Search arXiv⌕ Search

SEARCH · Search arXiv

Search Search arXiv

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

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 1,387 records · Page 77Linked to original sources
arXiv2026

The Bayesian Learning Rule Beyond KL Geometry

The Bayesian learning rule (BLR) exploits the relation between Bayesian updating and the Kullback-Leibler divergence (KLD) to derive efficient additive updates for exponential family variational posteriors. The KL geometry is powerful but restrictive: combined with posterior families with parameter-dependent support it can lead to undefined objectives. We ask which non-KL geometries preserve the additive BLR computation. Previous works that use $f$-divergences suggest that this is only possible with the KLD. Here, we show that this is not true. We show that additive information-geometric optimization algorithms can be derived for the class of mixed-space Bregman or Fenchel-Young discrepancies. These discrepancies induce generalized maximum-entropy families, dual coordinates, and generalized-BLR canonical gradient updates, recovering the classical BLR. The generalized BLR leads to new algorithms that \textit{cannot} be derived from the KL-based BLR. For the Tsallis negentropy with quadratic statistics, the generalized BLR yields bounded-support and heavy-tailed $q$-Gaussian posteriors and an easy Newton-like implementation. Experiments up to CIFAR-100 illustrate controllable support and tail behavior and competitive performance.

cs.LG↗
arXiv2026

Extreme Compression of Per-Layer Embeddings via Cross-Layer Latent Vector Quantization

Scaling embedding capacity alongside Transformer computation is an emerging trend in language models. This trend makes high compression rates increasingly useful. Hence, we study extreme compression of Per-Layer Embeddings (PLE), which follow this trend and account for 35.2% of the parameters of Gemma 4 E4B. At such rates, conventional quantizers give each token-layer vector its own code of only a few bytes. We instead develop OPTIC (Output-Path-fitted, Token-shared Indexed Codes), which stores one quantized code tuple per token and decodes it at every layer with a layer-specific affine decoder. Codes and decoders are fitted to the PLE output path rather than to the stored vectors. At about $100\times$ PLE compression of bfloat16 (BF16) E4B, OPTIC retains an aggregate 97.8% of the uncompressed model's performance across seven benchmarks, whereas product quantization, the strongest of six conventional baselines, retains 66.3%. At $594\times$ and $1,002\times$ PLE compression of BF16 E4B, OPTIC retains 90.5% and 84.1%. On the official E4B release with 4-bit linear weights and 16-bit activations (W4A16), OPTIC retains 96.9% at about $100\times$ PLE compression. During inference at about $100\times$ PLE compression, BF16 E4B uses 29.7% less GPU memory than with uncompressed PLE at the same key-value cache size, while running at most 0.40% slower.

cs.LG↗
arXiv2026

Identifiability from Stationary Covariance in Linear SDEs with Reciprocal Interactions and Latent Variables

Continuous-time stochastic differential equations (SDEs) provide a natural framework for modeling interacting dynamical systems. A central causal question is whether causal effects among observed variables can be identified from observational data drawn from the stationary distribution, given the underlying causal graph. For linear SDEs, existing identifiability results are largely restricted to fully observed systems whose causal graphs contain no directed 2-cycles. We first establish new graphical conditions for the generic identifiability of causal effects in systems that may contain directed 2-cycles, thereby allowing for reciprocal interactions. Next, we address the previously unexplored setting of latent variables by developing an instrumental-variable framework for identifying a target direct causal effect under partial observability.

cs.LG↗
arXiv2026

The Delayed Countdown Method for Validating Review-Oriented AI Workflows in eDiscovery

Review-oriented workflows have become the dominant approach to AI-assisted document review in electronic discovery (eDiscovery). How to validate the effectiveness of review-oriented workflows has, however, remained a confusing and controversial topic in technology-assisted review (TAR) in the law. We examine the sources of this controversy, and propose a simple and practical evaluation approach for review-oriented workflows. Our Delayed Countdown method is easily incorporated into existing review-oriented workflows and software, is compatible with arbitrary stopping rules, and produces valid frequentist confidence intervals on effectiveness. We explain the conceptual foundations of the method, provide example computations, discuss the handling of logically impossible recall values, and suggest future work on decreasing review costs.

