. Here, we investigate the robustness of this protocol against imperfections in its implementation. We consider several incoherent and coherent gate-error models, as well as measurement readout errors, and determine the regimes in which distillation enhances the fidelity with the target Bell state. We find that the protocol remains effective over substantial error ranges, although its robustness depends strongly on the type of error. We further analyze asymmetric errors between the two parties and identify conditions under which complementary postselection outcomes yield identical distillation dynamics, allowing the success probability to be enhanced.

quant-ph↗

arXiv2026

Unconditional scattering below the ground state for the three-dimensional focusing energy-critical generalized Hartree equation

Unconditional global existence and scattering are established for the three-dimensional focusing energy-critical generalized Hartree equation with $0<α<3$, below the ground-state energy and kinetic thresholds. The initial data lie in $\dot H^1$, with no radial symmetry, finite-mass, or finite-variance assumption. Annular frequency estimates and Gaussian interaction estimates rule out a compact critical trajectory by first controlling its nonlinear forcing and then recovering its conserved mass.

math.AP↗

arXiv2026

Degree-list Kempe equivalence for 3-connected graphs

In a proper coloring, a Kempe change interchanges two colors on a connected component of the subgraph they induce. For a fixed list assignment $L$, a Kempe change is $L$-valid if it transforms an $L$-coloring into another $L$-coloring. We prove that if $G$ is a $3$-connected graph other than a complete graph or the triangular prism, and $L$ is a list assignment with $|L(v)|\ge d_G(v)$ for every vertex $v$, then $G$ has a proper $L$-coloring and any two such colorings can be transformed into one another by a finite sequence of $L$-valid Kempe changes. This proves a conjecture of Chakraborty, Feghali and Mahmoud.

math.CO↗

arXiv2026

Weekend Crypto Volatility and Monday Market Risk

Cryptocurrencies trade continuously while major traditional markets close over the weekend. We examine whether weekend cryptocurrency volatility predicts Monday variance in equity, commodity and fixed-income markets. Using daily range-based variance measures for Bitcoin, Ether, XRP, the S&P 500, gold and a US Treasury ETF over 2017-2026, we augment heterogeneous autoregressive models with a common weekend cryptocurrency factor. Predictive content is weak before 2024 but significant thereafter. In the post-2024 sample, a one-standard-deviation increase in the factor is associated with 27-54% higher fitted Monday variance. Adding the factor reduces out-of-sample QLIKE losses by 11-22%, although statistical evidence varies across markets. Friday cryptocurrency volatility does not reproduce the result, and coin-specific variation adds no significant predictive information beyond the common factor. An endogenous procedure estimates a break in late 2023, during growing anticipation of US spot-Bitcoin ETF approval. However, ETF-flow interactions provide no evidence of amplification through greater measured Monday flow activity, leaving the mechanism unresolved. The findings suggest that calendar-specific volatility measures reveal predictive information that aggregate connectedness measures may obscure.

q-fin.ST↗

arXiv2026

pbhalo: computing primordial black hole and minihalo populations from a single primordial perturbation spectrum

An enhancement of the primordial curvature power spectrum on small scales can give rise to two populations of compact objects: primordial black holes (PBHs), which form during radiation domination, and dark matter minihalos, including ultracompact minihalos (UCMHs), which form from the same modes during matter domination. We present pbhalo, an open-source Python package that computes the PBH mass function, the halo mass function and the UCMH abundance from a single spectrum in a given cosmology and applies observational constraints on the PBH abundance and on the spectrum jointly. The package components are validated against colossus (a code for halo mass functions and cosmological quantities), pbhstat (a code for the PBH mass function from a given spectrum) and published results; joint constraints are illustrated for lognormal peaks. The code is distributed under the MIT license.

