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.
PLOS Computational Biology 2026-06-17

Deciphering cell type-specific causal genetic effects on brain imaging-derived phenotypes and disorders with single-cell Mendelian randomization

作者:

by Anyi Yang, Xingzhong Zhao, Xing-Ming Zhao, Yucheng T. Yang Reconstructing causality routes from genetic effects to complex phenotypes in particular cell types is crucial for understanding biological mechanisms underlying the brain-associated phenotypes including imaging-derived phenotypes (IDPs), and brain disorders and behaviors (DBs). Here, we develop a single-cell Mendelian randomization framework to infer cell type-specific causal relationships between gene expression and diverse brain-associated complex phenotypes by integrating single-cell expression quantitative trait loci (cis-eQTLs) and genome-wide association study findings. We identifiy a set of 254 and 217 cis-eQTL target genes (eGenes) that may have causal effects on 112 IDPs and 26 DBs in eight cell types, respectively. These causal eGenes exhibit strong cell type specificity and varied pleiotropy among different types of brain-associated phenotypes. Further integrative analysis reveals putative causality routes among cell type-specific causal eGenes and brain-associated complex phenotypes. Finally, we characterize the spatiotemporal expression patterns of these causal eGenes, and highlight the coordinated associations of the brain-associated phenotypes based on the expression of their causal eGenes. Overall, our study presents a large-scale analysis of the genetic effects of brain structures, disorders and behaviors, providing a catalog of cell type-specific causal eGenes.

02.
arXiv (CS.LG) 2026-06-15

Recovery thresholds for hidden weighted sparse graphs

arXiv:2606.14335v1 Announce Type: cross Abstract: Recovering structural information from noisy high-dimensional data is a fundamental task in statistical inference. We investigate the recovery thresholds for a graph hidden in a randomly weighted complete graph. Specifically, an unknown graph $H^* \in H_n$ is chosen uniformly at random, and hidden in a complete graph of $n$ vertices as follows: the weight of an edge $e \in H$ is distributed independently according to $P_n$; otherwise the weight is distributed independently according to $Q_n$. The goal is to recover almost all of $H$ from these edge weights. Assuming a local Lipschitzness of the Rényi divergence between distributions $P_n$ and $Q_n$, and a mild density condition for the graphs $H_n$, we give a unified characterization of the information-theoretic limit for recovering almost all of $H$ (also known as almost exact recovery). Our characterization connects the KL divergence between $P_n$ and $Q_n$ to the logarithm of the first moment threshold of $H$ in the Erdős-Rényi random graph model $G(n,p)$. Our lower bound also extends to the task of partial recovery, in which only a constant $\lambda$-fraction of $H$ needs to be recovered. Last but not least, for certain Bernoulli and Exponential regimes, and for Gaussian distributions, we are able to show an All-or-Nothing (AoN) threshold phenomenon at the exponential scale.

03.
arXiv (CS.AI) 2026-06-15

StreamMemBench: Streaming Evaluation of Agent Memory for Future-Oriented Assistance

arXiv:2606.14571v1 Announce Type: new Abstract: A central role of personal-agent memory is to turn stored information and prior interactions into future-oriented assistance. In daily use, useful cues come from what the agent observes and how the user interacts with the agent, and the agent must carry them forward from the current request to similar future tasks. Existing memory benchmarks usually test dialogue recall or task improvement in isolation, leaving the trajectory from streaming observations to later assistance largely untested. We introduce StreamMemBench, a streaming benchmark that constructs a two-step task sequence around each evidence anchor from EgoLife egocentric streams. The initial task tests evidence use, while the follow-up task tests whether feedback and interaction experience are reused. Four metrics diagnose evidence recall, initial evidence use, feedback incorporation, and follow-up reuse. Experiments with eight memory systems across two backbones show that current systems often fail to use observed evidence or turn feedback into reliable follow-up behavior, even when evidence is stored or feedback is incorporated locally. StreamMemBench is publicly available at https://github.com/landian60/StreamMemBench.

