Academic Intelligence · Curated Daily

探索全球前沿学术脉络

AcademicHub 汇聚顶级期刊与预印本平台的实时文献。定制您的专属科研雷达,利用大语言模型自动生成交叉领域文献分析简报。

01.
arXiv (math.PR) 2026-06-16

A small noise approximation for Muller's Ratchet

arXiv:2606.15842v1 Announce Type: new Abstract: We consider an infinite system of SDEs with Fleming-Viot noise indexed by $k=0,1,2,\dots$, whose parameters $\alpha,\lambda$, and $\nu$ are the (deleterious) selection coefficient, the (uni-directional) mutation rate, and a quantity which determines the size of the system's fluctuations. The SDE's unique weak solution $X(t) = (X_k(t))_{k=0,1,2,...}$ models what is known in population genetics as Muller's ratchet. Here, $X_k(t)$ stands for the frequency of individuals carrying $k$ deleterious mutations. Since the mutation process is uni-directional, $t\mapsto \inf\{k: X_k(t)> 0\}$ is non-decreasing for almost every path of $X$, and we refer to an increase as a click of Muller's ratchet. A long standing question concerns the clicking rate of Muller's ratchet. Using Duhamel's principle for semigroups, we give a partial answer by approximating $E(\sum_{k=1}^\infty kX_k(t) )$ and $E\big(X_0(t)\big)$ up to $O(1/\nu^2)$ for fixed $\alpha$, $\lambda$ and $t>0$. Our results suggest that $\psi:=\nu \alpha e^{-\lambda/\alpha}$ is a crucial quantity also when the mutation/selection ratio $\theta = \lambda/\alpha$ is moderately large: for large $\nu \alpha$, clicking of the ratchet on the time scale $\frac 1\alpha \log \theta$ becomes rare as soon as $\psi$ becomes large.

02.
arXiv (CS.CV) 2026-06-16

Planning with Unified Multimodal Models

With the powerful reasoning capabilities of large language models (LLMs) and vision-language models (VLMs), many recent works have explored using them for decision-making. However, most of these approaches rely solely on language-based reasoning, which limits their ability to reason and make informed decisions. Recently, a promising new direction has emerged with unified multimodal models (UMMs), which support both multimodal inputs and outputs. We believe such models have greater potential for decision-making by enabling reasoning through generated visual content. To this end, we propose Uni-Plan, a planning framework built on UMMs. Within this framework, a single model simultaneously serves as the policy, dynamics model, and value function. In addition, to avoid hallucinations in dynamics predictions, we present a novel approach self-discriminated filtering, where the generative model serves as a self-discriminator to filter out invalid dynamics predictions. Experiments on embodied decision-making tasks show that Uni-Plan substantially improves success rates compared to VLM-based methods, while also showing strong data scalability, requiring no expert demonstrations and achieving better performance under the same training-data size. This work lays a foundation for future research in reasoning and decision-making with UMMs.

03.
arXiv (CS.LG) 2026-06-11

A Multi-Modal Sensor Fusion Instrument for Measuring Regional Human Mobility: The Distributed Human Data Engine (DHDE)

arXiv:2603.21639v2 Announce Type: replace-cross Abstract: Accurately estimating human mobility in peripheral regional economies presents a fundamental measurement challenge: physical ground-truth sensors are sparse, behavioral intent signals are heterogeneous, and environmental friction introduces systematic bias into demand inference. We introduce the Distributed Human Data Engine (DHDE), a multi-modal sensor fusion architecture that addresses this challenge by integrating physical instrumentation (Edge-AI cameras), digital intent signals (route search impression metrics), behavioral records (90,350 spending records, 97,719 standardized survey responses), and meteorological data across four geographically distributed nodes in Fukui, Japan. The primary measurement-science contribution is the design, deployment, and cross-node validation of the DHDE as a sparse-sensor compensation instrument: a heterogeneous sensor fusion architecture that anchors non-stationary digital intent signals to concurrent physical ground-truth counts, correcting for systematic bias introduced by meteorological planning friction. The instrument is implemented as an ensemble inference pipeline (Random Forest and Ordinary Least Squares with Newey-West robust inference), calibrated across 397 daily observations and validated by chronological holdout replication across four geographically distinct node types. The primary OLS specification achieved an in-sample explanatory power of R2 = 0.810 and a chronological out-of-sample predictive performance of R2 = 0.683. Results identify an Under-Vibrancy Paradox where macro-regional visitor satisfaction correlates positively with crowd density (Spearman rank correlation rs = +0.150, p = 0.002). We estimate an annual proxy gap of 865,917 intent-implied visits, corresponding to JPY 11.96 billion (USD 72.6 million) in foregone revenue.

