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.
medRxiv (Medicine) 2026-06-22

Longitudinal multi-omics characterization of the malignant evolution in multirelapsing glioblastoma

Linking glioblastoma (GBM) evolution to clinical progression is challenged by multiple factors, including tumor location for repeated sample collection, and short patient survival. In a single individual, we collected and analysed samples from 11 operations distributed across 31 months of multi-relapsing and multifocal GBM, including terminal leptomeningeal progression. All samples shared genomic ancestry of the retinoblastoma protein 1 (RB1) and neurofibromin 1 (NF1) mutations while advanced progression and extracranial metastases featured mutations of tuberous sclerosis complex 2 (TSC2), PBRM1, CD22 and Fanconi anemia supplementation group I (FANCI), correlated with clinical resistance to immunotherapies and DNA-damaging agents. Single-cell analytics revealed distinct yet reversible shifts in response to the precision medicine arsenal. GBM parenchymal dissemination and extracranial progression were associated with strengthening of neuron-like cell phenotypes. Our multidimensional study describes GBM evolution over a rarely reported time scale, and provides a valuable resource linking genetic, molecular, cellular and clinical progressions.

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

From Drift to Coherence: Stabilizing Beliefs in LLMs

arXiv:2606.17832v1 Announce Type: new Abstract: Large language models (LLMs) are often hypothesized to perform implicit Bayesian inference, yet a key coherence condition, the martingale property of predictive beliefs, has been shown to fail in controlled synthetic in-context learning settings. We revisit this question in a more typical usage regime: generic multiple-choice question answering. Exploiting the discrete answer space, we compute exact predictive distributions and study belief dynamics induced by autoregressive answer resampling. We introduce prompted predictive resampling (PPR), where an LLM generates a sequence of answers to the same question. Empirically, PPR reveals early-stage belief drift, indicating martingale violations. However, after sufficient resampling steps, the belief process self-stabilizes and converges to a coherent predictive distribution. Based on this observation, we further propose (i) a seed-answer prompting strategy to accelerate stabilization, and (ii) a self-consistency loss that amortizes early-stage drift into the model via fine-tuning. Experiments on multiple-choice QA benchmarks show that our methods substantially reduce belief drift and improve predictive coherence without sacrificing accuracy.

03.
arXiv (quant-ph) 2026-06-16

Semiclassical Gravity Efficiently Solves $\mathsf{NP}$-Complete Problems

arXiv:2606.14806v1 Announce Type: cross Abstract: Assuming the gravitational field is classical and that it couples to quantum fields via the semiclassical Einstein field equations, we show that the weak-field dynamics of a massive and non-relativistic qubit can in principle be used to solve an $\mathsf{NP}$-complete problem in polynomial time. We attribute this vast computational power to the non-linear dynamics afforded by the semiclassical Einstein field equations. Consequently, the above two assumptions entail a violation of the Physical Extended Church–Turing Thesis, which we regard as evidence for the quantization of gravity.

04.
arXiv (CS.AI) 2026-06-12

Teach-and-Repeat: Accurately Extracting Operational Knowledge from Mobile Screen Demonstrations to Empower GUI Agents

arXiv:2606.12817v1 Announce Type: new Abstract: Understanding the digital world on mobile devices is shifting from static UI perception to dynamic action comprehension. This capability enables models to convert visual state transitions into operational knowledge, defined as short natural-language sentences that describe action types, target UI elements, textual arguments, and execution orders. However, due to the highly diverse and heterogeneous UI designs across applications, existing vision-language models (VLMs) struggle to accurately infer these underlying operations. To bridge this gap, we introduce Teach VLM, a core model designed to translate mobile screen trajectories into step-wise operational knowledge by extracting and analyzing operation-related keyframes from demonstration videos. To address the scarcity of aligned training data, we develop a systematic data flywheel for scalable data acquisition. We further introduce a novel Chinese Mobile Screen Teach Benchmark for fine-grained evaluation. Building upon Teach VLM, we propose the Teach-and-Repeat paradigm, where the generated operational knowledge serves as an interpretable procedural reference to guide downstream screen-based execution agents. Extensive evaluations demonstrate that Teach VLM significantly outperforms strong VLM baselines, achieving state-of-the-art performance in operation semantics prediction. Furthermore, experiments in Android World show that our paradigm yields consistent Task Success Rate improvements for downstream agents. Together, Teach VLM and the Teach-and-Repeat paradigm offer a practical pathway from raw demonstrations to reusable task automation.

