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 1,189 records · Page 66Linked to original sources

arXiv2026

Interpreting the top quark mass using simulations of top-flavoured mesons

Active work is being done on the issue of the interpretation of the top quark mass in terms of well-defined theoretical definitions. This work considers this issue by simulating fictitious heavy-light top-flavoured mesons with a fixed mass in Pythia 8.3. Within Heavy Quark Effective Theory, the mass of the top mesons can be related to the pole mass of the top quark. By fitting final-state observables in top pair production at current hadron and future lepton colliders, we extract a negative mass difference between the input Pythia mass for the top quark and the pole mass of around $240$-$330$ MeV, consistent with $\mathcal{O}(Λ_{\rm QCD})$.

hep-ph↗

arXiv2026

DeepOpen: Decision-Oriented Intent Classification on BANKING77 and CLINC150

We present DeepOpen, a decision-Oriented intent-classification framework with dataset- specific systems for BANKING77 and CLINC150. On BANKING77, a single English classi- fier continued with sharpness-aware minimization (SAM) achieves 94.1234% test accuracy (2,899/3,080). This is the best observed result among three fixed SAM seeds, selected for display after test evaluation; their mean is 94.0368%. On the 150 in-scope CLINC150 classes, probability fusion of one Laya specialist and three DeBERTa-v3-large classifiers achieves 98.0222% (4,411/4,500). We describe the two configurations, reproduce the reference com- parisons displayed in the accompanying evaluation explorer, and separate headline scores from evidence about improvement. Under a uniform BANKING77 evaluation path, the SAM mean exceeds the historical checkpoint by 0.0433 percentage points, with a shared-row 95% interval of [-0.2273, +0.3247]. Thus, the continuation does not establish a reliable accuracy gain. The report documents strong observed task-specific performance, selection effects, and unresolved historical execution-path differences, rather than a new backbone or a unified algorithmic advantage.

cs.AI↗

arXiv2026

Who Is the Influencer? Consensus Dynamics in Multi-Agent Systems

Multi-agent LLM systems are increasingly used in practice, but their behavior is complex and hard to understand. We design a simple task to ask which model has the most influence on the final consensus when several models are mixed in one system. Experimentally, across seven model families, we find that the opinions of the other models are absorbed into the opinion of the stronger model, the one with more parameters. This happens even though the task requires no pretraining knowledge and has no correct answer. To understand this phenomenon, we interpret the agents as weighted particles moving on a simplex. Larger models carry a larger weight and therefore occupy a deeper well in the energy landscape. A large weight means strong in-context learning: a model with a large weight amplifies the majority of its context. Larger models are generally better at in-context learning, and this is what makes the larger model the influencer. We hope that this work helps to interpret multi-agent systems, whose behavior is becoming too complex for humans to understand.

cs.AI↗

arXiv2026

Testing for peer effects in short panels without identifying the network structure

We propose an instrument-based test for peer effects in short-T panels when the underlying network is unknown. The test avoids estimating the network by aggregating pairwise comparisons between covariate-projected outcomes of potential peers and residualized own outcomes. The statistic is designed for fixed-T, large-N settings with individual fixed effects, heterogeneous slopes, heteroskedasticity, and serial dependence within units. We also develop a weighted extension that accounts for additive common time effects. The limiting distribution is standard Gaussian when the norm of the projected component of the outcomes is bounded away from zero, but can be nonstandard when it is local to zero. We derive a conditional Gaussian limit experiment and a feasible conditional Monte Carlo procedure to obtain uniformly valid critical values. Monte Carlo evidence shows that the conditional procedure leads to accurate size control while retaining power against latent peer-effect alternatives.

econ.EM↗

arXiv2026

Learning What to Ask for Tomorrow: Active Preference Acquisition for Long-Term Personalized Web Agents