04.
arXiv (CS.LG) 2026-06-16

Temporal Validation Changes the Apparent Public-Health Utility of Under-Five Mortality Prediction in Bangladesh: A Four-Round DHS Machine-Learning Study

arXiv:2602.03957v2 Announce Type: replace Abstract: Background: Under-five mortality in Bangladesh remains uneven despite national progress. DHS-based prediction models may guide targeted follow-up, but only if validation reflects future use. We examined how validation design changes apparent prediction performance. Methods: Four BDHS rounds (2011-2022; 33,962 children; 1,290 deaths) were analysed with a 26-feature pipeline and three model classes under four validation regimes, including cross-survey temporal validation (train 2011+2014, calibrate 2017, test 2022). A 32-unit ELU multilayer perceptron was selected via genetic-algorithm neural architecture search. AUROC used 2,000 bootstrap resamples; screening utility used sensitivity, PPV, and number needed to screen (NNS) at fixed capacity. Results: Validation regime altered public-health interpretation more than model class. NAS MLP AUROC ranged from 0.669 (2022-only random) to 0.775 (pooled random), with temporal AUROC 0.730. At the top-10% temporal threshold, NAS identified 152/355 deaths in 2022 (sensitivity 42.8%, PPV 13.2%, NNS 7.6). NNS across designs ranged from 5.6 to 11.0. Conclusions: Validation-regime choice changed screening workload and apparent policy value more than architecture. Temporal validation supports defensible estimates of follow-up and referral demand; DHS child-mortality studies should report sensitivity, PPV, and NNS before programmatic use.

05.
arXiv (math.PR) 2026-06-12

Diffusion approximations for interacting stochastic systems with reflection and control

arXiv:2601.05895v2 Announce Type: replace Abstract: We study diffusion approximations for a class of interacting stochastic systems with reflection and control. Motivated by interacting stochastic dynamics subject to feedback mechanisms and boundary constraints, we consider diffusion-scaled stochastic processes incorporating stochastic fluctuations, state-dependent interactions, and reflection. Under suitable assumptions, we establish convergence in distribution of the scaled processes to systems of interacting reflected stochastic differential equations of Ornstein-Uhlenbeck type. The limiting dynamics capture key features of constrained multi-agent systems, including mean-reverting behavior, interaction effects, and confinement within bounded domains through Skorokhod reflection. The analysis combines diffusion-scaling arguments, stability estimates, and continuity properties of the Skorokhod map to connect discrete stochastic systems with their reflected diffusion limits. To illustrate the framework, we present numerical examples motivated by crowd dynamics and neural population dynamics. The simulations demonstrate qualitative agreement between the finite stochastic systems and the corresponding reflected diffusion models and illustrate how diffusion approximations can provide tractable descriptions of interacting stochastic systems with constraints.

06.
arXiv (CS.CV) 2026-06-15

Context-Guided Semantic Alignment for Feature Fusion Networks

Feature fusion networks are fundamental components in modern object detectors, aggregating multi-scale features to detect objects of varying sizes. However, directly fusing features from different pyramid levels often introduces semantic inconsistency due to their heterogeneous representations. In this paper, we propose Feature Interaction NEtwork (FINE), a lightweight semantic alignment module that refines low-level features via high-level contextual guidance using cross-level attention prior to fusion. To bridge the structural gap and ensure computational efficiency, we introduce an Alignment-Aware Token Sampling that aligns corresponding spatial regions across scales, reducing the attention complexity by an order of magnitude. The resulting attention weights generate a spatial-channel modulation map that is upsampled and applied to the low-level features via residual element-wise modulation. This mechanism ensures that the network selectively enhances semantically relevant pixels while preserving the sub-pixel localization accuracy necessary for dense prediction tasks. FINE is generally applicable to various detectors and consistently improves detection accuracy without compromising efficiency.

