Academic Intelligence · Curated Daily

Explore the Frontier of Global Academia

AcademicHub aggregates real-time literature from top journals and preprint platforms. Build your personal research radar and let large language models compile cross-disciplinary analysis briefings automatically.

01.
arXiv (quant-ph) 2026-06-11

Testing Catability and Coherent Superposition of $2\mathcal{D}$ Graphene Quantum system

arXiv:2605.10967v2 Announce Type: replace Abstract: We develop a theoretical framework for describing superposed coherent states in graphene quantum systems using the concept of catability as a phase-sensitive metric functional measure. In this case, the formalism quantifies interference stability and coherence structure via phase-dependent contributions of quantum superposition states. Catability is defined as a functional measure sensitive to relative phase variations within coherent state combinations, serving as a diagnostic tool for quantum interference effects in graphene-based systems. Also, the formulation is extended using Lie algebra techniques, where the underlying symmetry structure of graphene quantum states is represented through operator algebras governing state transformations in quantum space. In this context, to describe nonlocal propagation and phase-resolved dynamics, a Green function approach is incorporated, enabling systematic treatment of quantum correlations in a spatially extended structures framework. A unified framework is constructed by combining Lie algebraic symmetry analysis with Green function propagation theory, yielding a consistent description of phase-sensitive catability in complex graphene quantum configurations within the framework approach. Results provide a structured route for testing coherence, interference stability, and quantum state control in low-dimensional quantum materials systems.

02.
arXiv (CS.CV) 2026-06-17

Pareto LoRA: Mitigating Modality Imbalance in Unified Multimodal Models via Pareto-Optimal Gradient Integration

Unified multimodal models (UMMs) have recently emerged as a promising paradigm for integrating multimodal understanding and generation within a single autoregressive transformer. However, during multimodal instruction tuning, these models often exhibit pronounced modality imbalance: language gradients dominate optimization, thus leading to lower image generation quality, especially under parameter-efficient fine-tuning such as LoRA. In this work, we systematically analyze modality imbalance in LoRA-based fine-tuning of UMMs for interleaved text-image generation. We show that vision modality performance degrades substantially more than text modality performance when compared to unimodal counterparts, and that modality-specific gradients can differ by orders of magnitude across various tasks and layers. Motivated by this observation, we reformulate the multimodal instruction tuning as a bi-objective optimization problem and propose Pareto LoRA, a Pareto-optimal gradient integration strategy that balances the text and image objectives by modulating the gradient direction and strength. Experiments on the CoMM benchmark with Emu2 demonstrate that Pareto LoRA consistently improves multimodal generation balance, achieving up to 44.9% gains in perceptual image quality over vanilla LoRA while maintaining comparable text performance.

03.
arXiv (math.PR) 2026-06-11

Hilbert space embeddings of independence tests and interaction measures of several variables

arXiv:2411.08653v2 Announce Type: replace-cross Abstract: We present a unified theoretical framework for kernel-based measures of dependence on product spaces. Building on the ideas underlying distance covariance, distance multivariance, and the Hilbert-Schmidt Independence Criterion (HSIC), we define a new family of kernels on an $n$-fold Cartesian product, termed positive definite independent of order $k$ (PDI$_{k}$ kernels). These kernels extend the concepts of positive definite and conditionally negative definite kernels to higher orders and provide the foundation for generalized independence and interaction tests, such as the generalized Lancaster interaction of order $k$ ($\Lambda_{k}^{n}$), and the Streitberg interaction ($\Sigma$). Our analysis focuses on the continuous setting, where we prove a Kernel Mean Embedding Theorem for PDI$_{k}$ kernels and establish the corresponding integrability restrictions. Based on these results, we characterize how the Kronecker products of PDI kernels behave.