05.
medRxiv (Medicine) 2026-06-17

Long-term mortality and cause-specific death after non-cardiac chest pain: a multicentre cohort study of 160,245 patients in China

Abstract Background Non-cardiac chest pain (NCCP) is commonly regarded as a low-risk condition. However, long-term mortality, cause-specific death, and high-risk subgroup characteristics remain poorly defined. Methods In this multicentre registry-linked cohort study, we linked the Chest Pain Center Registry from 101 hospitals in Hunan, China, with the Mortality and Cause of Death Registry. Adults diagnosed with NCCP from Jan 1, 2017, to Dec 31, 2021, were included. We assessed 3-year all-cause, cardiovascular, and non-cardiovascular mortality using Cox, restricted cubic spline, and Fine-Gray models. Findings Among 160,245 patients, 4674 deaths occurred within 3 years (2.9%). Mortality increased sharply after 60.5 years. Age [≥] 60.5 years (adjusted hazard ratio [aHR] 7.49 [95% CI 6.89-8.14]), rural residence (time-varying aHR 1.46 [1.35-1.57] in year 1 and 1.66 [1.46-1.89] in years 1-3), and male sex (aHR 1.47 [1.38-1.57]) independently predicted death. Three-year mortality ranged from 0.3% in younger urban women to 8.4% in older rural men. Cardiovascular diseases accounted for 56.4% of deaths among older patients, whereas other non-cardiovascular causes (22.8%) and malignancy (20.8%) were the largest categories among younger decedents. Interpretation NCCP is not uniformly benign. Age, rural residence, and sex identify patients who could benefit from risk-stratified follow-up, with cardiovascular prevention prioritised for older rural men and broader non-cardiovascular assessment considered for younger patients.

06.
arXiv (CS.CV) 2026-06-24

What Do Flow-Based Inverse Solvers Approximate? A Posterior-Transport View

A growing family of training-free solvers – FlowDPS, FLOWER, PnP-Flow and their diffusion ancestors (DPS, DAPS) – repurpose a pretrained flow-matching prior to solve imaging inverse problems by adding a measurement-guidance term to the deterministic probability-flow ODE. Despite strong empirical results, what these per-step corrections actually approximate – and how far the resulting samples are from the true posterior $p(x\mid y)$ – has not been characterized. We give a posterior-transport account of flow-based inverse problem solving. Our starting point is a simple but consequential fact: for a deterministic flow prior, Bayesian conditioning is realized entirely by a reweighting of the source distribution, not by a drift correction; pushing the reweighted source through the unmodified velocity field yields exact posterior samples. From this we show that trajectory-guidance solvers can be read as the minimum-kinetic-energy correction field needed to morph the unconditional source into the posterior, and that FlowDPS / FLOWER / PnP-Flow correspond to distinct zeroth-order / Gaussian / proximal approximations of this single object; we bound the resulting posterior bias in Wasserstein distance. A controlled $2$D study with a closed-form posterior confirms the theory decisively: source reweighting matches the true posterior to the Monte-Carlo floor on every metric, whereas trajectory guidance incurs $200$–$800\times$ larger error and collapses posterior modes, regardless of guidance strength. Guided by the analysis we propose a cheap, principled velocity-correction solver that is competitive across two in-domain priors (AFHQ, CelebA) and two out-of-distribution settings while, unlike point-estimate source-space optimizers, producing diverse posterior samples with uncertainty that correlates with reconstruction error.