07.
arXiv (CS.AI) 2026-06-11

Architecture-Aware Reinforcement Learning Makes Sliding-Window Attention Competitive in Math Reasoning

arXiv:2606.11634v1 Announce Type: new Abstract: The rapid progress of reasoning and agentic large language models (LLMs) has increased the demand for long-context inference, but self-attention (SA) scales quadratically with context length. To address this, we study SWARR (Sliding-Window Attention with Reinforced Adaptation for Math Reasoning), a practical recipe for adapting SWA models to mathematical reasoning. SWARR has two stages: (1) efficient conversion from a pretrained SA model to SWA with supervised fine-tuning (SFT), which avoids pretraining a new base model, and (2) policy adaptation with reinforcement learning (RL). We find that SWA still underperforms SA after SFT, and we hypothesize that this gap is caused in part by a data-architecture mismatch: most SFT data are prepared for SA models and may contain long-range dependencies that are difficult for SWA to model. Because on-policy RL optimizes self-generated trajectories under the SWA constraint, it can adapt trajectories to better match SWA. Experiments on mathematical reasoning benchmarks show that this recipe substantially narrows the gap between SWA and SA, recovering much of the accuracy lost during SWA conversion while preserving the efficiency benefits of linear-complexity attention. Our central contribution is the empirical finding that RL changes the conclusion one would draw from conversion and SFT alone about SWA's viability for math reasoning.

08.
medRxiv (Medicine) 2026-06-15

Using wastewater surveillance to explore community-level dietary intake in sewered and non-sewered sanitation systems in Malawi, Africa

Wastewater can be used to measure biomarkers that reflect population-level dietary intake and diversity; however, how this approach may apply in a low-income country remains a knowledge gap. This study aims to evaluate whether select dietary-related metabolites can be detected in wastewater and environmental surveillance (WES) samples from both sewered and non-sewered sanitation systems in Malawi, Africa. Fourteen WES samples were collected and analyzed from two university campuses in Mzuzu and Thyolo, Malawi. Four targets were analyzed: N-methyl-2-pyridone-5-carboxamide (2PY; a biomarker of vitamin B3), 4-pyridoxic acid (4-PA; a biomarker of vitamin B6), as well as enterodiol and enterolactone (biomarkers of dietary fiber and polyphenol consumption). An 18-question survey, paired spatiotemporally with the WES measurements, assessed self-reported daily dietary intake, food insecurity, and nutrient deficiency symptoms among 500 respondents. Among the 14 WES samples, 2PY, 4-PA, and enterolactone were detected, while enterodiol was not detected above the method limit (

09.
arXiv (CS.CV) 2026-06-12

ReFoCUS: Reinforcement-guided Frame Optimization for Contextual Understanding

Recent progress in Large Multi-modal Models (LMMs) has enabled effective vision-language reasoning, yet the ability to video understanding remains constrained by suboptimal frame selection strategies, albeit with the rapid development of video-specialized LMMs. Prior works attempted to solve this with static heuristics or external retrieval modules to feed frame-level information, but these approaches often fail to capture visual cues grounded to the given user queries conflating raw visual dynamics with true semantic relevance. In this paper, we introduce ReFoCUS (Reinforcement-guided Frame Optimization for Contextual UnderStanding), the first framework to integrate online policy-gradient reinforcement learning into frame-level optimization for video-LLMs. ReFoCUS aims to learn a frame selection policy, leveraging reward signals derived from reference models to capture their underlying scoring behavior over frame combinations that best support temporally grounded responses. To efficiently explore the large combinatorial frame space, we employ an autoregressive and query-conditional selection architecture that ensures contextual consistency while reducing complexity. Our policy learning removes the need for explicit frame-level supervision, as it implicitly discovers optimal and semantically consistent frame compositions. ReFoCUS consistently improves reasoning accuracy across multiple video QA benchmarks, demonstrating the advantage of aligning frame selection with model-internal utility.