04.
arXiv (CS.CL) 2026-06-11

Every Act Has Its Price: Compressed Moral Composition in Frontier LLMs

Existing LLM moral benchmarks usually ask which isolated moral act, value, or foundation a model prefers. This is useful but incomplete. Realistic judgments often require a model to combine several moral signals within the same option. We introduce **Moral Trolley Arena**, a two-stage blind ELO benchmark for measuring how LLMs compose moral evidence. The single-scene arena first calibrates individual moral acts from a 229-scenario corpus across five Moral Foundations Theory foundations; the composite arena then combines calibrated acts into two-act moral items over a controlled intensity grid and measures the resulting composite preferences. Across ten frontier models, composite judgments are largely predicted by component act strength, but the relation is consistently compressed rather than simply additive. Models also show non-additive intensity anchoring, bounded foundation-specific residuals after component control, and highly convergent composite preference surfaces across providers. These results suggest that moral audits should measure composition rules for moral evidence, not only rankings over isolated acts.

05.
arXiv (CS.AI) 2026-06-16

MR-GVNO: A Geometry-Aware Variational Physics-Informed Neural Operator for Mindlin-Reissner Plates on Irregular Domains

arXiv:2606.16624v1 Announce Type: new Abstract: Plate and shell structures are widely used in engineering, making rapid response prediction under varying geometries, materials, and loads highly desirable. However, conventional finite element methods require repeated modeling and solution, resulting in high computational costs. This study proposes a geometry-aware variational neural operator for Mindlin-Reissner plate problems, termed MR-GVNO. The method uses boundary point clouds to represent irregular geometries and employs separate encoders for spatially varying material fields, pressure loads, and scalar physical parameters. A cross-attention mechanism integrates these inputs with query point information to predict transverse deflections and rotations at arbitrary locations. MR-GVNO is trained without labeled solution data using a variational physics-informed loss derived from the discretized total potential energy. It directly processes irregular point clouds and allows different physical fields to be discretized independently, avoiding interpolation onto a common grid. Numerical experiments on single-hole, double-hole, and L-shaped plates demonstrate accurate response prediction under homogeneous and heterogeneous materials and uniform and random loads. The model also achieves millisecond-level full-field inference and favorable cross-geometry generalization.

06.
arXiv (CS.CV) 2026-06-18

SAMA: Semantic Anchor-aligned Augmentation for Unified Low-Resource Multimodal Information Extraction

Multimodal Information Extraction (MIE)-covering tasks such as Multimodal Named Entity Recognition (MNER), Relation Extraction (MRE), and Event Extraction (MEE)-is essential for understanding multimedia content but remains constrained by severe data scarcity. Although data augmentation is a promising remedy, existing approaches are impeded by coarse cross-modal alignment and fragmented, task-specific designs that fail to exploit shared semantic knowledge. To overcome these limitations, we introduce Semantic Anchor-aligned Multimodal Augmentation (SAMA), a unified framework for generating high-fidelity, task-aware synthetic data. SAMA constructs structured semantic anchors from ground-truth labels to guide a Collaborative Multi-Experts Multimodal Large Language Model (CME-MLLM), which integrates a Universal Adapter for shared semantics with Task-Specific Adapters to produce diverse yet constraint-compliant textual samples. For image synthesis, SAMA employs an Anchor-Preserving Diffusion mechanism that uses anchor-weighted prompts and latent conditioning to maintain critical semantic anchors while diversifying visual contexts. To eliminate the need for manual verification, SAMA further introduces a Dual-Constraint Filtering module that selects synthetic samples based on both cross-modal consistency and anchor fidelity. Extensive experiments across benchmark datasets for MNER, MRE, and MEE demonstrate that SAMA consistently outperforms state-of-the-art augmentation baselines under both fully supervised and low-resource settings, underscoring its versatility, robustness, and effectiveness.

