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 91 records · Page 5Linked to original sources

arXiv2026

Two-parameter continuous deformation of Starobinsky inflation as a bridge between Planck and ACT DESI data with $N_\star\in(50,60)$

We present a family of plateau-type inflationary potentials, eq.~\eqref{Vgeneral}, and analyze a two-parameter $\alpha\beta$-Starobinsky specialization that interpolates continuously between a \emph{maximal} plateau ($V\!\to\!V_0$) and a \emph{submaximal} plateau ($V\!\to\!V_\infty 0$ with $x_\star\gg 1/\beta$ the slow-roll scaling laws change to $n_s\simeq 1-\frac{4}{3N_\star},\, r\simeq\mathcal{C}(\alpha,\beta)\,N_\star^{-4/3},$ with an explicit coefficient $\mathcal{C}(\alpha,\beta)$ set by the plateau truncation. This deformation lifts $n_s$ at fixed $N_\star$ while further suppressing $r$, reconciling the Planck~2018 constraint $n_s=0.9649\pm0.0042$ (68\% CL) and BICEP/Keck18 data $r_{0.05}<0.036$ (95\% CL), with the higher central values $n_s\sim0.97$--$0.98$ preferred by ACT+DESI~DR2 (BAO), within the theoretically motivated interval $N_\star\in(50,60)$ and without exotic reheating. We provide an exact identity for $V/V'$ enabling analytic control of $N_\star$, a practical crossover criterion $\beta\,x_\star\ll1$ vs.\ $\gg1$, and a transparent mapping between $(\alpha,\beta)$ and the observables $(n_s,r,N_\star)$. These yield sharp, testable signatures, particularly the softened $N_\star$-scaling of $r$, that distinguish a maximal from a submaximal plateau with upcoming CMB and LSS data.

astro-ph.CO↗

arXiv2026

Demystifying the Privacy-Utility Trade-off in LLM Interactions

The integration of Large Language Models into daily tasks relies on context-rich instructions, inevitably exposing sensitive user information. Current privacy-preserving methods typically employ context-agnostic static rules, causing severe utility degradation. However, the specific mechanisms governing how sanitization impacts downstream performance remain largely underexplored. To address this, we conduct a systematic analysis to deconstruct the privacy-utility trade-off, uncovering three underlying mechanisms: (1) Context-Dependent Utility, which first establishes when to sanitize by revealing that data value shifts from critical constraints to dispensable noise based on user intent; (2) Strategic Adaptation, which subsequently determines how to sanitize by dictating that the choice between removal and replacement depends on the task's reliance on factual integrity versus structural coherence; and (3) Combinatorial Interplay, which finally extends the protection scope by demonstrating that attributes form a semantic web of synergistic dependencies or antagonistic redundancies. Guided by these insights, we introduce an intent-driven local protection framework. By distilling a lightweight model Veilmind-4B to drive a dynamic extraction-sanitization-restoration pipeline, our approach reaches a low-leakage privacy point while preserving substantially higher response utility than existing privacy-oriented baselines, advancing the privacy-utility trade-off toward the Pareto frontier.

cs.AI↗

arXiv2026

Distribution-aware Language Neuron Identification in Multilingual Large Language Models

Multilingual large language models (mLLMs) contain a small fraction of feed-forward neurons that are sensitive to particular languages, commonly termed language-specific neurons. Existing work measures language specificity using the entropy of each neuron's language-wise probabilities of being active, where a neuron is considered active when its activation value is positive. However, this approach may not fully capture the multilingual nature of mLLMs, where language representations are distributional and mutually related. We propose Distribution-aware Language Neuron selection, which leverages pairwise relationships between per-language activation distributions over the full activation range, including negative values. Specifically, we quantify each neuron's language specificity by clustering languages using pairwise overlap coefficients between their activation distributions. Across two mLLMs and two held-out corpora, our identifier more effectively isolates language-specific causal effects, yielding up to 4.9$\times$ higher on-target language damage per neuron while preserving off-target language performance.

cs.CL↗

arXiv2026

Importance Weighting for Unlabeled-unlabeled Learning under Distribution Shift

Unlabeled-unlabeled (UU) learning allows us to learn a binary classifier from two sets of unlabeled data with different class-priors. It is a general framework because it includes a wide variety of supervised learning such as positive-unlabeled (PU) learning, noisy label learning, and similarity-based learning. Existing UU learning assumes that the test and training distributions have the same class-conditional densities. However, this assumption rarely holds in practice due to distribution shifts. This paper proposes a distribution shift adaptation method for UU learning that uses UU data in the training distribution and a few UU data in the test distribution. The proposed method is based on the importance weighting, which minimizes the test risk by using training data with estimated importance weights. Although existing importance weighting methods cannot handle UU data, we show that it can be done in a principled manner. Thanks to the generality of UU learning, our method can handle various learning problems such as PU and noisy label learning under distribution shift within a single framework while existing methods are usually tailored to a specific problem. Moreover, it does not require any assumption of the shift types such as covariate shift. We experimentally demonstrate the effectiveness of the proposed method with real-world datasets.