10.
arXiv (CS.AI) 2026-06-24

Sol Video Inference Engine: Agent-Native Full-Stack Acceleration Framework for Efficient Video Generation

arXiv:2606.23743v1 Announce Type: cross Abstract: Modern video diffusion models achieve higher generation quality through scaling, but this also increases inference cost. Although many acceleration methods have been proposed, a central challenge is that the most effective acceleration strategy is highly instance-specific: a recipe that works well for one combination of model, hardware, and inference configuration often does not transfer to another. Different models vary in architecture, numerical sensitivity, and attention concentration patterns. Inference settings differ in spatial and temporal resolution and video duration, while hardware platforms differ in memory hierarchy, supported numerical formats, and kernel throughput. These factors create a large tuning space, making manual performance engineering costly. We present Sol Video Inference Engine, an agentic, native, training-free acceleration framework for video diffusion models. It organizes five broadly applicable techniques, cache, sparse attention, token pruning, quantization, and kernel fusion, into an agentic acceleration stack for instance-specific optimization. For a concrete deployment target defined by a model, hardware platform, and serving configuration, parallel skill agents optimize the implementation of each technique, an agent integrator composes them into a global acceleration stack, and a human validator provides feedback on generation quality. We instantiate this workflow on three video models with different sizes and architectures: 64B Cosmos3-Super, 22B LTX-2.3, and 2B SANA-Video. With little human effort, the full stack achieves more than 2x end-to-end acceleration while maintaining near-lossless VBench quality, demonstrating the effectiveness of the agent framework for video diffusion acceleration.

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

DuoBench: A Reproducible Benchmark for Bimanual Manipulation in Simulation and the Real World

arXiv:2606.11901v1 Announce Type: cross Abstract: Bimanual robot systems substantially expand manipulation capabilities, but coordinating two arms introduces additional control complexity and failure modes that are not well captured by existing benchmarks. We introduce DuoBench, an extensible benchmarking framework for bimanual manipulation policies on the FR3 Duo platform. DuoBench comprises eleven tasks spanning four coordination categories, implemented in simulation and partially reproduced in the real world through reproducible task recipes with 3D-printable assets. In addition, we propose a stage-based evaluation scheme that supports fine-grained semantic failure analysis beyond binary success and provide human-teleoperated datasets for all benchmark tasks. We benchmark several dual-arm imitation-learning and vision-language-action policies in simulation and on real hardware. Our results show that current policies remain challenged by bimanual manipulation, particularly in early interaction stages, parallel arm execution, and transfer between simulation and real-world settings. DuoBench provides a reproducible testbed for diagnosing these failure modes and studying future methods for dual-arm policy learning. Code, datasets, and videos are available at https://duobench.github.io/

12.
arXiv (CS.CL) 2026-06-17

Regression Language Models for Code

We study code-to-metric regression: predicting numeric outcomes of code executions, a challenging task due to the open-ended nature of programming languages. While prior methods have resorted to heavy and domain-specific feature engineering, we show that a single unified Regression Language Model (RLM) using a frozen LLM encoder can simultaneously predict directly from text, (i) the memory footprint of code across multiple high-level languages such as Python and C++, (ii) the latency of Triton GPU kernels, and (iii) the accuracy and speed of trained neural networks represented in ONNX. In particular, a relatively small 300M parameter RLM based on T5Gemma, obtains >0.9 Spearman-rank on competitive programming submissions from APPS, and a single unified model achieves >0.5 average Spearman-rank across 24 different programming languages from CodeNet. Furthermore, the RLM can obtain the highest average Kendall-Tau of 0.46 on five classic NAS design spaces previously dominated by graph neural networks, and simultaneously predict architecture latencies on numerous hardware platforms.