07.
arXiv (CS.CL) 2026-06-16

A Systematic Evaluation of Large Language Models for PTSD Severity Estimation: The Role of Contextual Knowledge and Modeling Strategies

Large language models (LLMs) are increasingly being used in a zero-shot (generative) fashion to assess mental health conditions, yet we have limited knowledge on what factors affect their accuracy. In this study, we use a clinical dataset of natural language narratives and self-reported PTSD severity scores from 1,437 individuals to comprehensively evaluate the performance of 11 state-of-the-art LLMs. To understand the factors affecting model's assessment accuracy, we systematically varied (i) contextual knowledge prompted to the models like subscale definitions, distribution summary, and interview questions, and (ii) modeling strategies including zero-shot vs few shot, amount of reasoning effort, model sizes, structured subscales vs direct scalar prediction, output rescaling and nine ensemble methods. Our findings indicate that (a) LLMs are most accurate when provided with detailed construct definitions and context of the narrative, even exceeding human raters agreement with self-reported scores; (b) increased reasoning effort leads to better estimation accuracy; (c) performance of open-weight models (Llama, DeepSeek) plateaus beyond 70B parameters while closed-weight (gpt-o3-mini, gpt-5) alternatives improve with newer generations; and (d) best performance is achieved when ensembling a supervised model with the zero-shot LLMs. Beyond agreement with self-reports, LLMs' estimates discriminated PTSD severity from depression, anxiety, and alcohol use, and prospectively predicted future mental healthcare expenditure. Together, these results suggest that contextual knowledge and modeling strategies meaningfully affect accuracy and clinical utility of LLM-based assessments of PTSD severity.

08.
arXiv (CS.CL) 2026-06-11

GrowLoop: Self-Evolving Conversation Evaluation Seeded by Human

With the rapid advancement of large language models, evaluating human-likeness in open-ended conversation has become increasingly important. However, human-likeness is a form of tacit knowledge that humans perceive intuitively, yet the underlying criteria resist explicit formulation. Human judgments vary widely, with strong agreement on some cases and legitimate disagreement on others. Meanwhile, the criteria behind human judgments remain implicit, leaving no clear basis for constructing cases. Further, what counts as human-likeness is not static, but evolving with model capability and human expectations. Despite progress in evaluation methods such as expert-authored benchmarks, Reward Models, and self-evolving benchmarks, none addresses all three challenges simultaneously. Therefore, we propose GrowLoop, a self-evolving conversation evaluation system that continuously adapts as models advance and scenarios shift. Starting from minimal human seed annotations, LLM agents iteratively extract and refine evaluation rubrics through Heuristic Learning. Human-AI agreement is required where annotators converge, while only plausibility is expected where they diverge. Moreover, the Rubric-Case co-evolution mechanism enables continuous evolution. When the evaluation target shifts, new human seeds expand the system's coverage accordingly. When applied to human-likeness evaluation in open-ended conversation, the AI judge guided by these rubrics not only substantially outperforms existing methods in alignment with human judgments, but also uncovers issues that annotators overlook. The resulting benchmark effectively discriminates models across capability tiers and reveals where they fall short, while generalizing to new scenarios and adapting as models advance. Our work shifts the benchmarking paradigm from manual updates or difficulty scaling to comprehensive, continuous self-evolution.

09.
arXiv (CS.LG) 2026-06-17

Questioning the Coverage-Length Metric in Conformal Prediction: When Shorter Intervals Are Not Better

arXiv:2601.21455v2 Announce Type: replace-cross Abstract: Conformal prediction(CP) has become a cornerstone of distribution-free uncertainty quantification, conventionally evaluated by its coverage and interval length. This work critically examines the sufficiency of these standard metrics. We demonstrate that the interval length might be deceptively improved through a counter-intuitive approach termed Prejudicial Trick(PT), while the coverage remains valid. Specifically, for any given test sample, PT probabilistically returns an interval, which is either null or constructed using an adjusted confidence level, thereby preserving marginal coverage. While PT potentially yields a deceptively lower interval length, it introduces practical vulnerabilities: the same input can yield completely different prediction intervals across repeated runs of the algorithm. We formally derive the conditions under which PT achieves these misleading improvements and provide extensive empirical evidence across various regression and classification tasks. Furthermore, we introduce a new metric interval stability which helps detect whether a new CP method implicitly improves the length based on such PT-like techniques. Code is available at https://github.com/benben-cd/PT-Conformal-Prediction.