astro-ph.CO↗

arXiv2026

Explainable Multimodal AI Approach to Reduce Unnecessary Breast Biopsies

Recently, there has been rise in multimodal vision-language models for computer-aided diagnosis (CAD), yet their application in breast diagnosis, particularly for predicting biopsy outcomes remains limited. This study presents a multimodal vision-language approach, incorporating Grad-CAM as an Explainable AI (XAI) technique, to predict benign tumors in patients recommended for biopsy, especially those classified as BI-RADS 4 and 5. A novel multimodal framework was devel- oped using Vision Transformer (ViT) for imaging feature extraction and DeBERTa-V3-Large for textual embeddings. The model was trained and evaluated on a comprehensive in-house dataset comprising 1,632 patients, each with four mammography views and a corresponding radiology report. The features of both modalities were extracted independently and fused by late fusion before classification. Grad-CAM was used to generate visual explanations highlighting key regions influencing model predictions. The proposed model demonstrated high sensitivity 90%, 95%, 99%, and 100% with corresponding specificities of 40%, 36%, 18%, and 11%, respectively. The best AUC achieved was 0.72 by multimodal approach on the test set, outperforming the image only (0.66) and text-only (0.68) baselines. The proposed multimodal architecture, enhanced with Grad-CAM, demonstrates a strong potential to identify benign tumors among high-risk patients, offering interpretable insights and thereby helping reduce unnecessary biopsies, patient anxiety, and healthcare costs.

cs.AI↗

arXiv2026

Valence-quark-core and meson-cloud effects on electromagnetic form factors of $Δ$ baryons

Electromagnetic form factors and electromagnetic properties of $Δ$ baryons are studied in the perturbative chiral quark model (PCQM) using quark wavefunctions predetermined from proton charge form factor data. Our predictions agree reasonably with lattice-QCD and other phenomenological results. For charged $Δ^{++}$, $Δ^{+}$, and $Δ^{-}$, observables are dominated by the three-quark core, with small meson-cloud corrections superimposed. For neutral $Δ^{0}$, the leading-order quark core contribution vanishes, so meson-cloud corrections fully determine its electromagnetic response. The PCQM gives a nonzero magnetic moment, a negative charge radius, and a sizable magnetic radius for $Δ^{0}$. This work provides consistent, parameter-free predictions for $Δ$ baryon structure and highlights the role of meson cloud fluctuations.

hep-ph↗

arXiv2026

Triage: Skip, Skim, or Refine Attention Tiles in Video Diffusion Transformers

Attention over tens of thousands of visual tokens dominates the inference of video diffusion transformers (DiTs). Existing methods reduce this cost through sparsity, low precision, or both, but use a single precision for all kept tiles. We observe that the output error introduced by approximation varies widely across tiles. Many tiles can be skipped with little error, others must be computed but gain little beyond low precision, and a few retain substantial error even at low precision. We present Triage, a training-free attention method that assigns each tile one of three approximation tiers: skipping it by replacing it with its block means, skimming it in NVFP4, or refining it with NVFP4 residual correction. A lightweight router predicts each tile's error under each tier from block statistics and moves a tile to a more accurate tier only when the predicted error reduction justifies the extra compute, with a single price setting this trade-off. A fused kernel executes all three tiers in a single pass on the FP4 tensor cores of RTX Blackwell GPUs. Across Wan2.2, HunyuanVideo, and LTX-2.5, Triage accelerates end-to-end generation by up to 2.32x over dense FlashAttention, and in each of five settings it runs faster than every baseline while reaching 0.3-3.0 dB higher PSNR than the highest-PSNR baseline.

cs.LG↗

arXiv2026

Curvature quotient flows and Alexandrov--Fenchel inequalities in hyperbolic space

We study quermassintegral-preserving curvature flows in $\mathbb H^{n+1}$, $n\geq3$, driven by the curvature quotient $(E_k/E_\ell)^{1/(k-\ell)}$. For $1\leq\ell<k$ with $k+\ell\leq n$, we prove that positive sectional curvature is preserved. Starting from a smooth closed strictly convex hypersurface with positive sectional curvature, the flow exists for all time and converges smoothly and exponentially to a geodesic sphere. As an application, we establish sharp Alexandrov--Fenchel inequalities between $W_k$ and $W_\ell$ for smooth bounded convex domains whose boundaries have nonnegative sectional curvature, whenever $0\leq\ell<k\leq n$ and $k+\ell\leq n$. Equality holds if and only if the domain is a geodesic ball. Together with previously known cases, these inequalities give the full family of Alexandrov--Fenchel inequalities for this class of domains in $\mathbb H^6$. We also construct counterexamples showing that $k+\ell\leq n$ is the optimal range for preservation of positive sectional curvature by these flows.