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

MobileFineTuner: A Mobile-Native Framework for On-Device LLM Fine-Tuning in Real-World Embedded AI Applications

arXiv:2512.08211v2 Announce Type: replace Abstract: Large language models (LLMs) are moving from cloud-centric services toward on-device embedded AI, where models interact with private, longitudinal signals sensed from users and their physical environments. Mobile phones are a natural platform for such applications because they are continuously carried by users, connected to wearable sensors, and deeply integrated with daily mobile applications. However, practical LLM fine-tuning on commodity phones remains difficult. Existing fine-tuning frameworks are largely Python-based and server-oriented, making them hard to deploy inside mobile applications. We present MobileFineTuner, a mobile-native open-source framework for end-to-end LLM fine-tuning on commodity mobile phones. MobileFineTuner is implemented in C++ and provides a reusable training stack. To make fine-tuning feasible under mobile resource constraints, MobileFineTuner integrates a resource-aware training runtime with memory-efficient attention, activation checkpointing, gradient accumulation, parameter sharding, and energy-aware scheduling. We evaluate MobileFineTuner on real mobile phones using GPT-2, Gemma 3, and Qwen2.5 models across multiple fine-tuning tasks. The results show that MobileFineTuner reproduces standard Full-FT and LoRA fine-tuning behavior, substantially reduces memory pressure and improves executability on memory-constrained phones. We further demonstrate MobileFineTuner through a private campus health-agent application, where a local LLM is fine-tuned on user-specific wearable-sensing records to provide more personalized responses while keeping raw records on the phone. These results establish MobileFineTuner as a practical toolkit for studying and building on-device LLM fine-tuning applications in embedded AI and sensing systems.

08.
arXiv (CS.CV) 2026-06-18

Structured Spectral Graph Representation Learning for Multi-label Abnormality Analysis from 3D CT Scans

With the growing volume of CT examinations, there is an increasing demand for automated tools such as organ segmentation, abnormality detection, and report generation to support radiologists in managing their clinical workload. Multi-label classification of 3D Chest CT scans remains a critical yet challenging problem due to the complex spatial relationships inherent in volumetric data and the wide variability of abnormalities. Existing methods based on 3D convolutional neural networks struggle to capture long-range dependencies, while Vision Transformers often require extensive pre-training on large-scale, domain-specific datasets to perform competitively. In this work, we propose a 2.5D alternative by introducing a new graph-based framework that represents 3D CT volumes as structured graphs, where axial slice triplets serve as nodes processed through spectral graph convolution, enabling the model to reason over inter-slice dependencies while maintaining complexity compatible with clinical deployment. Our method, trained and evaluated on 3 datasets from independent institutions, achieves strong cross-dataset generalization, and shows competitive performance compared to state-of-the-art visual encoders. We further conduct comprehensive ablation studies to evaluate the impact of various aggregation strategies, edge-weighting schemes, and graph connectivity patterns. Additionally, we demonstrate the broader applicability of our approach through transfer experiments on automated radiology report generation and abdominal CT data.

09.
arXiv (CS.CV) 2026-06-16

LOCUS: Local Visual Cue Search for Enhancing Fine-Grained Perception in Multimodal Large Language Models

Multimodal Large Language Models (MLLMs) remain unreliable on fine-grained visual perception, even when high-resolution inputs preserve the necessary local details. We identify this limitation as visual context rot: decisive evidence may exist in the full image, yet fail to be reliably selected and used amid redundant visual context. We propose LOCUS (LOcal visual CUe Search), a training framework that teaches MLLMs to internalize local evidence search through a verifiable proxy task. During training, LOCUS provides a local crop as a visual cue and optimizes the model to recover its spatial support in the full image using an IoU-based reward. The visual cue is used only during training, leaving the standard image-question inference interface unchanged. Experiments across fine-grained perception, hallucination, general understanding, and reasoning benchmarks show that LOCUS improves localization-sensitive visual understanding while preserving broad capabilities. Attention analyses further indicate stronger focus on task-relevant evidence regions, suggesting that training-time visual cue search provides an effective route to internalized fine-grained evidence selection.