10.
arXiv (CS.LG) 2026-06-16

Size Doesn't Matter: Cosine-Scored Sparse Autoencoders

arXiv:2606.15054v1 Announce Type: new Abstract: Sparse autoencoders (SAEs) detect features via inner product, so a feature's activation scales with both its directional alignment and the input's norm. Under BatchTopK, high-norm tokens inflate all pre-activations simultaneously, claiming dictionary slots regardless of content alignment. This matters because sublayer normalization has already discarded the magnitude the score measures, so the encoder detects a quantity the model does not read. We replace the score with a learned blend of cosine similarity and input magnitude, letting the optimizer choose how much norm to use; a per-feature extension lets each feature decide independently. In both regimes, training is free to recover inner product but never does, with no feature ever choosing more than half-magnitude dependence. At matched reconstruction, the cosine encoder learns features that align with human-recognizable concepts far more often than standard, filling dictionary slots that inner product wastes on norm detectors. Loss reweighting that equalizes gradients barely closes the gap, confirming forward-pass score geometry as the lever. The advantage is not universal across tasks or depths, but we believe cosine scoring should be the default for dictionary learning on normalized representations.

11.
arXiv (CS.AI) 2026-06-16

Rethinking Scaffolding in LLM Tutors: The Interactional Mismatch Between Benchmarks and Real-World Deployments

arXiv:2606.15766v1 Announce Type: new Abstract: A central pedagogical value evaluated in AI tutor benchmarks is scaffolding: guiding students through graduated steps toward a solution. Alignment and evaluation methods for embedding scaffolding behaviour into chatbots, however, rest on an implicit assumption: that students will take up the scaffolding and engage in the conversation. To examine whether this assumption holds, we introduce an evaluation pipeline around two metrics - Chatbot Scaffolding and Student Uptake - and apply them across nine datasets of 9,490 chats, spanning AI tutor benchmarks and real-world deployments of educational chatbots. Our analysis reveals that while benchmarks assume a high-scaffolding, high-student-uptake environment, students in real-world settings exhibit lower levels of uptake overall - frequently bypassing the chatbot's pedagogical framing to drive the interaction toward their own learning goals at little interpersonal cost. We argue that bypassing scaffolding is not necessarily detrimental; rather, it frequently highlights a mismatch between a chatbot's pedagogical framing and the student's learning goals. To meaningfully evaluate the effectiveness of a chatbot's assistance, future benchmarks must move beyond the assumption that students will simply take up the scaffolding, and instead evaluate how these chatbots navigate diverse learning contexts and student-driven interaction patterns.

12.
arXiv (CS.AI) 2026-06-12

Rethinking RAG in Long Videos: What to Retrieve and How to Use It?

arXiv:2606.13141v1 Announce Type: new Abstract: Retrieval-augmented generation is moving beyond text into long, egocentric video, where systems must select query-relevant chunks across multiple modalities and temporal granularities. Yet progress in VideoRAG is limited by two gaps: existing benchmarks allow queries to be answered without the video, obscuring retrieval errors, and prior methods apply a single modality-granularity configuration per query, ignoring chunk-level variability. We address both by introducing V-RAGBench, a benchmark of $\langle$query, evidence chunk, answer$\rangle$ triplets that enables faithful, decoupled evaluation of retrieval and generation, and CARVE, a simple method that runs parallel retrievers across configurations and employs chunk-adaptive reranking to identify the winning configuration for each chunk. Each chunk then enters the generator under its winning configuration selected during retrieval, yielding an interleaved evidence form where the chunk-level decision propagates across both stages. CARVE outperforms eight recent VideoRAG baselines, with the chunks supplied to the generator interleaving multiple configurations rather than sharing a single one, a behavior unattainable by query-level methods.