04.
arXiv (quant-ph) 2026-06-12

Quasi-local Edge Mode in XXX Spin Chain/Circuit with Interaction Boundary Defect

arXiv:2603.17835v2 Announce Type: replace-cross Abstract: We study the Heisenberg spin-1/2 model on a semi-infinite chain - or, equivalently, a trotterized unitary SU(2) symmetric six-vertex quantum circuit - with a boundary defect where the interaction between the two spins nearest the edge differs from that in the bulk. For sufficiently strong boundary interaction we explicitly construct a conserved operator quasi-localized near the boundary using a matrix-product ansatz. This quasi-local edge mode leads to non-decaying boundary correlation functions, corresponding to a nonzero boundary Drude weight. The correlation length of the edge mode diverges at a finite critical value of the boundary interaction, signaling a transition to ergodic boundary dynamics for subcritical interactions.

05.
arXiv (CS.CV) 2026-06-12

Why Commodity WiFi Sensors Fail at Multi-Person Gait Identification: A Systematic Analysis Using ESP32

WiFi Channel State Information (CSI) has shown promise for single-person gait identification, raising interest in its use for contactless biometrics, continuous authentication, and passive identification. However, the feasibility of multi-person identification on low-cost commodity devices remains unclear. A critical question is whether weak multi-person performance is primarily an algorithmic limitation, or whether it reflects a more fundamental sensing ceiling on commodity WiFi hardware. We address this question through a systematic empirical study using commodity ESP32 WiFi sensors. We evaluated six different signal separation methods–FastICA, SOBI, PCA-ICA, NMF, Wavelet, and Tensor decomposition–across seven scenarios spanning 1-10 people in both controlled and realistic indoor environments. To investigate beyond classification accuracy, we introduce three diagnostic metrics: intra-subject variability (ISV), inter-subject distinguishability (ISD), and performance degradation rate (PDR). In all methods, performance remains moderate (39%-56% accuracy), with limited evidence that algorithmic choice alone solves the problem. The best-performing method, NMF, reaches 56% accuracy, while all methods exhibit extremely high feature-space overlap (97%-99%), unstable within-subject representations, and marked environmental sensitivity. These findings suggest that, under commodity ESP32 CSI constraints, dense multi-person gait identification is limited more by sensing quality and spatial diversity than by the chosen separation algorithm. Our results have direct implications for security and privacy: they call into question the practicality of commodity WiFi CSI as a robust multi-user biometric primitive for authentication, while also placing important bounds on the passive identification capabilities achievable with low-cost off-the-shelf WiFi hardware.

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

GD$^2$PO: Mitigating Multi-Reward Conflicts via Group-Dynamic reward-Decoupled Policy Optimization

arXiv:2606.16771v1 Announce Type: new Abstract: As LLMs advance, post-training reinforcement learning (RL) increasingly relies on multi-dimensional rewards to cultivate comprehensive capabilities. This shift demands new algorithms capable of optimizing diverse and potentially competing objectives simultaneously. To address this, existing methods such as Group reward-Decoupled Policy Optimization (GDPO) decompose the overall score into independent reward groups, then compute the RL loss separately within each group. However, this strategy still encounters multi-reward conflicts: a single rollout can yield positive advantages on certain reward dimensions but negative ones on others, causing opposing signals to cancel each other out during aggregation, further hindering RL training efficiency. Inspired by Dynamic sAmpling Policy Optimization (DAPO), which improves RL training efficiency by filtering out ineffective rollouts with near-zero advantages, we propose Group-Dynamic reward-Decoupled Policy Optimization (GD$^2$PO). Specifically, GD$^2$PO employs a conflict-aware filtering mechanism to mask out rollouts suffering from severe reward-wise disagreement. By preventing conflicting signals from canceling each other out, this masking strategy preserves and enhances the magnitude of effective RL advantages, thereby significantly accelerating learning efficiency. Furthermore, we introduce query-level reweighting to dynamically adjust the update intensity of each query based on its overall reward consensus. Experiments on various multi-reward scenarios, including tool calling and human preference alignment, demonstrate that GD$^2$PO consistently and significantly outperforms existing baselines. The code is available at https://github.com/Qwen-Applications/GD2PO.