math.DG↗

arXiv2026

Positivity for a $t$-analogue of refined Littlewood-Richardson rules

In 2010, Haglund, Luoto, Mason and van Willigenburg proved that the product of a Schur function $s_λ$ and a Demazure atom $A_γ$ expands positively in terms of Demazure atoms, and interpreted the coefficients as the number of Littlewood-Richardson skew skyline tableaux of a given shape and a filling type. Extending this approach, they further established the positive expansion for the product of a Schur function $s_λ$ and a quasisymmetric Schur function $S_α$, or of a Schur function $s_λ$ and a Demazure character $κ_γ$. In particular, the latter provides a refinement of the classical Littlewood-Richardson rule. We present a natural $t$-analogue of these positive expansions by deriving the expansion for a product of a Hall-Littlewood polynomial $P_λ(t)$ and a $t$-Demazure atom $E_γ(t)$ in terms of $t$-Demazure atoms. The corresponding coefficients are $\mathbb{N}\,[t,1-t]$-positive; that is, they are polynomials with nonnegative integer coefficients in $t$ and $1-t$, and they are interpreted in terms of standard Young tableaux. We take the approach of connecting the multiplication coefficients with the LLT polynomials, which differs from the combinatorial insertion arguments of Haglund, Luoto, Mason and van Willigenburg.

math.CO↗

arXiv2026

Grid-Forming AC/DC Converters for Meshed DC Networks via Dual-Measurement Frequency Droop

This paper proposes and analyzes a dual-measurement frequency-droop (DMFD) controller for meshed MVDC networks. Using only two local DC measurements, bus-voltage deviation and a filtered DC-port current or power deviation, the controller enables grid-forming operation at both AC and DC ports across the mesh. Although the DMFD regulates neither the DC voltage nor the DC current directly and acts only on the terminal frequency, it is shown that AC-side synchronization closes an effective DC droop at every synchronized terminal, so that the steady operating point of the mesh exists and is unique. Constant frequency deviations at a terminal and single-terminal islanding enter the same DC equations as equivalent current offsets, with the island frequency fixed by the island power balance, allowing node voltages and branch currents to be obtained from one linear system. The steady-state model yields design-use envelopes to map the droop ratios and the net DC-injection levels into bounds on DC-voltage deviation and terminal currents. A small-signal dynamic model is developed and used to generate eigenvalue loci and damping-ratio charts that guide controller-gain selection. Averaged EMT simulations on a four-terminal ring corroborate the analysis: in regimes where PLL-based grid-following control is known to lose damping, the DMFD controller maintains stable operation and follows the predicted steady-state characteristics.

eess.SY↗

arXiv2026

Detecting bus delays and service anomalies in Niterói with autoencoder latent feature representations

Bus riders judge service by two failures: single trips that run late, and days on which the whole system misbehaves. We ask whether one unsupervised representation of raw GPS telemetry can serve both diagnoses on the NetMob 2026 Niterói bus dataset. An autoencoder compresses each trip into a 64-dimensional latent vector; the frozen encoder feeds supervised delay classifiers, and the reconstruction error serves as an anomaly score. Two findings are negative. First, the latent representation does not improve delay detection: gradient boosting on per-trip summary features reaches $0.846 \pm 0.013$ ROC-AUC against $0.838 \pm 0.016$ for the best latent setup, and the gap becomes significant once the preprocessing filters are relaxed. Second, two of the three external checks of the anomaly score were circular. Retraining without rainfall removes the rain association ($5.4\%$ against $5.0\%$ flagged, $p = 0.84$) and an apparent citywide disruption on 23 March. What survives is a day-level signal stable across three autoencoders (Spearman $ρ\geq 0.96$), driven mostly by day type, with degraded-telemetry days still separable within weekday and weekend strata. The alarm threshold limits detection more than the score does: a pooled percentile cut flags $5.7\%$ of trips on the worst degraded day, while re-scaling within each bus line flags $29.3\%$. The delay model rests almost entirely on one feature, no latent dimension is individually interpretable, and 22 chronic line-directions carry half of all delayed trips.

cs.LG↗