14.
arXiv (CS.LG) 2026-06-11

Trajectory Geometry of Transformer Representations Across Layers

arXiv:2606.09287v2 Announce Type: replace Abstract: Understanding how transformer representations evolve across layers, not merely what they encode, remains an open problem in mechanistic interpretability. We recast the transformer forward pass as a discrete population trajectory through a high-dimensional representation manifold, drawing on geometric tools from computational neuroscience. Rather than probing for pre-specified features, we characterize trajectory geometry using five metrics computed directly in the ambient space: trajectory length, curvature, a semantic convergence index, layerwise cosine similarity, and representational stability. Across three model families (GPT-2, TinyLlama, Qwen2.5) and five controlled prompt families, we report four findings. First, semantically related prompts converge significantly in middle-to-late layers (peak CI 0.41–0.58, p

15.
medRxiv (Medicine) 2026-06-23

Uptake of minimal intervention dentistry among Romanian dental professionals and trainees: an exploratory cluster and network analysis

Background Minimal intervention dentistry (MID) is promoted as a prevention-oriented approach to caries management, but its integration into routine practice remains uneven. Existing research often examines MID-related knowledge, attitudes, or practices separately, offering limited insight into how these dimensions co-occur within individuals or are conditionally associated. Methods This exploratory cross-sectional survey examined multidimensional MID uptake among 327 Romanian dental students, residents, and specialists from five university centers. Ten MID-related scores were analyzed, including nine formative composites and one single-item peer-norm indicator. K-means clustering examined uptake profiles, and Gaussian graphical model network analysis with stepwise BIC selection examined conditional associations among constructs. Results A two-cluster solution was highly reproducible but modestly separated (n = 144 vs n = 183; average silhouette width = 0.13; mean Jaccard similarities = 0.92 and 0.94). The profiles reflected broadly lower versus higher uptake across knowledge-, belief-, and practice-related dimensions, while perceived peer norms for hygiene instruction showed the opposite pattern. Profile membership was not clearly patterned by gender, age band, professional status, or clinical experience. The primary network included 14 non-zero edges out of 36 possible edges, all positive; the strongest partial association linked diagnostic knowledge to diagnostic methods used in practice (partial r = .22). Familiarity, diagnostic knowledge, and general practices occupied more interconnected positions descriptively, but limited centrality stability precluded interpreting them as intervention targets. Conclusions MID uptake in this sample was better represented as a continuum of modestly differentiated profiles than as sharply separated participant types. The findings provide an exploratory map of multidimensional MID uptake and may inform future survey validation, implementation research, and dental education studies. Because the study was cross-sectional, convenience-sampled, and based on self-report, findings should be interpreted as hypothesis-generating rather than causal or population-representative.

16.
arXiv (CS.LG) 2026-06-19

Tracking Representation Dynamics in Large Language Models with Persistent Homology

arXiv:2606.19542v1 Announce Type: new Abstract: Large language models are commonly aligned through supervised fine-tuning, yet little is known about how their internal representations evolve during this process. We study alignment dynamics using persistent homology by tracking the topology of activation spaces throughout fine-tuning. Across four transformer language models ranging from 1B to 7B parameters and three alignment objectives corresponding to helpful, harmless, and mixed training data, we find that the majority of topological reorganization occurs during the earliest stages of training. A dense checkpoint analysis reveals a transient peak in topological activity followed by rapid stabilization. We further show that different alignment objectives induce distinguishable topological trajectories, while instruction-tuned and pretrained models exhibit qualitatively different patterns of evolution. Our results suggest that persistent homology provides a complementary perspective on alignment, revealing representation-level changes that are not apparent from behavioral metrics alone.