cs.LG↗

arXiv2026

U-Shaped Fermi-Level Dependence of Point-Defect Aggregation Enthalpy in Silicon

Point defects govern the electronic and structural behavior of silicon, yet their configuration and energy depend on both the charge state and the Fermi level (EF). Using an unbiased global structure search with first-principles calculations, we screened more than 1,500 candidate structures, covering intrinsic defects and C, H, O, N, P, and B impurities in charge states from -2 to +2. The reaction enthalpy of defect aggregation depends on EF in a U shape, being strongest where the net charge transferred during the reaction vanishes and weaker toward both band edges. This U arises because complexes are more charge-neutral than their isolated constituents, so the net charge transfer changes sign across the gap. The point where this sign change occurs is set by the transition levels of the specific defects. For typical reactions, the driving force is tunable by 0.3-0.5 eV, which shifts the equilibrium complex concentration by five to eight orders of magnitude at room temperature. Our results make the Fermi level a practical lever for defect engineering in as-grown and irradiated silicon.

cond-mat.mtrl-sci↗

arXiv2026

Rethinking Verbalized Confidence for LLM-as-a-Judge: A Compatibility Shift on Post-2025 Proprietary Models

Verbalized confidence, long dismissed as overconfident, coarse, and prone to round-number clustering, is now the more robust soft-scoring mechanism for LLM-as-a-Judge on top-tier proprietary models. Across SummEval, AggreFact, and HelpSteer2, spanning up to 18 LLMs, we show that the standard advice to prefer log-probabilities no longer holds on post-2025 models, where verbalized confidence is the better signal. We call this a compatibility shift. On top of a standard verbalized-confidence baseline, we introduce two new ingredients: an overconfidence advisory and self-debate. Together they improve calibration, score-distribution spread, and robustness to task subjectivity. We further observe a generation effect: post-2025 models accommodate these two additions with little balanced-accuracy cost, whereas pre-2025 models pay a measurable penalty. Compared with logprob-based G-Eval, verbalized confidence is the more subjectivity-robust soft signal on GPT-family top-tier releases. The shift is invisible under accuracy-only reporting. Rather than defaulting to hard predictions, we recommend broader use of soft scoring in LLM-as-a-Judge. More broadly, verbalized confidence has moved from a weaker substitute for logprobs to a practical soft-scoring mechanism for contemporary LLM judges.

cs.CL↗

arXiv2026

Structured Stochastic Representations of Integrated Dynamic Strategies

Dynamic allocation decisions couple present resource use to evolving internal conditions, delayed returns, and future costs. We represent this interaction by four probability localizations linked through regime-indexed, graph-constrained column-stochastic operators. Pre-action state or context selects a locally affine model, while action-dependent changes update subsequent regimes, yielding a causal switched representation of nonlinear evolution. We characterize operator identifiability relative to the graph, the stochastic constraints, and the sampled embedding, separating coefficient recovery from predictive equivalence on the decision domain. Decision making is then formulated through implementable return--cost acceptability regions. Finite-horizon error propagation supplies conservative classification margins, and simultaneous intervals distinguish model-relative near-optimality from certified $\epsilon$-optimality over a declared finite policy class. Regime-indexed stochastic feedback is admitted when it satisfies the same certification test. Reproducible synthetic laboratories for personal preparation, supplier participation, and customer retention illustrate exact, operator-supplied, and noisy feedback cases. Multinomial experiments show improving recovery of the feedback function and fewer unresolved decisions with increasing sample size, while unrestricted off-policy recovery remains limited. The contribution is a structure-preserving representation--identification--decision workflow, not a domain-specific physiological or commercial calibration.

eess.SY↗

arXiv2026

The $q$-deformed cross-ratio: modular invariants and Coxeter friezes

We introduce and study a scalar $q$-deformation of the cross-ratio on $\mathbb P^1(\mathbb Q)$. Our construction is based on the notion of $q$-deformed rational numbers due to Morier-Genoud and the author. The $q$-cross-ratio is invariant under $\mathrm{PSL}(2,\mathbb{Z})$, while elements of determinant $-1$ of $\mathrm{PGL}(2,\mathbb{Z})$ act by $q\mapsto q^{-1}$. A principal result is its relation to $q$-deformed Coxeter friezes associated with rational polygons. The expansion at $q=e^h$ yields an algebraically independent sequence of modular invariants and relative invariants, although this sequence does not separate modular orbits. We compute the first two nonconstant coefficients of this expansion explicitly.