13.
medRxiv (Medicine) 2026-06-23

The Target ALS Global Natural History Study: Cross-platform proteomics to accelerate biofluid biomarker and drug target discovery in amyotrophic lateral sclerosis

Amyotrophic lateral sclerosis (ALS) is a fatal, rapidly progressive neurodegenerative disease of motor neurons for which therapeutics are limited. Improved biomarkers are imperative to improve patient care and therapeutic development. Here, we employed 35-plex isobaric tandem mass tag labeling based on isobutyl-proline reporter group (TMTpro) to perform unbiased proteomic analysis of cerebrospinal fluid (CSF) and plasma from control (n= 28, n= 31) and sporadic ALS (sALS) (n= 39, n= 41), from the Target ALS Global Natural History Study (TALS GNHS). We identified 2,875 proteins in CSF and 1,118 proteins in plasma and identified known and novel differentially expressed proteins (DEPs) between controls and sALS, some of which were orthogonally validated using immunoassay. Comparison of TMTpro-MS and Olink proximity extension assay proteomics revealed common and non-overlapping differentially expressed proteins illustrating strengths unique to each platform. This initial cross-sectional proteomic study of biofluids from the TALS GNHS, with unrestricted availability of study results to the research community, highlights the potential of this resource as a potent platform for ALS biomarker discovery.

14.
PLOS Medicine 2026-06-04

Comparative impacts and cost-effectiveness of tuberculosis systematic screening strategies in prisons in Brazil, Colombia, and Peru: A mathematical modeling study

Authors:

by Yiran E. Liu, José Victor Bortolotto Bampi, Ronan F. Arthur, Argita D. Salindri, Caroline Busatto, Pedro Avedillo Jiménez, Daniele Maria Pelissari, Fernanda Dockhorn Costa Johansen, Robert Arana-Narvaez, Alvaro Fernando Moreno Roca, Wilfredo Santos Solís Tupes, Esther Mori Jiu, Christian Alfredo Moreno Roca, Erika Albertina Abregú Contreras, Valentina Antonieta Alarcón Guizado, Julián Trujillo Trujillo, Belkys Marcelino, Mónica Alonso Gonzalez, Mayra Cecilia Córdova Ayllon, Ted Cohen, Moises A. Huaman, Jeremy D. Goldhaber-Fiebert, Julio Croda, Jason R. Andrews Background Incarceration is a leading driver of tuberculosis in Latin America. Systematic screening in prisons may reduce tuberculosis burden, but optimal strategies and cost-effectiveness remain uncertain. We examined the population-wide health impacts and cost-effectiveness of systematic screening in prisons in Brazil, Colombia, and Peru, comparing different timepoints, frequencies, and screening algorithms. Methods and findings Using dynamic transmission models calibrated to Brazil, Colombia, and Peru, we simulated annual or biannual (twice-yearly) prison-wide screening, alone or combined with entry and exit screening from 2026 to 2035. We evaluated four algorithms: (1) symptom screening, (2) chest X-ray with computer-aided detection (CXR-CAD), (3) symptoms and CXR-CAD (follow-up testing if either is positive), and (4) GeneXpert Ultra (Xpert) with pooled sputum. Individuals screening positive then received individual Xpert. We projected impacts on within-prison and population-level tuberculosis incidence in 2035, along with discounted costs (2023 US dollars) and disability-adjusted life years (DALYs). Model projections showed that combined entry, exit, and biannual screening with CXR-CAD was highly impactful and cost-effective across countries, reducing tuberculosis incidence by 61%–87% in prisons and 18%–28% population-wide. Compared to only biannual CXR-CAD (the next best strategy), the incremental cost per DALY averted of adding entry and exit screening was $2,984 (Brazil), $2,925 (Colombia), and $645 (Peru). Adding symptom screening to CXR-CAD marginally increased benefit and was only cost-effective in Peru’s higher-incidence prisons. Biannual screening alone remained cost-effective at prison incidence levels well below national averages, as well as at far lower willingness-to-pay thresholds. In settings without CXR-CAD, pooled Xpert was an impactful, cost-effective alternative. Key limitations include the model’s simplified representation of tuberculosis disease states and lack of stratification by age, gender/sex, HIV, or drug resistance. Conclusions These modeling results support immediate national-level adoption of prison-wide tuberculosis screening twice-yearly and at entry and exit, using CXR-CAD or pooled Xpert.