Web agents increasingly serve the same user across repeated online tasks such as booking travel, shopping, and scheduling, so their long-term usefulness depends on how well they know the user's preferences. Passive memory and personalization methods cannot obtain preferences the user never reveals, and clarification methods that value a question by its benefit to the current task overlook the benefit of reusing the answer in later tasks. We study Active Preference Acquisition (APA): which missing preferences are worth the cost of asking? We define the Prospective Personalization Value (PPV) of a missing preference as its expected cumulative utility gain over future tasks minus the effort and disclosure-risk cost of asking, and select questions under a per-interaction budget. We characterize when cumulative reuse justifies acquisition despite insufficient single-task benefit, and evaluate this regime with an initial-profile diagnostic. APA computes the counterfactual gain of each candidate preference on public task prototypes with the agent's own task solver and aggregates it by the expected reuse of each task family. We evaluate it on ActiveWebPref, a controlled longitudinal benchmark with hidden but queryable user profiles and exact task utilities. At a horizon of 20 tasks, 34.4% of budget-feasible valuable missing preferences are amortization-dependent; APA acquires 80.3% of them, whereas current-task value of information acquires 5.1%. Under per-user asking-cost caps matched to APA's realized expenditure, APA achieves higher net utility than four heuristic selection baselines. In our cost setting, asking about every unknown preference yields lower net utility than not asking, while APA recovers 84.2% of its utility gain with 27.6% of its questions. The value of acquiring a user preference should account for its reuse across future tasks, not only its benefit to the task at hand.

cs.CL↗

arXiv2026

DRAFT: Decoding and Reconstructing 3D Room Layouts from fMRI and Camera Trajectories

While functional magnetic resonance imaging (fMRI) decoders reconstruct images and 3D objects, recovering the metric geometry of rooms from a moving viewpoint remains an open challenge. We introduce, to our knowledge for the first time, the task of reconstructing 3D layouts of held-out scenes using only the known camera trajectories and the fMRI signals of participants watching indoor walkthroughs, without any visual frames. The task is hard because of the temporal mismatch between rapid visual changes and blurred fMRI signals. Our method, DRAFT (Decoding and Reconstructing 3D room lAyouts from fMRI and Trajectories), extracts a compact layout representation from the early visual cortex and fuses this neural code with the geometric free space derived from camera motion. To evaluate it, we propose layout r, the Pearson correlation between predicted and ray-cast log depths. On ten held-out scenes (20 participants), DRAFT reaches a layout r of 0.480, outperforming the camera path alone (0.446) and seven fMRI decoders (at most 0.429). DRAFT also works as a plug-in that improves all seven decoders without retraining, raising their layout r from 0.171-0.429 to 0.487-0.493. Brain activity and camera motion can thus recover the metric layout of unseen rooms, a first step toward reconstructing the 3D spaces people see.

cs.CV↗

arXiv2026

Zero Forcing, Support Varieties, and Boundary Complexes

Let a finite simple graph be given. We study symmetric matrices with this graph pattern that admit a null vector with prescribed exact support. Such a vector exists precisely when its support is a fort, meaning that no vertex outside the support has exactly one neighbor inside it. Over the complex numbers, we prove that the Zariski closure of each exact-support coefficient locus is irreducible and compute its codimension in terms of boundary vertices and a labeled forest on the connected components of the support. The corresponding forest deficiency also determines the generic restricted nullity. An incidence parametrization yields a prime decomposition of the reduced zero-forcing failure locus. For a prescribed common kernel, we describe the coefficient space through equilibrium stresses, exterior affine dependencies, and free coefficients. We then construct boundary complexes recording local solution values and equation residuals. Their gluing maps are defined over the universal coefficient ring and commute with arbitrary base change. By retaining the complete collection of support signatures associated with each seed set, we obtain a composition formula for the minimum zero-forcing number and an algorithm for its computation.

math.CO↗

arXiv2026

On Modified Newton Methods for Unconstrained Optimization Based on Lanczos Tridiagonalization