math.DG↗

arXiv2026

ShellVis: Sandboxed Live Programming for Shell Scripts

Live programming provides visibility to programmers by running and tracing programs as they are edited. However, for programs with potentially harmful side effects, liveness can turn mistakes into disasters. We propose enabling live programming in environments with side effects via sandboxing: confining effects to a simulation of the true environment. We apply sandboxed live programming in the challenging context of shell scripting: a ubiquitous and powerful---yet notoriously opaque and error-prone---tool. ShellVis provides line-by-line feedback on a shell script's run-time behavior, with file operations sandboxed via a safe overlay of the file system. A qualitative user evaluation finds ShellVis to be helpful to participants, replacing tedious existing practices and instilling confidence. Participant responses also reveal areas for future research, particularly bridging the gulf of execution alongside the gulf of evaluation. ShellVis serves as a case study of how sandboxing can bring live-programming techniques into the many real-world programming contexts where side effects are important.

cs.HC↗

arXiv2026

Janus Dipoles: Fundamentals, Realizations, and Emerging Applications

The Janus dipole - featuring orthogonally oriented electric and magnetic dipoles with a 90-degree phase difference - has emerged as a powerful paradigm for wave manipulation. Unlike traditional Huygens dipoles used for directional control, this unique configuration exhibits strongly asymmetric, face-selective near-field behavior while maintaining a quasi-isotropic far-field radiation pattern. These remarkable properties make the Janus dipole an essential platform for directional wave shaping, with wide-ranging applications in on-chip photonics, quantum interactions, and wireless power transfer. This review systematically traces the rapid development of the Janus dipole from its foundational theoretical inception to its diverse implementation platforms across optical, microwave, and acoustic frequencies. In this paper, we explore the governing principles, classify realization strategies into passive Janus dipoles, active Janus dipoles, and advanced near-field coupling control, and highlight emerging frontiers. By bridging foundational electrodynamics with advanced device engineering, this paper serves as an essential reference and roadmap for researchers designing next-generation, highly integrated, and compact wave-manipulation systems.

physics.app-ph↗

arXiv2026

Averaging principles for nonautonomous multiscale stochastic Burgers equations with reflection

In this paper, we study averaging principles for nonautonomous multiscale stochastic Burgers equations with reflection. First, we derive a general averaging principle applicable to such equations under minimal assumptions. Subsequently, since the coefficients of the obtained averaged equation still depend on the small scaling parameter $\e$, we impose either periodic or asymptotic conditions on the coefficients, thereby obtain two distinct averaged equations whose coefficients are independent of $\e$ and establish two averaging principles. Stopping times and Khasminskii's time discretization schemes play an important role. Finally, a concrete example is provided to illustrate the applicability and validity of the theoretical results.

math.PR↗

arXiv2026

The Cartan-Hadamard conjecture in dimension five

We show that the sharp Euclidean isoperimetric inequality holds for domains in complete simply connected Riemannian $5$-manifolds of nonpositive sectional curvature, which establishes the Cartan-Hadamard conjecture in that dimension. The main step is a sharp inequality for constant-mean-curvature hypersurfaces, proved via integrals over pairs of boundary points, in the spirit of Banchoff-Pohl, together with an estimate for Jacobi fields along geodesic chords. The inequality persists for boundaries of isoperimetric regions in geodesic balls, whose mean curvature is constant only on the free part. An isoperimetric-profile argument, after Kleiner, completes the proof. Our method also gives a new proof in dimension $3$.

math.DG↗

arXiv2026

Counting Near-Spanning Matchings in Latin Squares and Steiner Triple Systems

Montgomery recently proved that for sufficiently large $n$, every Latin square of order $n$ has a partial transversal with $n-1$ cells, and every Steiner triple system of order $n$ has a matching with $\lfloor n/3\rfloor-1$ edges, thus confirming the Ryser--Brualdi--Stein conjecture for even $n$ and the conjecture of Brouwer. We prove sharp enumerative refinements of these results: there is an absolute constant $c>0$ such that, for sufficiently large $n$, 1) every Latin square of order $n$ has $ \left((1\pm n^{-c})\frac{n}{\mathrm {e}^2}\right)^n$ partial transversals with $n-1$ cells; 2) every Steiner triple system of order $n$ has $ \left((1\pm n^{-c})\frac{n}{2\mathrm {e}^2}\right)^{\lfloor n/3\rfloor}$ matchings with $\lfloor n/3\rfloor-1$ edges. The first estimate confirms predictions of Montgomery and Kelly.