15.
arXiv (CS.LG) 2026-06-18

Predicting the Neutrino Mass Ordering Using Neural Networks

arXiv:2606.03745v1 Announce Type: cross Abstract: Determining the neutrino mass ordering remains a central open problem in particle physics. While next-generation long-baseline experiments are expected to resolve this question, current data provide limited sensitivity because the spectral differences between normal and inverted ordering are subtle and entangled with parameter degeneracies. We investigate a machine-learning strategy for mass-ordering determination using a feed-forward neural-network classifier trained on synthetic long-baseline datasets generated with three-flavour oscillation probabilities, matter effects, and statistical fluctuations. We evaluate the classifier against standard $\chi^2$ and $\log\mathcal{L}$ approaches using common discrimination metrics, including receiver-operating-characteristic curves, to quantify sensitivity and to illustrate how operating points can be selected to prioritise purity or efficiency. We find that the neural network achieves performance comparable to conventional fits for the scenarios studied, providing a flexible, independent cross-check of established analyses. The framework can be extended to incorporate systematic uncertainties and to explore joint inference of oscillation parameters, and it may also serve as a pedagogical tool for introducing machine-learning methods in neutrino physics.

16.
arXiv (CS.CL) 2026-06-11

Fine-tuning Multi-modal LLMs with ART: Art-based Reinforcement Training

There are two main Parameter-Efficient Fine-Tuning (PEFT) techniques for Large Language Models (LLMs). While Low-Rank Adaptation (LoRA) introduces additional weights between the LLM layers, Soft Prompting introduces additional fine-tuning-specific raw tokens to an LLM input. However, both require modification to the computational graphs of precompiled, preoptimized LLMs. As a result, neither is fully supported in high-throughput engines like vLLM. We propose fine-tuning with ART (Art-based Reinforcement Training). The method injects information into a frozen Multimodal Large Language Model (MLLM) by optimizing only its raw visual input, thus enabling the soft-token approach on pre-compiled computational graphs. It relies on backpropagation of gradients back into a plain pixel array and thus supports any fine-tuning objective. Moreover, the optimized visual input can be stylized as task-relevant computational artworks. The approach's effectiveness is confirmed for different sizes of a popular open Qwen architecture and for several textual benchmarks. Specifically, ART reaches accuracy competitive with LoRA across mathematics and structured-tool-use benchmarks.

17.
arXiv (math.PR) 2026-06-15

Mixing Times for the Facilitated Exclusion Process

arXiv:2402.18999v2 Announce Type: replace Abstract: The facilitated simple exclusion process (FEP) is a one-dimensional exclusion process with a dynamical constraint. We establish bounds on the mixing time of the FEP on the segment, with closed boundaries, and the circle. The FEP on these spaces exhibits transient states that, if the macroscopic density of particles is at least $1/2$, the process will eventually exit to reach an ergodic component. If the macroscopic density is less than $1/2$ the process will hit an absorbing state. We show that the symmetric FEP (SFEP) on the segment $\{1,\ldots,N\}$, with $k>N/2$ particles, has mixing time of order $N^{2}\log(N-k)$ and exhibits the pre-cutoff phenomenon. For the asymmetric FEP (AFEP) on the segment, we show that there exists initial conditions for which the hitting time of the ergodic component is exponentially slow in the number of holes $N-k$. In particular, when $N-k$ is large enough, the hitting time of the ergodic component determines the mixing time. For the SFEP on the circle of size $N$, and macroscopic particle density $\rho \in(1/2,1)$, we establish bounds on the mixing time of order $N^{2}\log N$ for the process restricted to its ergodic component. We also give an upper bound on the hitting time of the ergodic component of order $N^{2}\log N$ for a large class of initial conditions. The proofs rely on couplings with exclusion processes (both open and closed boundaries) via a novel lattice path (height function) construction of the FEP.