For a strongly convex unconstrained minimization problem, Newton's method with a line search offers rapid convergence to a minimizer by incorporating second-order information. However, for a general unconstrained problem the Hessian need not be positive definite, which can lead to the inability to find a search direction or to search directions that fail to produce descent. Modified Newton methods are designed to handle this situation. This paper considers modifications to a Newton-conjugate-gradient approach based on Lanczos tridiagonalization. Existing methods are extended using dynamic modification schemes that provide a descent direction while keeping the modification bounded. A Lanczos-based scheme and a conjugate-gradient (CG)-based scheme are formulated and analyzed. These methods are mathematically equivalent in exact arithmetic, but their practical behavior differs in finite precision. Finally, it is shown how the convergence and computational cost are affected by varying the amount of additional curvature introduced and by allowing the early termination of the inner iterations.

math.OC↗

arXiv2026

Where Do Long Captions Fail? Position-Aware Diagnosis and Reinforcement Learning for Detailed Image Captioning

Detailed image captioning is usually framed as an amount problem: a model should mention more visual facts while avoiding hallucination. However, autoregressive VLMs generate captions as ordered sequences, and later generations may be produced under weaker or noisier visual grounding. We therefore argue that long-form captioning should be studied not only by what facts are mentioned, but also by where they appear. We introduce a position-aware rubric representation that decomposes dense human references into atomic visual claims, verifies generated captions against those claims, and induces a quartile-wise composition of supported, hallucinated, and generic content. This lens reveals two qualitatively different failures in long captions: models tend to produce unsupported specificity in the middle, while the final portion often escapes into vague, non-evidential language rather than simply accumulating more hallucination. A same-reference human control does not exhibit these patterns; additional human audits examine the visual validity of the induced labels. Motivated by this diagnosis, we propose PAGER (Position-Aware Grounded Evidence Reinforcement), a GRPO training objective with separate signals for supported fine-grained middle evidence, evidence-bearing tail continuation, claim coverage, and content-aware length control. The result is a closed loop: the same ordered claim representation identifies where long captions fail and supplies the position-specific reward needed to repair those failures. Experiments on detailed-caption benchmarks show that PAGER improves position-wise grounding and pairwise caption preference while preserving competitive global caption quality.

cs.CV↗

arXiv2026

Irrationality Exponents of Real CM Elliptic Periods and Powers of the Lemniscate Period

We give a determinant proof that every nonzero real period of a nonsingular elliptic curve over $\mathbb Q$ with complex multiplication, measured with a specified nonzero rational invariant differential, has irrationality exponent two. The proof uses an integral origin chart, matched local lengths, an actual compactification and weighted packet map, and arithmetic and analytic estimates for the same selected minor on a rectangular sublattice of the full CM period lattice. For the integral-defined lemniscate period $Ω=2\int_0^1(1-t^4)^{-1/2}\,dt$, we also give the uniform bound $μ(Ω^n)=O(n\log(2n))$ for integers $n\geq1$. Its new arithmetic ingredient clears integration denominators prime by prime while retaining the coefficient order; a one-root norm argument and an analytic estimate at algebraic centers complete the power proof. The two arguments are presented with their different charts and exact dependencies. Lean verifies the new CM determinant argument and rational normalization relative to explicit classical uniformization, CM-lifting and theta inputs; the lemniscate-power endpoint is also verified. The formalization record states the exact classical boundary and theorem references.