math.CO↗

arXiv2026

Diarization Error Decomposition Under Pause Annotation Ambiguity

Speaker diarization evaluation is sensitive to ambiguity in pause annotation, which can inflate diarization error rate (DER) or obscure genuine model errors. We show that morphological closing, which has been used for pause-tolerant diarization evaluation, discards segment-level distinctions. Instead, we propose an exact, overlap-aware decomposition of standard DER into a pause-attributable component, consisting of errors compatible with pause filling, and a residual core component that can serve as a proxy for intrinsic diarization errors. The decomposition leaves DER unchanged, while the pause-attributable and core components vary monotonically with the pause threshold and eventually saturate. Experiments spanning synthetic transformations, annotation mismatch, cross-domain evaluation, and tight-boundary diarization show that the decomposition reveals error sources not apparent from standard DER.

eess.AS↗

arXiv2026

DeFiFusion: Combining Transaction Events with Smart Contracts to Detect Price Manipulation Attacks

Decentralized Finance (DeFi) has emerged as a rapidly growing blockchain-based financial service, where market transaction dynamics and underlying smart contract logic are intricately intertwined. This autonomous interplay, while eliminating centralized intermediaries, significantly expands the vulnerability surface of DeFi protocols to Price Manipulation Attacks (PMAs), which have already inflicted catastrophic financial losses. Despite their gravity, existing detection paradigms suffer from fundamental limitations. Transaction-centric methods lack awareness of contract execution semantics, making them prone to false positives under legitimate market volatility, while static contract analyses ignore real transaction behaviors and frequently report vulnerabilities that are infeasible to exploit in practice. We present DeFiFusion, a dual-modal PMA detection framework that closes this gap by jointly modeling transaction events and smart contract semantics within a unified pipeline. Our core insight is that PMA maliciousness emerges only from the interaction between transaction behaviors and the contract logic they exploit; neither signal suffices in isolation. Accordingly, we derive price-manipulation-aware event encoding for extracting fine-grained temporal and economic features tailored to manipulation patterns. We further introduce LLM-based contract semantic extraction to supply the execution-logic context that prior behavioral methods lack. To fuse these modalities, we propose a Dual-Modal Projection-Fusion Transformer with T5-style relative positional encoding, capturing the cyclic multi-stage execution structures that distinguish PMAs from benign market activity. Extensive experiments demonstrate that DeFiFusion consistently achieves state-of-the-art detection performance, effectively recalling 222 of the 225 PMA cases while maintaining a precision of 96.10%.

cs.CR↗

arXiv2026

Bar cohomology of links: beyond Milnor invariants

We develop bar cohomology of link complements as an invariant of links in homology spheres. In this setting, bar cohomology is a Hopf algebra which is calculable using surfaces and their intersection curves in a link complement. In this first in a sequence of works, we introduce the invariant and show that it defines a canonical subspace of the tensor Hopf algebra, which already encodes information about Milnor's link invariants and provides geometrically significant information beyond them.

math.GT↗

arXiv2026

Axionic Wormholes in Metric-Affine Gravity

The axion is a promising candidate for solving the strong CP problem. To solve this problem, the global U(1) symmetry must be preserved to a high degree of accuracy. However, it is well known that global symmetries are explicitly violated by quantum gravity effects, giving rise to what is referred to as the axion quality problem. In this paper, we investigate axionic wormholes as a source of explicit U(1) violation in Metric-Affine Gravity. This framework allows for spacetime torsion and non-metricity, which accommodate additional curvature-like and topological terms, such as the Holst and Nieh--Yan terms, that are absent from the metric and Palatini formalisms. We show that non-minimal couplings to these terms modify the wormhole dynamics and enhance the Euclidean wormhole action, thereby alleviating the axion quality problem. We also find that the viable parameter space is enlarged when two of these couplings are simultaneously present. We further identify representative parameter regions where the alleviation of the axion quality problem is compatible with inflationary constraints.

hep-ph↗

arXiv2026

Spectral properties of Random Matrices

We give the theoretical foundations of random matrix theory through the definitions of a random matrix, a random probability measure and the corresponding empirical spectral distribution. The technical tool we use is the Stieltjes transform method through which we prove optimal convergence of the empirical spectral distribution of random sample covariance matrices to the deterministic Marchenko-Pastur distribution. We also give new results about the rigidity of the eigenvalues of this random sample covariance matrix and the rate of their convergence. We then define the Dyson equation method to prove new local laws about a random matrix model that interpolates between the Marchenko-Pastur distribution, the elliptical law and the circular law. Through our work these local laws can be considered universal.

math.PR↗