10.
Nature (Science) 2026-06-18

Daily briefing: The proteins that protect us from deadly mutations

作者:

Proteins that ‘buffer’ the effects of mutations could help to treat diseases such as cancers. Plus, goats can follow human voices and the battle over a key ocean observatory project in the United States. Proteins that ‘buffer’ the effects of mutations could help to treat diseases such as cancers. Plus, goats can follow human voices and the battle over a key ocean observatory project in the United States.

11.
Nature (Science) 2026-06-22

Isotopic evidence for a cold and distant origin of 3I/ATLAS

Interstellar objects provide the only directly observable samples of icy planetesimals formed around other stars, and can therefore provide insight into the diversity of physical and chemical conditions occurring during exoplanet formation1−3. Here we report isotopic measurements of the interstellar comet 3I/ATLAS, which reveal an elemental composition unlike any Solar System body. The water in 3I/ATLAS is enriched in deuterium, at a level of D/H = (0.98 ± 0.06)%, which is more than an order of magnitude higher than in known comets, while its range of 12C/13C ratios (141–191 for CO2 and 123–172 for CO) exceeds typical values found in the Solar System, as well as nearby interstellar clouds and protoplanetary disks. Such extreme isotopic signatures indicate formation at temperatures  ≲ 30 K in a relatively metal-poor environment. When interpreted with respect to models for Galactic chemical evolution, the carbon isotopic composition implies that 3I/ATLAS may have accreted as long ago as 12 billion years, following a period of intense, early star formation. 3I/ATLAS thus represents a preserved fragment of an ancient planetary system.

12.
arXiv (CS.AI) 2026-06-16

Defending against Adaptive Prompt Injection Attacks via Reasoning-enabled Task Alignment

arXiv:2606.15441v1 Announce Type: cross Abstract: Indirect prompt injection attacks hijack LLM-based agents by embedding malicious instructions in third-party data that the agent retrieves during task execution. Existing defenses report near-zero attack success rate on static benchmarks, yet recent adaptive evaluations show that these results collapse once the attacker is allowed to optimize against the deployed defense. In this work, we trace this collapse to two failure modes. First, existing defense methods are confined to recognizing specific attack patterns, rather than assessing whether the intent of every embedded instruction is relevant to the user task. Second, training-based defenses, which otherwise offer the strongest safety-utility trade-off, assemble their adversarial examples from a handful of hand-crafted templates, and the resulting defender fails to generalize outside that narrow strategy distribution. To address these gaps, we propose RETA, a training-based method that grounds defense decisions on the user tasks rather than attacker-controlled data. At each tool-output step, the defender undertakes chain-of-thought reasoning verifying that its actions are consistent with the user task. Leveraging red-teaming, a simulated attacker synthesizes adversarial training data and receives a dictionary-learning diversity reward, achieving broad coverage of injection-reformulation strategies. Together, these allow the defender to be optimized via multi-objective reinforcement learning and achieve better safety-utility trade-off. Across six black-box adaptive attacks, RETA keeps every per-attack ASR below 10%, with average ASR of 2.92% and 3.75% on the two target models, while preserving most utility under attack and on clean inputs.

13.
arXiv (quant-ph) 2026-06-16

Forbidden transitions in superconducting artificial atoms

arXiv:2606.06069v2 Announce Type: replace Abstract: Artificial atoms built from Josephson junctions have become a powerful tool to explore the limits of quantum optics due to their strong coupling to electromagnetic fields and their sensitivity to changes at the single-photon level. This sensitivity to quantum fluctuations complements their metrological and computational use, which are based on the precise oscillating frequency of the underlying supercurrents. We present here a theory for Josephson junctions immersed in electromagnetic fields where focus is shifted from temporal correlations and towards spatial ones. Unlike the commonly used circuit and black-box descriptions, our work is based on a microscopic model that enables systematically accounting for the effect of the spatial and vectorial profile of an electromagnetic field over a junction. As an example of the interactions that emerge in such a setup, we investigate the possibility of driving a junction via a quadrupole transition, using typical experimental parameters in existing devices. With the transition being dependent on the gradient of the electric field – rather than its intensity – the junction can be excited in a region where the electric field vanishes.