math.NT↗

arXiv2026

Can AI Co-Design Structures? An AI Agent for Early-Stage Structural Design

Recent advances in large language models and agentic workflows offer a new paradigm to early-stage structural design translating design intent into structural models. Although text-to-3D approaches have demonstrated that natural language can be used to produce visual, mesh-based geometric models, structural design requires domain-specific geometric arrangements and semantically rich data types. This paper introduces a proof-of-concept workflow for prompt-to-structure generation using an AI agent connected to a COMPAS-based structural modelling API through a Model Context Protocol (MCP) server. Through controlled modelling operations, the agent creates and modifies a graph-based structural data model composed of explicit building elements. The resulting artifacts can be inspected, modified, serialised, and further connected to downstream AEC workflows. The approach is demonstrated through selected prompt-to-structure examples and by progressive and editable design sequences that resemble typical iterative structural design workflows. The results are preliminary, but demonstrate the possibility of combining natural language, agentic workflows, and domain-specific structural data structures as a first step toward AI-assisted structural co-design.

cs.AI↗

arXiv2026

What Comes Next and Why: Interpretable Next-Event Prediction with Neuro-Symbolic Rules

In real-world applications, being able to explain why a prediction is made is often as important - if not more so - than achieving high accuracy. However, existing methods for sequential event prediction typically prioritize accuracy at the expense of interpretability. Neuro-symbolic approaches offer a promising direction to improve interpretability by learning symbolic rules, but extending them to next-event prediction is challenging due to the presence of temporal dependencies. We introduce RULES, a neuro-symbolic model that learns sequential rules directly from data, and combine it with a residual network to form RUNES (Rule Network for Event Sequences). RULES accurately recovers ground truth rules from synthetic data and discovers compact, interpretable rule sets on real data. RUNES combines these rules additively with a residual network, achieving competitive accuracy while providing transparent access to active rules and how residuals adjust predictions.

cs.LG↗

arXiv2026

The limits of magic-state distillation

Magic-state factories are a major cost of fault-tolerant quantum computing, yet how few noisy inputs they need is known only for restricted protocols. We show that, for tolerating input faults, any stabilizer protocol, however adaptive, postselected or multistage, is no better than a projection onto one stabilizer code, and that for rational targets a fault pattern is harmful exactly when its complement is. Searches over all codes show that quadratic distillation of one T state needs exactly nine inputs and that the eight-input CCZ factory is optimal. At order $d$, one T state needs $n\geq2d$.

quant-ph↗

arXiv2026

A Grammar of French Civil Cassation: A Ground-Level Dataset and Tool

This paper presents a dataset of 121,536 ground-reasons pairs across 86,464 rulings of the five civil chambers of the French Court of Cassation (2016-2025). An appeal on points of law cannot itself be appealed and often decides a case for good. Understanding how the Court rules is thus crucial and can help lawyers who draft those appeals. Inside a ruling, every ground of appeal (a distinct legal complaint against the appealed decision) has its own answer from the Court. This answer is called the reasons and it belongs to one of a few doctrinal families. Using a deterministic pipeline, we pair each ground with its reasons and attach the pair to its family, according to the established French taxonomy. We obtain a descriptive grammar of civil cassation that shows which grounds succeed, which legal texts they refer to, and how both evolve over the decade. We release the dataset, the code and an interactive consultation tool.

cs.CL↗

arXiv2026

Through a PRISM, Clearly: Step-Sequence Reasoning for Mistake Recognition in Procedural Videos

Procedural mistake recognition requires identifying deviations from valid multi-step activities. We study post-completion recognition from videos with benchmark-provided step boundaries. We introduce PRISM (Procedural mistake Recognition via Interpretable Semantic Modeling), which grounds video segments in language and maps their descriptions to procedural states using a lightweight hidden Markov model trained on correct executions. Reference alignment and semantic checks identify errors in step content and order. To support evaluation across heterogeneous benchmarks, we map dataset-specific annotations to four edit-based error types: insertion, deletion, substitution, and transposition. Across three benchmarks, we compare gold and VLM-generated descriptions in raw, semantic-role, and action-object forms. Decoder controls show consistent gains from the trained local decoder over nearest-candidate matching, with additional, dataset-dependent benefits from sequence decoding. Semantic role labeling improves Any-F1 in all six matched grounded comparisons, although the gains vary across datasets and do not translate into uniform improvements under separate configuration selection. These results establish language grounding as an effective interface for explicit procedural mistake reasoning and characterize the contributions of representation and decoding.