cs.IR↗
arXiv2026

Lepton number and flavor violation mediated by doubly charged scalars at $μ$TRISTAN $μ^+ μ^+$ collider

Charged lepton flavor violation is an important probe of physics beyond the Standard Model. Although $ΔL_μ= -ΔL_e = \pm 1$ lepton flavor violating (LFV) processes, such as $μ\to eγ$ and $μ\to 3e$, are severely constrained experimentally, there is still plenty of room for LFV processes involving lepton number violation (LNV). In this paper, we investigate such processes at a $μ^+μ^+$ collider that remain even if processes with $ΔL_μ= -ΔL_e = \pm 1$ are forbidden by discrete symmetries. We consider processes induced by doubly charged scalars, which can be either an $SU(2)_L$ singlet or a component of an $SU(2)_L$ triplet. The cross sections of these processes at the collider depend on the representation of the scalar unless the muon beam is unpolarized. In the case of $ΔL_μ= -ΔL_e = 2$, we discuss the combination of the muonium-to-antimuonium transition and the $μ^+μ^+ \to e^+e^+γ/Z$ processes at the collider, and how the model parameters, such as the scalar couplings and mass, can be determined from experimental measurements. Doubly charged scalars are often utilized in models for generating neutrino masses, and they can also be used to construct models with LNV processes at a muon collider. If neutrino masses are generated at tree level, the LNV processes at the collider are suppressed, and therefore, we consider models that do not induce neutrino masses at tree level. LNV processes, such as $μ^+μ^+ \to t\bar b t\bar b$ and $μ^+μ^+ \to μ^+ e^- t\bar b t\bar b$, are possible in these models. Even if LNV is indicated by searches at hadron colliders, it is worthwhile to pursue these processes as clean LNV signals at a lepton collider.

hep-ph↗
arXiv2026

A New Twist on Astigmatic Transformation: Beyond Laguerre-Gaussian Beams

Astigmatic transformation is widely used to convert the azimuthal phase structure of Laguerre-Gaussian vortex beams into intensity patterns from which the topological charge can be read out. Here, we show that this phase-to-intensity conversion is not just restricted to the Laguerre-Gaussian radial profile. We derive analytical expressions for the astigmatically transformed high-order Bessel--Gauss and perfect optical vortex fields, and find that, despite their substantially different radial structures, the number of intensity minima is directly set by the accumulated azimuthal phase winding. Numerical simulations and experimental measurements on both beam families confirm this result, establishing astigmatic transformation as a direct probe of the topological charge of scalar structured optical fields. We further extend the approach to non-separable vector Bessel-Gauss beams, in which the spatial and polarization degrees of freedom are coupled, and show that polarization-resolved detection allows the topological charge of the constituent spatial modes to be determined independently.

physics.optics↗
arXiv2026

When LLMs Forget What They Believe: Characterizing State Integrity Collapse in Long-Horizon Dialogues

Large language models can repeat an instruction while failing to keep obeying it. We study this recall-compliance dissociation in long-horizon dialogue and define State Integrity Collapse: a model preserves a stated constraint in memory but stops applying it to later turns. We introduce StateTrack-Bench, a controlled suite of conversations covering numerical, format, role/persona, factual-memory, and logical-rule constraints. Across seven provider-pinned model snapshots and five dialogue-depth probes, strict joint retention ranges from 62.4% to 100.0%, with degradation concentrated in symbolic constraints and in one especially brittle model. A forked recall audit shows that 88.6% of observed failures are behavioral-drift failures rather than ordinary forgetting; excluding the brittle outlier preserves this diagnosis (84.4% of failures). Extended evaluations at depths up to 50 turns confirm that the dissociation persists beyond the core benchmark, while Structured State Anchoring recovers up to 24.4 points on immutable constraints. The results show that reliability cannot be inferred from single-turn instruction following alone; evaluation must track whether remembered constraints remain operational as dialogue state grows. Code and artifacts are available at: https://github.com/xzzvsxd/StateTrack.