14.
arXiv (CS.LG) 2026-06-18

Dimension-Free Convergence of Discrete Diffusion Models: Adjoint Equations Induce the Right Space

arXiv:2605.17232v2 Announce Type: replace Abstract: Discrete diffusion has become a leading framework for generative modeling in various applications including language, vision, and biology. Existing convergence theory, however, exhibits fundamental limitations. KL-based analyses diverge under singular priors such as the masked distribution, while bounds in total variation (TV) depend on the state space size $S$ and become vacuous for modern language tasks, where vocabularies contain hundreds of thousands of tokens. We develop a unified adjoint-equation-based framework that establishes dimension-free convergence guarantees in any integral probability metric (IPM). To the best of our knowledge, our bounds are the first to be entirely free of $S$ and applicable to both masked and uniform priors. Importantly, our theory relies only on a single standard rate-matrix regularity assumption and applies to general priors. Five novel techniques drive our improvements: working in the space of observables via adjoint equations rather than directly with probability measures, a regularity analysis that yields bounds on any IPM, a coupling argument that removes $S$-dependence under uniform transitions, and score-marginal cancellation and exit-routing techniques that remove $S$-dependence under masked transitions. Our framework thus sharply departs from prior analyses and avoids the shortcomings of pathspace-KL and existing TV-based approaches. Beyond convergence bounds, our framework provides a versatile toolkit for further theoretical study of discrete diffusion models, including principled choices of loss functions and dimension-free step complexity.

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

BCL: Bayesian In-Context Learning Framework for Information Extraction

Existing information extraction (IE) tasks increasingly adopt in-context learning (ICL) with large language models. However, current approaches either show inconsistent performance across model scales or lack systematic optimization and generalizability. Building on this, we propose BCL (Bayesian In-Context Learning Framework for Information Extraction), the first optimization framework that uses particle filtering with Bayesian updates to systematically refine label representations across IE tasks. Through four steps initialization, observation, weight update, and resampling, BCL generalizes to both sequence labeling and relation classification paradigms. Extensive experiments demonstrate substantial and consistent improvements over existing approaches.

16.
arXiv (CS.AI) 2026-06-12

Valid Inference with Synthetic Data via Task Exchangeability

arXiv:2606.13629v1 Announce Type: cross Abstract: There is a proliferation of work arguing for the use of synthetic data in scientific research. For example, social scientists are arguing for the use of LLM-generated "silicon samples" in pilot studies; AI evaluations increasingly rely on "LLM-as-a-judge" outputs; and proteomics research is accelerated by generative models that produce synthetic protein structures. These developments raise an intriguing possibility: synthetic data may help researchers ask more questions, run more studies, and accelerate discovery. But they also raise a fundamental concern: synthetic data can be biased, noisy, and misspecified. In this work, we propose statistical principles for using synthetic data in scientific research with provable validity guarantees. The key insight is a new technical condition that we call task exchangeability. Informally, this is a requirement that the researcher can identify historical tasks, for which real data is available, such that their current task of interest is exchangeable with the historical tasks in an appropriate mathematical sense. We develop methods for valid inference under task exchangeability, together with extensions that provide guarantees even beyond exchangeability. We demonstrate the framework on public opinion surveys with silicon samples and AI evaluation with autoraters.

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

The Standard Model, The Exceptional Jordan Algebra, and Triality

作者:

arXiv:2006.16265v2 Announce Type: replace-cross Abstract: Jordan, Wigner and von Neumann classified the possible algebras of quantum mechanical observables, and found they fell into 4 "ordinary" families, plus one remarkable outlier: the exceptional Jordan algebra. We point out an intriguing relationship between the complexification of this algebra and the standard model of particle physics, its minimal left-right-symmetric $SU(3)\times SU(2)_{L}\times SU(2)_{R}\times U(1)$ extension, and $Spin(10)$ unification. This suggests a geometric interpretation, where a single generation of standard model fermions is described by the tangent space $(\mathbb{C}\otimes\mathbb{O})^{2}$ of the complex octonionic projective plane, and the existence of three generations is related to $SO(8)$ triality.