07.
arXiv (CS.AI) 2026-06-12

How AI Agents Reshape Knowledge Work: Autonomy, Efficiency, and Scope

arXiv:2606.07489v2 Announce Type: replace Abstract: Frontier AI systems are bridging the gap between intelligence and utility by shifting from conversational assistants to autonomous agents that execute tasks end to end. Using production data from Perplexity's Search and Computer products, we study this transition by examining how AI agents accelerate and reshape knowledge work. Three key empirical findings emerge. First, using sessions with near-identical initial query pairs as natural experiments for the same underlying task attempted with both products, Computer performs 26 minutes of autonomous work per user session, versus 33 seconds for Search. Computer automates task decomposition and execution that Search users might otherwise manually orchestrate and implement. As a result, Computer shifts follow-up query distribution toward higher-order work such as verification and extension. Autonomy also increases execution quality, with per-query dissatisfaction rates 55% lower on Computer than on Search. Second, due to its autonomy advantage, Computer reduces completion time from 269 to 36 minutes on matched tasks, lowering estimated time and cost by 87% and 94%, respectively, compared to humans equipped with Search alone. Third, Computer changes the scope of work that users attempt: Computer queries more often cross occupational boundaries, require higher-order cognition, draw on broader expertise, take the form of composite tasks that bundle interdependent subtasks into a single query, and unlock work activities that are essentially absent from Search usage among the same users. Together, the evidence indicates that AI agents accelerate workflows, enhance output quality, reduce costs, and expand the breadth and depth of automated work.

08.
arXiv (CS.AI) 2026-06-16

RaBiT: Residual-Aware Binarization Training for Accurate and Efficient LLMs

arXiv:2602.05367v3 Announce Type: replace Abstract: Efficient deployment of large language models (LLMs) requires extreme quantization, forcing a critical trade-off between low-bit efficiency and performance. Residual binarization enables hardware-friendly, matmul-free inference by stacking binary ($\pm$1) layers, but is plagued by pathological feature co-adaptation. We identify a key failure mode, which we term inter-path adaptation: during quantization-aware training (QAT), parallel residual binary paths learn redundant features, degrading the error-compensation structure and limiting the expressive capacity of the model. While prior work relies on heuristic workarounds (e.g., path freezing) that constrain the solution space, we propose RaBiT, a novel quantization framework that resolves co-adaptation by algorithmically enforcing a residual hierarchy. Its core mechanism sequentially derives each binary path from a single shared full-precision weight, which ensures that every path corrects the error of the preceding one. This process is stabilized by a robust initialization that prioritizes functional preservation over mere weight approximation. RaBiT redefines the 2-bit accuracy-efficiency frontier: it achieves state-of-the-art performance, rivals even hardware-intensive Vector Quantization (VQ) methods, and delivers a $4.49\times$ inference speed-up over full-precision models on an RTX 4090. Code is available at https://github.com/SamsungLabs/RaBiT.

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

How Far Can Machine Translation Quality Take You? Extrinsic Discourse Evaluation in Goal-Oriented Setups

Existing machine translation (MT) metrics and discourse-focused evaluations primarily assess translation quality intrinsically, without measuring the downstream consequences of translation errors. In this work, we focus on extrinsic discourse evaluation of machine translation under two distinct regimes: static and interactive. Under the static regime, we propose an entity counting task as a probe of referential consistency in discourse. We show that high intrinsic MT quality does not reliably predict downstream discourse success and strong MT systems still produce referential inconsistencies. For the interactive regime, we study the goal-oriented multi-agent Welfare Diplomacy game as a probe of long-horizon communication and coordination. We find that interaction-specific translation failures impact downstream coordination. Our results highlight goal-oriented environments as a viable framework for discourse-sensitive extrinsic MT evaluation.

10.
arXiv (CS.CL) 2026-06-15

Benchmarking Web Agent Safety under E-commerce Deceptive Interfaces

As autonomous web agents are increasingly deployed to perform real-world tasks, ensuring their safety has become a critical concern. In this work, we study web agent behavior under realistic deceptive interfaces in the e-commerce domain. We introduce WebDecept, a lightweight and configurable plugin framework that enables controlled injection of deceptive interface patterns into existing web environments. Using WebDecept, we instantiate seven deceptive patterns commonly observed on the open web, including targeted advertisements, domain redirection, and shopping manipulation. By injecting these patterns into the frontend during task execution, we perform controlled evaluation of multiple multimodal web agents. Our results show that current web agents are highly susceptible to multiple classes of deceptive interfaces, and that prompt-based constraints are often insufficient to mitigate these failures. We further analyze how the design choices of deceptive patterns influence the success of such manipulations. These findings highlight safety challenges that should be addressed as web agents are scaled toward real-world deployment.