cs.CL↗
arXiv2026

The Value of System 1 Models in Medicine

Many clinical decisions are a choice among a few fixed options, taken quickly and in large numbers. Such decisions also need an uncertainty estimate to know when to defer to a clinician. Generative language models answer slowly in free text and state their confidence poorly. In contrast, System 1 models answer zero-shot with a probability for each answer. However, their performance across clinical tasks is unknown. Four System 1 models, two proprietary and two with open weights, were evaluated zero-shot on 86 public tasks from 14 clinical domains in 20 languages, spanning examination questions and medical imaging. Tasks were assigned to six capability tiers before analysis. The primary outcome was the area under the receiver-operating-characteristic curve (AUROC). Secondary outcomes included calibration, language effects and comparisons with published results and generative models. Across 179,111 cases, the mean AUROC of the best model per task was 0.94 for medical knowledge, 0.93 for text understanding, 0.83 for clinical judgement from text, 0.89 for image perception, 0.84 for 2D image findings and 0.69 for volumetric findings. Imaging tasks scored 0.09 (95% CI, 0.06 to 0.11) lower than text tasks (P < 0.001). The best model matched or exceeded the best published result on 13 of 26 text tasks and on 8 of 35 imaging tasks. Triage nurses outperformed the best model (accuracy, 0.85 vs. 0.56). One decision took 0.34 to 0.44 seconds and 24 to 81 joules on open-weight models run locally. Accuracy on identical questions fell by up to 39 percentage points from English to Yoruba. Generative models of the same family were 11 to 76 times slower, with 3 to 5 percentage points higher balanced accuracy. System 1 models decide quickly, cheaply and with usable confidence when the answer is stated in a text or recalled from medical knowledge, but not when it must be judged or read from an image.

cs.LG↗
arXiv2026

Charmed baryon-to-meson ratios inside jets as a discriminator of hadronization mechanisms: from color reconnection in p+p to coalescence in Pb+Pb

Jets separate the products of hard scatterings from the underlying event and thus help to clarify the origin of the enhanced baryon-to-meson ratios observed from small to large collision systems. We investigate the charmed baryon-to-meson ratio $Λ_c^+/D^0$ within jets in p+p and Pb+Pb collisions at $\sqrt{s_{NN}}=5.02$~TeV to discriminate between candidate hadronization mechanisms. In p+p collisions, simulated with PYTHIA8 using color reconnection (CR-BLC Mode 2) as a representative beyond-leading-color scenario, the in-jet ratio is enhanced at low-to-intermediate $p_T$ relative to the inclusive one---opposite to the light-flavor $Λ/K^0_S$ case---because the color-reconnection-induced excess of the ratio remains correlated with the jet and survives the jet requirement, while the fragmentation baseline shows no such enhancement. In Pb+Pb collisions, described by the SHELL model with a coalescence plus fragmentation hybrid hadronization, the in-jet ratio is instead suppressed moderately at intermediate $p_T$, because coalescence hadrons assembled from soft thermal partons carry no correlation with the jet direction and are preferentially discarded by the jet selection. The sign of the in-jet modification thus discriminates between the two mechanisms. Both mechanisms shift charmed hadrons toward small values of the parallel jet momentum fraction $z^{ch}_{||}$, with a stronger effect on $Λ_c^+$, so that the $z^{ch}_{||}$ spectrum probes non-fragmentation hadronization independently of parton energy loss. The radial profile of the in-jet ratio rises with the distance from the jet axis in both systems, with the lowest-$p_T$ interval displaced farthest for color reconnection and the intermediate-$p_T$ interval for coalescence. These results provide the baseline and predictions needed for future jet-tagged charmed-hadron measurements in both collision systems.