18.
arXiv (CS.AI) 2026-06-24

InSight: Self-Guided Skill Acquisition via Steerable VLAs

arXiv:2606.24884v1 Announce Type: cross Abstract: Vision-language-action (VLA) models can learn manipulation skills from demonstrations, but their capabilities are bounded by the skills in the training data. We present InSight, a framework that unlocks autonomous skill acquisition by rendering VLAs steerable at the primitive-action level (e.g., "move gripper to the bowl", "lift upward", "pour the bottle"). InSight consists of two primary stages: (1) an automated segmentation pipeline that partitions demonstrations into labeled primitives via VLM plan decomposition and end-effector poses to enable VLA primitive steerability, and (2) a VLM-guided data flywheel that identifies missing primitives required to accomplish a novel task, autonomously attempts demonstrations of the missing primitives with VLM-proposed low-level control, and automatically labels, stores, and integrates successful demonstrations into the VLA training set. We evaluate InSight across simulation and real-world manipulation tasks, including block flipping, drawer closing, sweeping, twisting, and pouring, without any human demonstrations of these target skills. Once learned, these primitives can be composed to execute novel, long-horizon tasks without additional human demonstrations. Our findings demonstrate that primitive steerability provides a practical foundation for continual skill acquisition in VLA policies. Project website: https://insight-vla.github.io.

19.
medRxiv (Medicine) 2026-06-22

AFFORDABILITY OF INTOXICATION FROM CHEAP ETHANOL: EVIDENCE FROM RETAIL ALCOHOL MARKETS IN UGANDA

Background: Alcohol affordability is a determinant of consumption and alcohol-related harm. In many low- and middle-income countries (LMICs), informal production, variable alcohol strength, and non-standard packaging complicate conventional affordability measures, limiting evidence on the economic accessibility of alcohol and the cost of intoxication. Objective: To assess the affordability of intoxication in Uganda by estimating the cost of obtaining ethanol to reach intoxication across alcohol products, packaging types, and retail contexts. Methods: Data were collected on 824 alcoholic beverages from urban, rural, and urban-slum retail markets. Ethanol-standardized pricing (price per gram of alcohol) was calculated, and the cost of consuming 60 g of ethanol was estimated. Multivariate regression identified determinants of ethanol affordability. Results: Affordability varied by product type and packaging. Opaque beers and illicit spirits provided the cheapest pathways to intoxication, with median costs of UGX 1,200-1,500 per 60 g of ethanol. Plastic packaging was associated with lower ethanol costs than glass packaging. Ethanol prices differed across formal and informal markets (p < 0.01), while rural areas and urban informal settlements had 20-25% lower costs than urban areas. Regulatory status alone did not predict affordability. Conclusions: In Ugandas diverse alcohol market, affordability is driven by access to ethanol rather than beverage price alone. Low-cost, high-strength alcohol sold through informal channels enables intoxication at minimal expense, among disadvantaged populations. Implications: Alcohol policies should target ethanol content through minimum unit pricing, alcohol-content-based taxation, and regulation of informal markets and packaging practices to reduce harmful consumption and inequities.