17.
arXiv (quant-ph) 2026-06-12

Continuum Neural Momentum Eigenstate for Variationally Solving Quasiparticles

arXiv:2606.12928v1 Announce Type: cross Abstract: We design the first neural quantum state for continuum particles that, for any chosen allowed momentum $\mathbf{k}$, is by construction an exact eigenstate of total momentum with eigenvalue $\mathbf{k}$. Our architecture, EVE, enables off-the-shelf VMC to solve for momentum-sector ground states. We test EVE on 2D bosons with mutual $1/r$ interactions, finding that a single unified ansatz is capable of describing four qualitatively different states: superfluid, roton, crystal, and phonon. At different densities, we extract the underlying phase of matter from the dispersion's shape. At $r_s = 20.0$, we see the roton minimum at finite $k$ expected of a superfluid. At $r_s = 100.0$, we see striking zone folding indicative of crystalline order, with periodically spaced minima representing floating crystals connected by phonon arcs in between. Using density-density correlation functions, we confirm the phase diagnoses and probe the excitations' correlation structures. Finally, we analyze the roton's phase texture and find unexpected multi-particle phase strings, formed when several vortex dipoles merge, leaving two vortices connected by a phase slip.

18.
arXiv (math.PR) 2026-06-12

Mixing times of one-sided $k$-transposition shuffles

arXiv:2112.05085v2 Announce Type: replace Abstract: We study mixing times of the one-sided $k$-transposition shuffle. We prove that this shuffle mixes relatively slowly, even for $k$ big. Using the recent ``lifting eigenvectors'' technique of Dieker and Saliola and applying the $\ell^2$ bound, we prove different mixing behaviors and explore the occurrence of cutoff depending on $k$.

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

LADBench: A Benchmark for Logical Fault Detection in Images

Large Vision Language Models (VLMs) excel at visual question answering and semantic grounding, but their capacity for autonomous logical reasoning remains underexplored. Existing anomaly benchmarks emphasize visual errors or direct prompting rather than the physical and social common sense needed for open-world deployment. To address this, we introduce LAD-bench, a benchmark of more than 1,000 curated synthetic images with logical anomalies across four domains: Residential, Urban, Collaborative, and Nature. We further propose a Tiered Prompting Protocol based on progressive disclosure, which measures how much explicit assistance a model needs to localize and reason about a logical fault. Evaluating leading foundation models reveals substantial weaknesses: even the best achieves only 70.11% overall accuracy, showing that implicit logical fault detection remains unsolved. Crucially, models often fail to identify anomalies even after receiving explicit hints in deeper tiers. By surfacing these limitations in sequential multimodal reasoning, LAD-Bench offers a rigorous framework for advancing the safety, reliability, and cognitive alignment of autonomous visual systems. Dataset and Code: https://huggingface.co/datasets/SahasraK/LADBench

20.
arXiv (CS.CL) 2026-06-11

When Does Language Matter? Multilingual Instructions Reveal Step-wise Language Sensitivity in Vision-Language-Action Models

Vision-Language-Action (VLA) models have shown strong performance in language-conditioned robotic manipulation, yet their robustness to linguistic variation remains poorly understood. In this work, we present the first systematic multilingual evaluation of VLA models by translating the LIBERO benchmark into ten languages, revealing severe performance degradation under non-English instructions, with success rates dropping by 30-50%. Through fine-grained analysis of task executions, we find that language influence is highly non-uniform across steps: certain steps exhibit strong language dependence and dominate overall task failure, while others are largely language-agnostic. Based on this insight, we propose a step-wise inference-time intervention that aligns representations according to step language sensitivity, substantially improving performance under linguistic variation. Our results indicate that language robustness in VLA models is fundamentally a step-wise control problem, highlighting the importance of temporally structured analysis for reliable embodied agents.

21.
PLOS Medicine 2026-06-16

The data transparency crisis in research: Lessons from systematic reviews and meta-analyses

by Saul Martin-Rodriguez, Rodrigo Fernandez-Gonzalo, David Moher Summary points Systematic reviews and meta-analyses underpin clinical guidelines and health policy, yet their validity may be compromised by limited access to underlying datasets and associated analytical code. Reliance on incomplete or inconsistently reported summary statistics forces researchers to use imputation and unverifiable assumptions, which can distort effect estimates and mislead clinical decision-making. The consequences extend beyond methodology: flawed evidence synthesis can influence treatment recommendations, healthcare spending, and patient safety, as illustrated by historical cases such as hormone replacement therapy. Despite widespread data-sharing policies, compliance remains low, enforcement weak, and monitoring almost non-existent, with many datasets remaining unavailable or inaccessible. This Policy Forum argues for strengthening enforceable data-sharing mechanisms, including clearer enforcement and pragmatic verification approaches within editorial workflows.