hep-ph↗
arXiv2026

SurveyGraph: Generating Academic Surveys via Methods-Entity Knowledge Graphs

Most automated survey generation systems retrieve papers as flat documents, limiting their ability to capture research structure. We present SurveyGraph, a framework built around a Methods-Entity Knowledge Graph (MEKG) encoding method-level follow-on, competition, and shared-resource relations. The graph supports core-paper discovery, retrieval, hybrid outlining, and evidence-gated narrative synthesis. We evaluate retrieval, semantic edges, citation grounding, collection coverage, and generation-stage components. In a controlled seven-method equal-budget benchmark, MEKG-anchored fusion balances recall and structural density, with an independent co-citation check. In a separate archived graph-native diagnostic, graph-aware retrieval nearly doubles recall over flat lexical baselines across eight runs (p = 0.004). A semantic evidence gate achieves AUC = 0.75. Audits of 160 generated sentences find 80.6% anchored-or-partial citation support. Human adjudication of 108 stratified semantic edges finds 68.5% supported-or-partial relations (binary-support Cohen's kappa = 0.64), with improves strongest (80.6%) and extends weakest (52.8%). A SurveyBench-facing collection probe reaches mean reference coverage of 0.521 across four topics with an eight-paper budget. Same-backend Qwen diagnostics on a 30-paper/74-relation graph show reduced organization signals when hybrid outlining is removed. Concurrent-system comparisons provide bounded context rather than matched-backend end-to-end evidence. The results position MEKG as an auditable structural layer for typed diagnostics, relation-specific filtering, and controlled survey generation. Code and artifacts: https://github.com/xzzvsxd/SurveyGraph.

cs.CL↗
arXiv2026

Transfer Meta-Learning for State Space Autoencoders

Nonlinear system identification often requires costly experimental campaigns and separate model fitting for each different instance (target system) of a given system family. To reduce this burden, meta-learning approaches reuse information acquired across related systems. In particular, in-context meta-learning enables zero-shot prediction from a brief target system context (i.e., a short input-output trajectory), but generally yields large black-box sequence predictors without exposing an explicit state space model. Conversely, compact structured state space models are suitable for simulation, estimation, and control, but are usually fitted or adapted separately for each target system. To bridge this gap, we introduce a transfer meta-learning extension of state space autoencoders. The method reuses a pretrained encoder, whose weights remain fixed, to extract a compact representation of the target system from its context. An affine parameterization maps this representation to the parameters of a compact neural state space model without target-specific adaptation; the resulting model contains orders of magnitude fewer parameters than the reference in-context predictors from the literature. We evaluate zero-shot prediction and simulation performance on held-out system instances from Buck-Boost converter and Bouc-Wen oscillator system families, comparing performance against individually fitted baseline models.

eess.SY↗
arXiv2026

Atomic limits of empirical measures of interacting particle systems with self-duality

We study the fixed-particle limit of empirical measures for a unified self-dual family of conservative particle systems on $\mathbb{Z}^d$: independent random walkers, the symmetric exclusion process, and the symmetric inclusion process. For a fixed number of particles $k$, under diffusive space-time scaling parametrised by $n$, we prove that as $n\to\infty$ the normalised empirical measure converges in $D([0,T],\mathcal{M}_1(\mathbb{R}^d))$ to an atomic empirical measure carried by $k$ independent diffusions, with constant covariance determined by the microscopic jump kernel. Equivalently, the limit is the unique atomic solution of the Dean-Kawasaki martingale problem with mass quantisation $1/k$. In particular, exclusion and inclusion have the same fixed-$k$ atomic limit as independent walkers: the interaction contributes only through microscopic near-collisions, whose rescaled occupation time vanishes. The proof combines self-duality, consistency-based reduction to two particles, and Green-function estimates for finite perturbations of symmetric random walks.

math.PR↗
arXiv2026

Stream4D: Streaming 4D Reconstruction with Training-free Motion Decoupling via Causal KV-Cache Attention