11.
arXiv (CS.AI) 2026-06-18

Vibe Coding Ate My Homework: An evaluation of AI approaches to greenfield software engineering and programming

arXiv:2606.18293v1 Announce Type: cross Abstract: Thanks to rapid developments in generative AI, we are in the midst of a paradigm shift that may change how we interact with computers forever. We have observed a growth in the use of natural language prompts to build applications and coding infrastructures without underlying knowledge of the field, and this practice has been dubbed `vibe coding.' It arguably represents what the field of programming has been building towards since the beginning, with every higher level of abstraction that is conceived. Vibe coding promises to be the endpoint for the meta of high-level programming as far as method of input is concerned: eliminating a human's use of code syntax entirely in favour of programming in their mother tongue. This paper aims to evaluate the viability of vibe coding for greenfield software engineering tasks, as well as analyse the benchmarks that have been used to measure its software engineering prowess. To this end, we have developed an evaluation suite for analysing an LLM's proficiency in carrying out simple, isolated greenfield programming tasks in Python to provide scoped insight on the matter.

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

"Did you lie?" Evaluating Lie Detectors across Model Scale and Belief-Verified Model Organisms

arXiv:2606.12618v1 Announce Type: new Abstract: Robust lie detectors for language models could enable powerful techniques for auditing, monitoring, and post-hoc investigation of model behaviour, but evaluating them requires testbeds where models verifiably believe the opposite of what they say. We show that existing trained model organisms often fail this requirement, leaving prior positive and negative detection results difficult to interpret. We address this with 13 reasoning model organisms whose hidden beliefs are verified in chain-of-thought and shown to generalise to held-out tasks, alongside Varied Deception, a prompted-lying testbed covering a broad range of lie-inducing motivations. On these testbeds we evaluate four detectors: a chain-of-thought judge, a logprob classifier, and two activation probes, including Did-You-Lie (DYL), a new method for training follow-up probes. On prompted lying, across 31 open-weight models spanning 2B to 1T parameters, all four detectors show positive scaling with model capability. However, every activation- and logprob-based detector drops sharply on our trained model organisms, with DYL retaining the most signal; only the chain-of-thought judge remains strong, with 0.82 balanced accuracy, partly as an artefact of our verification process favouring CoT-readable beliefs. Current lie detectors therefore cannot support high-confidence claims about model beliefs, and we suggest research directions that may address some of their current limitations. We release our datasets, model organisms, and trained detectors.

13.
arXiv (CS.CV) 2026-06-18

Low-Cost Neuromorphic Fall Detection Using Synthetic Event Data and Hybrid SNNs

This work presents the development of hybrid models that integrate spiking neural networks (SNNs) with components of convolutional neural networks (CNNs) to learn from simulated event-based camera data (Dynamic Vision Sensor, DVS) generated from conventional smartphone videos. Aimed primarily at human fall detection, the approach leverages the energy efficiency and spatio-temporal processing capabilities of SNNs by converting video frames into event-based data. The proposed models are evaluated through simulations on multiple datasets, comparing their performance to that of traditional machine learning models. Results demonstrate significant gains in efficiency without sacrificing accuracy, underscoring the potential of combining SNNs and DVS technology for complex tasks in real-world environments.

14.
arXiv (CS.CV) 2026-06-16

Attention-Based Prototype Calibration for Multi-Rater Few-Shot Medical Image Segmentation

Few-shot medical image segmentation methods typically assume a single ground-truth annotation, overlooking systematic variability across expert raters commonly observed in clinical datasets. We propose an attention-based prototype calibration framework for few-shot multi-rater segmentation that models rater-specific deviations from a consensus representation in prototype space. A lightweight yet principled attention operator directly refines rater prototypes without modifying the backbone feature extractor, making the approach fully compatible with existing prototype-based few-shot segmentation methods. This design preserves semantic consistency while enabling personalized segmentation outputs with minimal computational overhead. Experiments on multi-rater medical imaging datasets demonstrate consistent improvements over baseline prototype approaches, highlighting the effectiveness of structured prototype calibration for modeling annotation variability. Our code is available at https://github.com/truong2710-cyber/JAPC.