18.
arXiv (CS.AI) 2026-06-16

CoffeeBench: Benchmarking Long-Horizon LLM Agents in Heterogeneous Multi-Agent Economies

arXiv:2606.16613v1 Announce Type: new Abstract: As LLM agents become capable of increasingly long-horizon tasks, evaluating their performance in economic systems is becoming increasingly important. Unlike existing benchmarks that primarily evaluate a single agent interacting with a passive environment, economic systems are inherently multi-agent, requiring autonomous agents to communicate, negotiate, and transact while pursuing their own objectives over extended periods. We introduce CoffeeBench, a benchmark for evaluating LLM agents in a long-horizon multi-agent economy composed of heterogeneous firms. In CoffeeBench, two farmers, two roasters, and two retailers autonomously operate their businesses over a 90-day simulation, each seeking to maximize cumulative net income through communication and transactions while managing cash, inventory, and pricing. The evaluated model controls one coffee roaster, while the remaining firms are controlled by fixed reference agents. Across several recent open-weight and proprietary LLMs, all models outperform a passive baseline that takes no actions, with most achieving positive net income. Analysis of agent behavior reveals substantial differences in long-horizon economic interaction: higher-performing models communicate more actively with other firms, whereas Claude~Haiku~4.5 exhibits an idle-drift failure mode, repeatedly choosing inaction despite producing coherent assessments and plans. We release our code and agent trajectories to support future research.

19.
arXiv (CS.CL) 2026-06-17

Dissociating Decodability and Causal Use in Bracket-Sequence Transformers

When trained on tasks requiring an understanding of hierarchical structure, transformers have been found to represent this hierarchy in distinct ways: in the geometry of the residual stream, and in stack-like attention patterns maintaining a last-in, first-out ordering. However, it remains unclear whether these representations are causally used or merely decodable. We examine this gap in transformers trained on the Dyck language (a formal language of balanced bracket sequences), where the hierarchical ground truth is explicit. By probing and intervening on the residual stream and attention patterns, we find that depth, distance, and top-of-stack signals are all decodable, yet their causal roles diverge. Specifically, masking attention to the true top-of-stack position causes a sharp drop in long-distance accuracy, while ablating low-dimensional residual stream subspaces has comparatively little effect. These results, which extend to a templated natural language setting, suggest that even in a controlled setting where the relevant hierarchical variables are known, decodability alone does not imply causal use.

20.
arXiv (CS.CV) 2026-06-17

Geometric Consistency Protocol for Foundation Model Features in Multi-View Satellite Imagery

Standardized evaluation protocols are indispensable for robust benchmarking in remote sensing, particularly as foundation features are increasingly transferred across diverse sensors and complex imaging geometries. In satellite multi-view reconstruction, conventional evaluations relying on unconstrained 2D global matching are often misleading. The Rational Function Model (RFM) and its Rational Polynomial Coefficients (RPC) dictate a curved, height-dependent epipolar geometry that render flat 2D search spaces physically inconsistent. We propose a geometry-faithful and reproducible protocol tailored for the RPC framework. Our approach integrates an RPC-projected 3D consistency metric with a geometry-constrained dense matching proxy, specifically evaluating whether similarity responses remain localized and unique under physically plausible search manifolds. A pivotal finding of our joint reporting strategy is the decoupling of semantic agreement and geometric localization: high cross-view similarity at a projected 3D point does not guarantee reliable matchability in practical inference. Our benchmark demonstrates that incorporating geometric constraints is fundamental to the problem definition in satellite imagery. Furthermore, we show that state-of-the-art 2D backbones remain remarkably competitive against specialized 3D-aware models when subjected to this RPC-consistent evaluation.

21.
arXiv (CS.CL) 2026-06-11

Reassessing High-Performing LLMs on Polish Medical Exams: True Competence or Bias-Driven Performance?