4D scene reconstruction requires robust handling of dynamic objects, yet existing streaming methods commonly assume static scenes, leading to pose drift and ghosting artifacts over long sequences. We introduce Stream4D, a training-free, motion-decoupled streaming reconstruction framework that can be directly integrated into causal Transformer backbones (e.g., StreamVGGT, Stream3R). Our key observation is that causal attention implicitly encodes rich temporal motion cues within the KV cache. We adaptively aggregate these cues from multiple attention layers and across temporal windows, and fuse them with depth priors to handle challenging cases such as occlusions and textureless regions, producing precise dynamic masks that decouple static and dynamic regions. We then apply a soft key masking scheme that leverages these dynamic masks to suppress interference from dynamic regions in causal attention, thereby improving camera pose estimation accuracy. Stream4D inherits the efficiency of causal attention and remains compatible with existing streaming architectures. On diverse dynamic-scene benchmarks, Stream4D significantly improves camera pose estimation and static-scene reconstruction quality, yields reliable dynamic masks, and effectively mitigates long-sequence ghosting.

cs.CV↗
arXiv2026

MVLoMa: Multi-View Local Feature Matching with Any Geometry Foundation Model

Local feature matching is typically formulated for image pairs, although structure-from-motion and visual localization commonly match one reference image against many retrieved or covisible neighbours. Applying a pairwise matcher independently to these centre-neighbour edges duplicates reference-side computation and prevents the matcher from exploiting complementary multi-view geometric evidence from the remaining observations. To address these limitations, we introduce MVLoMa, a sparse star-graph matcher that extends LoMa from image pairs to multiple views. MVLoMa first uses a frozen multi-view geometry model to jointly describe all images in a star, then injects the resulting cross-view geometric context into local descriptors. It then matches the centre image to all neighbours in a single forward pass through per-view self-attention and bidirectional centre-neighbour cross-attention. For supervision, we derive positive correspondences only from depth-valid, mutually visible dense warps and explicitly penalize high-confidence matches that violate epipolar geometry. On 25-view star graphs from HyperSim, ScanNet++, Mip-NeRF 360, and Tanks and Temples, MVLoMa obtains the best macro-average relative-pose AUC among the compared methods at all three thresholds (58.65/85.41/90.99). Its matcher is 2.20 times faster than sequential evaluation of the 24 centre-neighbour edges. The same design supports four tested frozen geometry encoders -- DA3, $π^3$, VGGT, and VGGT-$Ω$ -- with DA3-Giant performing best on all reported encoder-ablation metrics. Controlled ablations further isolate the effects of geometry-aware descriptors, visibility filtering, geometric negative supervision, and joint multi-view inference.

cs.CV↗
arXiv2026

Hyperreduction accelerates finite element model updating for the discovery of neural materials

Finite element model updating is among the most general approaches to discover constitutive models from structural experiments: since the displacement field is computed rather than measured, it accommodates arbitrary geometries, boundary conditions, and sparse or noisy data. However, every loss evaluation requires a nonlinear finite element solve per load step, and training neural materials requires many such evaluations. Intrusive model order reduction lowers this cost while retaining boundary conditions and variational structure. How can a reduced model remain accurate while the network weights change at every training step? We combine proper orthogonal decomposition and energy-conserving sampling and weighting with adjoint sensitivities of the reduced model, and adapt both the trial space and the sampled elements along the optimization path. The constitutive law is a polyconvex neural network: a logarithmic normalization term, together with a zero-anchored input convex architecture and a growth term, renders the energy polyconvex, coercive, and non-negative by construction, independent of the trained weights. We demonstrate the approach on two experimental datasets and one synthetic three-dimensional example. For a two-phase specimen, training time decreases from nine hours to 41 minutes compared with a full-order identification, without loss of accuracy. In the three-dimensional example, six modes replace 48 672 degrees of freedom, 116 of 15 625 elements suffice to evaluate the internal forces, and training takes 13 minutes. In all cases, the identified material reproduces the measured reaction force and predicts the displacement field from the force alone. Hyperreduction makes the discovery of neural materials from structural data affordable while retaining the advantages of finite element model updating.

cs.CE↗
arXiv2026

Bridging the Data Gap: Source-Informed Task-Space Transport for Few-Shot Diffusion