cs.CV↗

arXiv2026

ProductWebBench: Evaluating Website Continuity Beyond Task Completion

Frontend coding agents are usually graded on one question: did the requested target get built. Real engineering changes carry a second obligation, because the rest of the site has to keep working. We call that second obligation website continuity, and introduce ProductWebBench, a benchmark that scores both on living frontend repositories: 320 Change tasks over existing code and 80 long-horizon Build tasks. Every task freezes a repository snapshot, staged requirements, browser evidence and deterministic checks before evaluation, and enters only once a reference passes every check and plausible-wrong ones fail. A simulated user reveals the requirements one stage at a time, and every check cleared earlier stays armed for the rest of the run. Across 8,737 runs from 13 models the two come apart: of the Change runs passing every requested-work check, 28.2% still fail a continuity check -- a blank route, horizontal overflow, console errors -- and 10.8% push a browser state they had already passed back to failing. Because the same checks also run on the untouched repository, each regression carries exact provenance, not a post-hoc label: on the 586 runs where both readings exist, 18.3% are caught breaking a verified state. Interaction is a further blind spot, failing in 29.5% of runs whose static DOM already passes. ProductWebBench therefore scores a patch against the living system it lands in, not the target alone.

cs.SE↗

arXiv2026

Constraint Reasoning and Large Language Models Join Forces: Providing Contrastive Explanations for Real-World Optimisation-Based Applications

We focus on optimisation-based Decision Support Systems in real-world applications. In this context, we aim to provide contrastive explanations that answer users' questions about recommendations based on the solution of an optimisation problem, specifically a Mixed-Integer Program (MIP). Addressing this endeavour is crucial to ensure human oversight of the decisions implemented, as required by international guidelines on trustworthy and responsible automated decision-making systems. We propose a novel methodology that combines constraint reasoning techniques and Large Language Models (LLMs) to generate contrastive explanations for optimisation problems encoded as MIPs. Although LLMs have shown impressive capabilities in Natural Language processing, they can produce incorrect answers due to reasoning limitations. To overcome these limitations, we leverage constraint reasoning techniques to distil the information necessary to produce an explanation, which is then passed as input to an LLM as the so-called graph of reasons. Such a graph captures the semantics and the relationships among constraints of the optimisation problem. Results on real-world scheduling domains show that incorporating constraint reasoning improves explanations generated by the LLM. In particular, using the graph of reasons enables the LLM to produce natural-language explanations that include more relevant reasons while reducing errors compared to explanations generated without this structure.

cs.AI↗

arXiv2026

SetCommit: Coordinated Token Commitments for Few-Step Diffusion Language Model Decoding

Diffusion language models generate multiple token positions in parallel, but few-step decoding makes parallel commitment difficult: tokens that appear reliable individually may be jointly inconsistent, while premature commitments can prevent later context-dependent refinement. In this paper, we introduce SetCommit, a training-free method for coordinated parallel token commitment. Instead of selecting tokens independently by confidence, SetCommit selects a commitment set that balances token reliability against local joint commitment risk, preserving conditioning opportunities for subsequent denoising steps. It further performs budget-neutral local exchanges, replacing vulnerable earlier commitments when newly revealed context supports more reliable alternatives. SetCommit requires no retraining or architectural modification and is compatible with existing diffusion decoding strategies. Experiments with LLaDA-8B and Dream-7B on GSM8K, MATH-500, MBPP, and HumanEval show accuracy/pass@1 gains of up to 7.4 percentage points over native decoding at 32 model evaluations. At common quality targets, geometric-mean incremental speedups range from $1.44$ to $1.74\times$. Code is available at https://github.com/Kissmetothemoon/SetCommit.

cs.AI↗