22.
arXiv (math.PR) 2026-06-15

Laws of Large Numbers for Non-Independent Random Variables on Hyperspaces with respect to the Hausdorff Metric

arXiv:2011.07199v5 Announce Type: replace Abstract: This paper investigates the limit behavior of the Minkowski sums for sequences of set-valued random variables. When the underlying space is finite dimensional, by using the support function, we establish the weak and strong laws of large numbers for non-independent random variables in the hyperspace with respect to the Hausdorff metric $d_H$.

23.
arXiv (CS.CL) 2026-06-18

Notation Matters: A Benchmark Study of Token-Optimized Formats in Agentic AI Systems

Large language models in Agentic AI systems consume tool schemas and execution results and emit tool invocations as structured data. The default language for that exchange, JSON, was designed for application-to-application interchange rather than token efficiency, so its structural elements impose substantial token overhead. Recent work proposes token-optimized alternatives such as TOON (Token-Oriented Object Notation) and TRON (Token Reduced Object Notation) as more compact replacements, but these formats have been evaluated only on isolated comprehension or generation tasks. Whether their token reductions hold inside end-to-end agentic loops therefore remains an open question. We evaluate TOON and TRON on four agentic benchmarks (BFCL, MCPToolBenchPP, MCP-Universe, StableToolBench) and five open-weight LLMs, decoupling input compression from output compression to measure comprehension and generation independently. TRON reduces tokens by up to 27% with accuracy within 14pp of the JSON baseline. TOON achieves up to 18% reduction at a similar 9pp accuracy cost, but additionally cascades on multi-turn parsing failures and collapses parallel tool-call output for most models. The code is available at: https://github.com/lkutschka/notation-matters

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

Latent Thought Flow: Efficient Latent Reasoning in Large Language Models

arXiv:2606.16222v1 Announce Type: new Abstract: Large Language Models (LLMs) increasingly rely on intermediate reasoning, yet explicit Chain-of-Thought (CoT) suffers from a linguistic space bottleneck: each thought must be decoded into tokens, causing high inference overhead. Latent reasoning moves deliberation into continuous space, but existing methods mostly learn deterministic or reward-maximizing paths, lacking a principled way to allocate probability across trajectories with different correctness and costs. We propose Latent Thought Flow (LTF), which models reasoning as variable-length continuous trajectories and trains a sampler to match a reward-induced posterior over answer quality and computation cost. We instantiate this with a continuous GFlowNet using stochastic latent transitions. To handle sparse answer supervision, we introduce an Entropy-Weighted Subtrajectory Balance objective for intermediate rewards and a reference-prior regularizer to anchor exploration. Experiments under finetuning and transfer learning settings show that LTF outperforms explicit CoT and latent reasoning baselines, improving accuracy by 9.5% while reducing reasoning length by 27.2% on average compared with strong latent reasoning baselines.

25.
arXiv (CS.AI) 2026-06-16

SorryDB: Can AI Provers Complete Real-World Lean Theorems?

arXiv:2603.02668v2 Announce Type: replace Abstract: We present SorryDB, a dynamically-updating benchmark of open Lean tasks drawn from 78 real world formalization projects on GitHub. Unlike existing static benchmarks, often composed of competition problems, hillclimbing the SorryDB benchmark will yield tools that are aligned to the community needs, more usable by mathematicians, and more capable of understanding complex dependencies. Moreover, by providing a continuously updated stream of tasks, SorryDB mitigates test-set contamination and offers a robust metric for an agent's ability to contribute to novel formal mathematics projects. We evaluate a collection of approaches, including generalist large language models, agentic approaches, and specialized symbolic provers, over a selected snapshot of 1000 tasks from SorryDB. We show that current approaches are complementary: even though an agentic approach based on Gemini Flash is the most performant, it is not strictly better than other off-the-shelf large-language models, specialized provers, or even a curated list of Lean tactics.