Many deployments require a model to move rapidly from a data-rich source to a related target observed only sparsely. This is especially consequential where target data are difficult to collect, as in healthcare settings with persistent population or acquisition imbalances. Retraining or fine-tuning a high-capacity generator for every target is costly and can be statistically unreliable when target data are scarce. We introduce Meta-Diff, which instead learns a recurring source--target relation once in a compact task space. A shared encoder maps each distribution to a low-dimensional coordinate; a learned relation transports a new source coordinate to its predicted target, and a shared diffusion backbone generates from that prediction without target-specific retraining. Controlled Gaussian-mixture experiments identify when this works: task identity must survive the representation, the relation must generalise across tasks, and the source--target shift must leave useful headroom for transport. On held-out CIFAR-100 validation tasks, correct source $+$ transport achieves 13.9% lower mean SW than transport from the wrong source. With validation-selected adaptation settings in the final-test one-target comparison, its mean SW is 14.7% lower than target-only modelling and 32.7% lower than full fine-tuning; the main route itself uses no target image and assumes the applicable relation is known. On matched CIFAR test episodes, generating from the clean-source coordinate and then applying the known corruption yields lower SW than task-space transport. Office-Home exposes a complementary boundary: transport can align target-domain statistics without reliable object identity. Together, these results recast extreme-few-shot adaptation as reusable transport in task space, while making the conditions for its success and failure directly measurable.

cs.LG↗
arXiv2026

Multigroup Radiation Diffusion with Compton Scattering on a Moving Mesh: Implementation in RICH and Application to Tidal Disruption Events

Compton scattering exchanges energy between radiation and hot gas and changes the spectrum in high-energy astrophysical systems such as tidal disruption events (TDEs). Because it couples all photon energies and can act faster than the hydrodynamic time step, many radiation-hydrodynamics (RHD) simulations omit its spectral redistribution or treat its energy exchange in the gray approximation. We extend the multigroup flux-limited diffusion module of the moving-mesh RHD code RICH to include Compton scattering with the exact thermal Compton kernel. In our semi-implicit scheme for the thermal radiative transfer equation, the temperature dependence of the kernel enters a generalized Fleck factor, and the group Compton operator conserves photon number and satisfies detailed balance. In equilibration tests, the scheme agrees with independent solutions, relaxes without absorption and emission to a Bose-Einstein spectrum, and remains stable for time steps that do not resolve the Compton equilibration timescale. We apply it to a three-dimensional simulation of the disruption of a $0.5\,M_\odot$ star by a $10^6\,M_\odot$ black hole. We find that during the X-ray flash, inverse Compton scattering dominates the coupling between the gas and the radiation near the nozzle shock, where it adds X-ray energy much faster than thermal emission and may affect the dynamics. Outside the X-ray thermalization radii, its instantaneous X-ray production rate during the flash is comparable to or larger than the escaping X-ray luminosity, so it may contribute substantially to the X-ray emission from TDEs.

astro-ph.HE↗
arXiv2026

Consistent but Biased: Chair Personas Trade Noise for Bias in Multi-Agent Clinical Decisions

In clinical systems built from several large language model (LLM) agents, one agent often chairs the discussion and makes the final decision. Whether the chair's persona prompt changes that decision, or the agreement between repeated runs that some systems treat as reliability, has not been tested with the case held constant. We ask whether the persona changes the decision or only its timing, how it affects agreement between runs, and which behaviors appear when the prompt describes only personality traits. A simulated department of seven agents discussed 20 published cases of degenerative cervical myelopathy, with the case reports replayed unchanged in every run. Only the prompts of the chair and its deputy changed (no persona, or a consultative, autocratic or avoidant one), written either as rules of conduct or as personality traits. With personas written as rules, the chair chose to operate now in 93% of runs when autocratic and in 2% when avoidant, against 57% without a persona. Surgery was still planned in nearly all decisions, so the personas mainly changed whether to operate now or after more tests. These effects held with a second model and reworded prompts. The autocratic and avoidant personas made the three runs of a case agree more often, but only about as often as chance predicts, and Fleiss $κ$ did not rise. Prompts that described only personality traits kept more than half of the autocratic and avoidant effects on operating now. They reproduced quick decisions, delays and extra tests, but not ignoring red flags, dropping backup plans or passing the decision to the family. Studies should report and vary the chair's persona as they do the model, and correct agreement between runs for chance before calling it reliability.

cs.MA↗