15.
arXiv (CS.CL) 2026-06-17

Fine-tuning LLMs for Passive Depression Severity Estimation from AI Mental Health Dialogue

Depression is the leading cause of disability worldwide, and early detection of symptom change is essential for timely intervention. Validated instruments such as the Patient Health Questionnaire-9 (PHQ-9) support symptom monitoring at scale, but real-world completion rates are low, introducing response bias and systematic missingness. Passive approaches that infer severity from routinely generated data could close this gap. We address this by predicting PHQ-9 total scores directly from transcripts of conversations between users and an AI mental health application, requiring only conversation text and no additional clinical data. We fine-tune a Qwen3.5-27B backbone with a regression head, augment 3,111 ground-truth labels with pseudolabels generated by a reasoning model (Claude Opus) and iteratively trained intermediate models, for a combined dataset of 6,283 users. On a held-out test set of 842 users, our best model achieves MAE = 2.6, RMSE = 4.0, Pearson r = 0.80, and AUC = 0.91 at the PHQ-9 >= 10 clinical threshold. We also find AUC > 0.87 at every severity threshold from PHQ-9 >= 3 to PHQ-9 >= 24, demonstrating that the model captures depression severity across the full clinical spectrum. This work opens the door to passive, continuous symptom monitoring in AI mental health platforms, without requiring users to complete self-report measures.

16.
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.

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

Divide, Deliberate, Decide: A Multi-Agent Framework for Fine-Grained Egocentric Action Recognition

Fine-grained action recognition in egocentric video is challenging for Vision-Language Models (VLMs): actions often differ only in small visual cues, and a single model tends to be biased toward a subset of these cues. We propose Divide, Deliberate, Decide, a fully-local, zero-shot multi-agent framework in which (i) a VLM orchestrator chunks the video and proposes a top-k candidate label list per segment, (ii) an ensemble of heterogeneous VLM specialists, drawn from different open model families, engages in a structured deliberation that includes a peer-consultation round of questions, and (iii) agent rankings are aggregated with a Borda count and the orchestrator re-ranks its own prediction in light of the specialists' evidence. The entire pipeline runs locally with no fine-tuning. Experiments show that our method positively improves zero-shot action recognition performance over the baseline, highlighting the influence of a heterogeneous deliberation step, showing that the gain stems from decorrelated model priors rather than from additional compute.

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

A Multi-Domain Benchmark for Detecting AI-Generated Text-Rich Images from GPT-Image-2

Text-rich images often contain privacy-sensitive, transactional, or decision-relevant information. As recent multimodal image generation models become increasingly capable of synthesizing realistic textual content and structured visual designs, detecting AI-generated text-rich images has become an important challenge for digital trust and content authenticity. Existing benchmarks, however, largely focus on object-centric images and provide limited coverage of scenarios where textual semantics and layout organization are central. In this paper, we introduce a multi-domain benchmark for detecting text-rich images generated by OpenAI's GPT Image 2. The benchmark contains 8,602 images across six representative categories: commercial posters, infographics, academic posters, receipts, tables, and UI screenshots. Using this benchmark, we evaluate five representative AI-generated image detectors in a zero-shot setting and analyze their overall, category-wise, and post-processing robustness. Our results show that detector performance is highly domain-dependent: methods that perform well in some categories often fail on others, and even the strongest conventional detector exhibits severe sensitivity to JPEG compression. We further conduct an exploratory evaluation with a multimodal vision-language model, revealing both its promise and its limitations on structured formats. These findings highlight the need for text- and layout-aware detection methods for modern AI-generated images. Our dataset is released at XXX.

19.
arXiv (CS.AI) 2026-06-11

Position: Hippocampal Explicit Memory Is the Cornerstone for AGI

作者:

arXiv:2606.11245v1 Announce Type: new Abstract: Large Language Models (LLMs) have demonstrated remarkable capabilities across various tasks, raising expectations for Artificial General Intelligence (AGI). This position paper argues that integrating explicit memory is the cornerstone for advancing LLMs toward AGI. The key reason is that the underlying learning mechanism of LLMs is highly analogous to human implicit memory. However, higher-order cognitive functions necessary for AGI, such as long-term strategic planning, metacognition, and symbolic reasoning, heavily rely on hippocampal explicit memory and cannot arise solely from implicit statistical learning. Drawing on findings from neuroscience, I advance this perspective and complement it with computational requirements for artificial explicit memory systems, hoping to foster further research and lay the groundwork for explicit memory integration.