20.
arXiv (CS.LG) 2026-06-18

Signature filtering: a lightweight enhancement for statistical watermark detection in large language models

arXiv:2606.18430v1 Announce Type: new Abstract: Statistical watermarks help organizations attribute large language model (LLM) outputs, yet existing detectors often struggle when watermark signals are weak, texts are repetitive, or watermarks are edited. We propose signature filtering, a detection-time module that enhances watermark detection without modifying watermark embedding and text generation. It learns a small set of ``signature'' tokens whose presence makes watermark tests unreliable, and removes these tokens before detection. The signatures are obtained by solving a mixed-integer linear program on a small training set, with constraints that maximize the true positive rate. We additionally derive finite-sample and asymptotic bounds under several attacker models (color-blind, color-adaptive, and distributionally correlated). On four well-known watermark families (Kgw, Sweet, Unigram, Exp), four benchmark corpora (C4, MBPP, HumanEval, Code-Search-Net), and six LLMs (Opt-1.3b, Opt-6.7b, Llama2-13b, Llama3.1-8b, Qwen2.5-14b, Phi-3-medium-14b), 2- and 3-gram signatures raise detection rates in weak-signal and low-entropy settings from 8~31% without filtering to 78~99% with filtering, while keeping false positives controllable and often negligible. In stress tests where we scramble sentences and perturb 25~50% of tokens by dilution, deletions, and substitutions, 2-gram filters for Kgw-style watermarks preserve most of the clean-text detection gains, often matching or outperforming the advanced WinMax watermark detector. Signature filtering thus provides a simple, scalable, and model-agnostic add-on to strengthen watermark-based provenance checks for LLM text in information processing workflows.

21.
arXiv (CS.AI) 2026-06-16

TERMS-Bench: Diagnosing LLM Negotiation Agents Beyond Deal Rate

arXiv:2605.13909v2 Announce Type: replace-cross Abstract: Negotiation is a central mechanism of economic exchange, shaping markets, procurement, labor agreements, and resource allocation. It is also a canonical testbed for agentic language models, requiring multi-turn interaction under hidden preferences, strategic communication, and binding constraints. These properties make negotiation hard to evaluate: unlike math or code, it has no intrinsic verifier. Existing LLM negotiation evaluations rely on LLM-vs.-LLM interaction or aggregate outcomes such as deal rate, leaving failures opaque. We introduce Terms-Bench, short for Testbed for Economic Reasoning in Multi-turn Strategy, a Bayesian-game framework that makes the environment itself the verifier by specifying the counterpart's latent type, policy, and payoff structure. We instantiate it in bilateral price negotiation, where the counterpart's private state and simulator policy are hidden from the agent but observable to the evaluator. This turns the counterpart from a black-box opponent into a diagnostic instrument, enabling agent-attributable failure analysis and oracle-reference optimality gaps. Evaluating 13 LLM agents spanning frontier systems from major providers, Terms-Bench turns negotiation evaluation from aggregate ranking into actionable diagnosis: where agents fail, why they fail, and what to strengthen. Empirically, frontier models saturate deal rate yet diverge in surplus extraction, cue use, belief calibration, and compliance, revealing agent-specific bargaining bottlenecks masked by prior benchmarks.

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

CoCoGEC: Counterfactual Generation for Robust Grammatical Error Correction

Grammatical error correction (GEC) systems are usually trained and evaluated on GEC benchmarks, but their performance often drops sharply once the surrounding context is slightly perturbed or extended. This indicates that the existing GEC models usually fail to understand the error patterns in the varying contexts. In this paper, we thoroughly investigate the counterfactuals for GEC tasks, where the subtle changes to the contexts could lead to the label flipping issue. We propose CoCoGEC, a counterfactual generation framework that creates copies of training instances with error-irrelevant contexts altered. Our framework systematically generates counterfactuals by (1) generating intra- and inter-sentence counterfactuals that maintain the error patterns as well as syntax of the original instances by altering the word-level and sentence-level contexts; (2) revising the generated counterfactuals by selecting the instances with flipped labels and high GEC Mutual Information (MI) coefficient. Extensive experiments show that our method substantially improves the stability of GEC models, outperforming a set of data augmentation baselines. Particularly, it could achieve absolute F0.5 gains of +9.9, +11.3, and +20.8 points on the perturbed BEA-19*,CoNLL-14*, and TEM-8* data set.Our code is released at https://github.com/Quinnok/CoCoGEC