Large language models (LLMs) in medicine are mainly evaluated using multiple-choice question answering (MCQA), which can overestimate real clinical ability due to guessing strategies and answer biases. To address these limitations, we introduce an expanded and more challenging benchmark based on Polish medical exams, adding over 15,000 questions, two new domains, and four structural modifications that reduce MCQA-specific artifacts and better test reasoning. We evaluate 21 LLMs and show that evaluation design strongly affects results. Under our harder setup, the best model (Qwen3.5-122B) drops by 28.4 and 31 pp on English and Polish exams, respectively. Despite low evidence of data contamination, standard MCQA scores do not reliably reflect true medical competence. To facilitate further research, we make our benchmark publicly available.

22.
arXiv (CS.CV) 2026-06-16

Divide-and-Denoise: A Game-Theoretic Method for Fairly Composing Diffusion Models

The abundance of pre-trained diffusion models provides an opportunity for composition. Combining several models, however, runs the risk of one model dominating or models disagreeing with each other. Here, we propose Divide-and-Denoise, a method for coordinating multiple pre-trained diffusion models during sampling. Much like managing a specialized workforce, our method creates a fair but efficient division of labor across models. Central to our method is the notion of an allocation which defines the responsibility of each model to every region of the noisy sample. At every timestep, we then denoise by (i) updating the allocation by solving a fair division game, where we divide the sample into regions that maximize total utility under fairness constraints, and (ii) aligning the models with this allocation, where we guide each model to denoise within its assigned region. This leads to a new composite denoising process that evolves in tandem with a division process. We evaluate Divide-and-Denoise on conditional image generation. Across several quality metrics, including the GenEval benchmark, our method outperforms baselines and resolves common failures including missing objects and mismatched attributes. Experiments show that Divide-and-Denoise utilizes each model's expertise without neglecting any other model.

23.
arXiv (CS.CL) 2026-06-16

SCAR: Semantic Continuity-Aware Retrieval for Efficient Context Expansion in RAG

Fixed-length chunking in Retrieval-Augmented Generation (RAG) often leads to boundary fragmentation, where critical evidence is split across segments, degrading retrieval recall. While static windowing and parent retrieval improve recall, they introduce significant token overhead. We propose SCAR (Semantic Continuity-Aware Retrieval), an adaptive retrieval policy that selectively expands neighboring chunks by weighing query-neighbor relevance against a structural continuity penalty. SCAR uses a relative expansion threshold tied to each retrieved chunk's own query-relevance, yielding an approximately scale-invariant decision rule that transfers across embedding models without recalibration. Across four diverse corpora (RFC, GDPR, a 10-K report, and a Merger agreement; N=320 queries; 160 boundary-fragmented), SCAR achieves 92.8% recall on boundary-fragmented queries with only 7.84 chunks, a 22.9% reduction compared to static windowing (10.16 chunks). Paired bootstrap tests (B=10,000) confirm the chunk reduction is highly significant (p

24.
arXiv (CS.LG) 2026-06-16

Benchmarking Instance-Dependent Label Noise with Controlled Corruptions

arXiv:2606.14965v1 Announce Type: new Abstract: Synthetic instance-dependent label noise (IDN) benchmarks are widely used to evaluate noisy-label learning methods, yet existing approaches typically generate noise through imperfect annotators or classifier raters, leaving the source of ambiguity implicit. We introduce CILN, a benchmark generation framework that creates IDN through controlled input corruptions. A diverse voter pool labels corrupted instances, producing benchmark datasets in which both the source and severity of ambiguity are explicit and controllable. Using CIFAR10, MNIST, and Adult, we construct 90 benchmark settings spanning multiple corruption families and severity levels. Our experiments show that the resulting benchmarks exhibit genuine instance-dependent noise, provide diverse confusion structures, and, on CIFAR-10, can produce label distributions that are closer to human uncertainty than an existing synthetic IDN benchmark. We further demonstrate that corruption-mediated IDN can expose failure modes of popular noisy-label learning methods, including Co-Teaching and DivideMix, that are not observed under comparable levels of rater-fallibility noise. These findings suggest that noise structure, not only noise rate, plays an important role in benchmark difficulty and algorithm behavior. By making ambiguity generation explicit and controllable, CILN provides a complementary benchmarking framework for studying noisy-label learning under diverse sources of instance difficulty.