20.
arXiv (quant-ph) 2026-06-11

A Pfaffian quantum Hall state of ultracold bosons

arXiv:2606.12409v1 Announce Type: cross Abstract: Fractional quantum Hall states are a cornerstone of topological physics, hosting fractionally charged quasiparticles with exotic statistics that promise to enable topologically protected quantum information processing. Among these, the Pfaffian state introduced by Moore and Read implements a p-wave pairing structure that supports excitations with non-Abelian exchange statistics. Despite extensive study in electronic systems, direct access to its pairing structure has remained limited. Here we realize a three-particle bosonic Pfaffian state of ultracold $^{87}\mathrm{Rb}$ atoms in an optical lattice subject to a Floquet-engineered synthetic magnetic field. Using a Bayesian-optimized adiabatic protocol, we prepare a state exhibiting Pfaffian pairing correlations. Site-resolved measurements of multi-point density correlations reveal a pronounced suppression of short-range three-body coincidences, reflecting the underlying pairing structure. We further probe the state's transport response through Hall drift measurements. Our results establish a bottom-up approach to engineering non-Abelian topological order and lay the groundwork for future explorations of anyonic braiding in synthetic matter.

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

Temporal2Seq: A Unified Framework for Temporal Video Understanding Tasks

With the development of video understanding, there is a proliferation of tasks for clip-level temporal video analysis, including temporal action detection (TAD), temporal action segmentation (TAS), and generic event boundary detection (GEBD). While task-specific video understanding models have exhibited outstanding performance in each task, there remains a dearth of a unified framework capable of simultaneously addressing multiple tasks, which is a promising direction for the next generation of AI. To this end, in this paper, we propose a single unified framework, coined as Temporal2Seq, to formulate the output of these temporal video understanding tasks as a sequence of discrete tokens. With this unified token representation, Temporal2Seq can train a generalist model within a single architecture on different video understanding tasks. In the absence of multi-task learning (MTL) benchmarks, we compile a comprehensive co-training dataset by borrowing the datasets from TAD, TAS, and GEBD tasks. We evaluate our Temporal2Seq generalist model on the corresponding test sets of three tasks, demonstrating that Temporal2Seq can produce reasonable results on various tasks and achieve advantages compared with single-task training on this framework. We also investigate the generalization performance of our generalist model on new datasets from different tasks, which yields superior performance to the specific model.

22.
arXiv (CS.AI) 2026-06-19

Process-Verified Reinforcement Learning for Theorem Proving via Lean

arXiv:2606.20068v1 Announce Type: new Abstract: While reinforcement learning from verifiable rewards (RLVR) typically has relied on a single binary verification signal, symbolic proof assistants in formal reasoning offer rich, fine-grained structured feedback. This gap between structured processes and unstructured rewards highlights the importance of feedback that is both dense and sound. In this work, we demonstrate that the Lean proof assistant itself can serve as a symbolic process oracle, supplying both outcome-level and fine-grained tactic-level verified feedback during training. Proof attempts are parsed into tactic sequences, and Lean's elaboration marks both locally sound steps and the earliest failing step, yielding dense, verifier-grounded credit signals rooted in type theory. We incorporate these structured rewards into a GRPO-style reinforcement learning objective with first-error propagation and first-token credit methods that balances outcome- and process-level advantages. Experiments with STP-Lean and DeepSeek-Prover-V1.5 show that tactic-level supervision outperforms outcome-only baselines in most settings, delivering improvements on benchmarks such as MiniF2F and ProofNet. Beyond empirical gains, our study highlights a broader perspective: symbolic proof assistants are not only verifiers at evaluation time, but can also act as process-level reward oracles during training. This opens a path toward reinforcement learning frameworks that combine the scalability of language models with the reliability of symbolic verification for formal reasoning.

23.
arXiv (CS.CL) 2026-06-15

SIMMER: Benchmarking Latent Failures in LLM Executable Planning with a World Model

Large language models (LLMs) are increasingly deployed as planners for autonomous agents in household environments. While existing benchmarks evaluate whether LLM-generated plans execute successfully, they overlook a critical type of failure: latent failures. Unlike immediate failures that trigger instant feedback at execution time and enable timely correction, latent failures do not immediately halt plan execution but silently compromise goal achievement. In severe cases, they cause irreversible harm. To address this gap, we introduce SIMMER, a benchmark for evaluating latent failures in LLM planning through a human-curated symbolic world model grounded in the kitchen domain. SIMMER defines a world model comprising 77 actions, 262 unique objects, and approximately 46,800 possible interactions that are semantically realistic, derived from real-world cooking scripts. It then leverages a state machine executor that validates plans against the world model and detects immediate precondition violations, latent hazards, and irreversible failures. Experiments across six LLMs show that even frontier models achieve at most 17% error-free plans. Moreover, up to 56% of plans contain latent failures, the majority of which lead to irreversible consequences. We further demonstrate that explicit state reasoning via counterfactual foresight simulation can reduce latent failures by up to 72% and irreversible cases by up to 75%, suggesting a promising direction for more robust LLM planners.