23.
arXiv (CS.AI) 2026-06-24

An Introduction to Causal Reinforcement Learning

arXiv:2606.24160v1 Announce Type: new Abstract: Causal inference provides a set of principles and tools that allow one to combine data and knowledge about an environment to reason with questions of counterfactual nature, i.e., what would have happened had reality been different, even when no data of this unrealized reality is currently available. Reinforcement learning provides methods to learn a policy that optimizes a specific measure (e.g., reward, regret) when the agent is deployed in an environment and pursues an exploratory, trial-and-error approach. These two disciplines have evolved independently and with virtually no interaction between them. We note that they operate over different aspects of the same building block, counterfactual relations, which makes them umbilically connected. Based on these observations, novel learning opportunities arise when this connection is explicitly acknowledged and mathematized. To realize this potential, we note that any environment where the RL agent is deployed can be decomposed as a collection of autonomous mechanisms with different causal invariances, parsimoniously modeled as a structural causal model; any standard RL setting implicitly encodes such a model. This formalization allows us to put under a unifying treatment different modes of learning, including online, off-policy, and causal calculus learning, which appear unrelated in the literature. However, these modalities are not exhaustive: we introduce several natural and pervasive classes of learning settings that entail novel dimensions of analysis. Specifically, we introduce and discuss through causal lenses generalized policy learning, where to intervene, imitation learning, and counterfactual learning. These tasks lead to a broader view of counterfactual learning and suggest great potential for studying causal inference and reinforcement learning side by side, which we call causal reinforcement learning (CRL).

24.
arXiv (CS.LG) 2026-06-11

Bootstrapped Monitoring: Leveraging Transparent Reasoning to Oversee Stronger AI Agents

arXiv:2606.11998v1 Announce Type: new Abstract: Trusted monitoring is a cornerstone of AI control. However, as frontier models grow more capable, the increasing capabilities gap between trusted and untrusted models may render trusted models unreliable monitors. We introduce bootstrapped monitoring, a protocol that addresses this by inserting a stronger, intermediate untrusted model with transparent chain-of-thought reasoning into the oversight chain. The untrusted monitor ($U_m$) evaluates the agent's actions, while a weaker trusted model ($T$) oversees $U_m$'s reasoning to detect collusion. We evaluate bootstrapped monitoring on multi-turn software engineering tasks (BashArena) across multiple agents and monitors. Bootstrapped monitoring substantially improves catch rates over trusted-only monitoring, even when the untrusted monitor actively colludes with the agent, provided we have access to its raw chain-of-thought. Our results suggest that bootstrapped monitoring can extend the useful lifetime of trusted models in control as AI capabilities advance.

25.
arXiv (CS.CV) 2026-06-18

Beyond the Current Observation: Evaluating Multimodal Large Language Models in Controllable Non-Markov Games

Deploying multimodal foundation models as closed-loop policies increasingly requires conditioning actions on observations that are no longer visible. However, existing benchmarks either expose the full state, conflate hidden-state reconstruction with other agent skills, or test recall only after an episode has ended. We introduce RNG-Bench (Reconstructive Non-Markov Games), a benchmark suite designed to isolate a base model's ability to reconstruct past observations and act on them during multi-step interaction. RNG-Bench includes two complementary games: Matching Pairs, where card identities briefly revealed at specific locations must later be recalled, and 3D Maze, where egocentric views must be integrated into a spatial map. Both games are evaluated under a unified harness with three controlled difficulty axes: grid size, visual pattern, and observation modality. The benchmark further introduces a head-to-head duel protocol to control for instance-level variance and a Memory Gap metric that disentangles forgetting from poor action selection. The hardest configurations require contexts of roughly 128K tokens and 350 image inputs per episode, and remain far from saturated by frontier MLLMs. Memory Gap analysis shows that most residual errors stem from forgetting earlier observations rather than from suboptimal decision making. Finally, fine-tuning Qwen3.5-9B on optimal-policy rollouts and filtered model demonstrations improves performance on RNG-Bench and transfers to existing benchmarks without degrading general multimodal capability.