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

Calibrated Triage, Not Autonomy: Confidence Estimation for Medical Vision-Language Models

A vision-language model can answer a question about a medical image fluently and confidently while barely using the image, leaning instead on language priors. In medicine this is the failure that matters most, because the answer looks trustworthy and is not, and the only protection is a confidence score reliable enough to tell the system when to abstain. We ask a deployment question rather than an accuracy one: how much imaging work a model can safely handle alone, and which confidence signal makes that possible. We evaluate seven confidence estimators across five open-weight LVLMs and three medical visual-question-answering datasets spanning broad clinical imaging, radiology, and pathology, with every probe trained only on natural images and applied without adaptation. Recast as bounded selective prediction (automate a case only when confidence clears a threshold, defer the rest), the comparison is cautionary. The standard metrics are poor guides: discrimination barely separates the methods, and the weak calibration of a cheap self-report is cheaply removed by off-domain temperature scaling without changing deployable yield. What distinguishes a usable estimator is the high-confidence region a clinician acts on: the weakest baselines are confidently wrong on 41 to 45 percent of their errors against 1 to 4 percent for the best probe, and no estimator is reliably best across domains or models. Safe handoff is governed at two levels: base-model competence sets a ceiling, so a well-calibrated score recovers roughly a third of radiology cases at a 20 percent error tolerance but almost none of pathology; the confidence layer then decides how much of that ceiling is reachable. The usable role today is calibrated triage, not autonomy: automate the cases a calibrated score marks safe, route the rest to a clinician. We release all outputs, correctness judgments, and confidence scores, with code.

25.
medRxiv (Medicine) 2026-06-19

Validation of an Artificial Intelligence-Assisted Mobile Application for Dietary Oxalate Assessment in Kidney Stone Prevention

Background: Calcium oxalate nephrolithiasis is the most common type of kidney stone disease. Dietary oxalate intake is an important modifiable factor. Assessing dietary oxalate exposure in clinical practice poses challenges due to limitations of traditional dietary recall tools and variability in food composition data. Artificial intelligence (AI) applications in mobile health may offer scalable solutions for better dietary monitoring and kidney stone prevention. We examined the ability of StoneFree AI to estimate dietary oxalate from verbal and image-based food inputs. Objective: To evaluate the accuracy and limitations of StoneFree AI, for estimating dietary oxalate intake from verbal food descriptions and meal images, and to evaluate errors from entries that may inform future clinical use in kidney stone prevention. Methods: StoneFree AI is a cross-platform mobile application that uses a multimodal large language model (Google Gemini) to interpret verbal food descriptions and visual food images. The identified foods were mapped to oxalate values using the Harvard Oxalate Database. System performance was evaluated using 804 verbal food entries and 276 portion-size food images obtained from the ASA24 dietary assessment database. Verbal inputs were compared with reference oxalate values using absolute error and predefined agreement thresholds ({+/-}1, {+/-}5, {+/-}10 mg). Image-based inputs were evaluated against mutually exclusive primary error categories, including food identification, portion estimation, ingredient recognition, oxalate reference selection, and non-analyzable cases. Results: For verbal food entries, the AI system showed strong agreement with reference oxalate values. Overall, 82.1% of estimates were within {+/-}1 mg, 91.5% within {+/-}5 mg, and 94.5% within {+/-}10 mg of reference values. The mean absolute error was 3.32 mg, the median absolute error was 0.10 mg, and the concordance correlation coefficient (CCC) was 0.860. Image-based inputs showed a higher overall error rate of 63.0%, primarily due to food identification errors (33.0%), inaccurate portion estimation (11.0%), and ingredient recognition errors (9.8%). Most errors occurred with visually complex meals, such as mixed dishes and grain-based foods. Conclusions: AI-assisted estimation of dietary oxalate intake demonstrated high accuracy when structured verbal inputs were used but was less reliable for image-based meal analysis. These findings suggest AI-enabled mobile tools may support dietary monitoring for kidney stone prevention, particularly when user input is structured. Further refinement of computer vision models and prospective clinical validation are required before widespread clinical implementation.