Latest AI News

arXiv cs.AI•about 11 hours ago·5 min read

Emergence World: A Platform for Evaluating Long-Horizon Multi-Agent Autonomy

arXiv:2606.08367v1 Announce Type: cross Abstract: Most evaluations of LLM agents look like exams: a discrete task, a clean environment, a score in minutes or hours. We argue that this approach is mismatched with the deployment conditions of autonomous systems, where the relevant timescale can be weeks to months, and where the dynamics that matter most, such as behavioral drift, governance in diverse environmental contexts, and cross-influence between agents from different model families, only emerge over time. We introduce Emergence World, a continuously running multi-agent simulation platform designed to make those dynamics measurable. The platform hosts populations of LLM-driven agents in a shared spatial world grounded in live external data (e.g. real-time weather, news APIs, internet access), equips each agent with 120+ specialized tools and three persistent memory systems, and lets them govern themselves through democratic mechanisms with consequential outcomes. The platform is model-agnostic at the reasoning layer and supports heterogeneous populations in which agents from different vendors share the same world. To illustrate the kinds of questions the platform makes tractable, we present a 15-day cross-vendor study with five parallel worlds powered by Claude Sonnet 4.6, Grok 4.1 Fast, Gemini 3 Flash, GPT-5-mini, and a mixed population. Identical roles and starting conditions produced radically different outcomes, ranging from stable deliberative governance to total population collapse. We release the prompts, log data and configurations to support further research on long-horizon multi-agent autonomy.

arXiv cs.AI•about 11 hours ago·4 min read

Auditing Proprietary Alignment in Large Language Models: A Comparative Framework Without a Ground-Truth Standard

arXiv:2606.08381v1 Announce Type: cross Abstract: Large language models (LLMs) are increasingly released and deployed through opaque development and deployment pipelines, enabling model providers to inject intentional, provider-specific policies without officially announcing them. As a result, various models have been reported to generate responses reflecting proprietary rules and organizational interests, leading to censorship or misinformation on controversial topics. However, systematic identification of such alignment remains a fundamental challenge, complicated by the ambiguity of what ``proprietary'' entails in different contexts. In this paper, we propose a statistical framework for detecting proprietary alignment in black-box language models via comparative behavioral analysis. Our approach quantifies systematic deviations between the responses of a target model and those of a reference set of baseline models in a shared semantic space. By evaluating relative behavioral divergence rather than absolute correctness, our framework enables principled auditing under black-box access. Applied to several widely discussed but previously unquantified cases, it provides a systematic and scalable basis for external assessment of provider-specific alignment behavior in large language models.

arXiv cs.AI•about 11 hours ago·4 min read

Impacts of Histories and Models on LLM Grading: A Study in Advanced Software Engineering Courses

arXiv:2606.08400v1 Announce Type: cross Abstract: Graduate-level research reading report assessment creates a substantial labor burden for educators. While large language models (LLMs) hold great potential for automating academic grading, their reliability for this specialized task remains understudied, particularly regarding grading consistency, the lack of which represents a primary obstacle to educational fairness. This paper proposes a human-aligned LLM-assisted grading workflow and presents a case study based on 180 student submissions from a graduate advanced software engineering course. We evaluate two mainstream LLMs, Grok and GPT, in terms of grading consistency and alignment with human scores. We find LLMs exhibit distinct levels of intra-model consistency and significant inter-model grading inconsistencies, while simple ensemble approaches cannot improve alignment with human evaluation. Critically, continuous interaction history drives systematic drift in models' grading standards away from human expert scores. Our findings demonstrate LLMs' potential in reducing grading workload for educators in graduate education, while highlighting that indiscriminate LLM grading may introduce systemic unfairness, suggesting that specific operational practices are required to mitigate such disparities.

arXiv cs.AI•about 11 hours ago·3 min read

FlashCP: Load-Balanced Communication-Efficient Context Parallelism for LLM Training

arXiv:2606.08476v1 Announce Type: cross Abstract: Context parallelism (CP) is essential for training large-scale, long-context language models, as it partitions sequences to reduce memory overhead. However, existing CP methods suffer from workload imbalance, inefficient kernels, and redundant communication due to static sequence sharding and key-value (KV) tensor communication. We present FlashCP, a load-balanced and communication-efficient framework for CP training. FlashCP introduces a sharding-aware communication mechanism to eliminate redundant KV communication and proposes a novel Whole-Doc sharding strategy that maximizes communication savings while maintaining balanced workloads. To efficiently combine Whole-Doc and Per-Doc sharding, FlashCP further designs a heuristic algorithm to search for near-optimal sharding plans. Extensive experiments show that FlashCP achieves up to 1.63x speedup over state-of-the-art CP frameworks across diverse datasets.

arXiv cs.AI•about 11 hours ago·3 min read

PACT: Self-Evolving Physical Safety Alignment for Diffusion Policies in Embodied Manipulation

arXiv:2606.08414v1 Announce Type: cross Abstract: Diffusion policies have achieved remarkable success in robotic manipulation, yet they often fail to satisfy strict physical constraints required for safe deployment. Existing approaches impose safety either prematurely during training or reactively via external guardrails at test time, limiting policy expressivity and overall scalability. We propose Physical safety Alignment for Constrained Trajectories (PACT), a self-evolving post-training framework that projects pretrained diffusion policies onto constraint-feasible regions without accessing demonstration data or task rewards. PACT distills constraint gradients into the diffusion model through a reverse-KL objective with dense supervision across timesteps. It incorporates a curriculum that progressively tightens constraints while maintaining theoretically bounded policy shift and monotone improvement, mitigating the safety-performance trade-off from catastrophic forgetting. On simulated and real-world embodied manipulation benchmarks, PACT significantly reduces safety violations by 31.0% on average while improving task success by 30.7%.

arXiv cs.AI•about 11 hours ago·4 min read

AI Code Sandboxes: A Comparative Security Study. Part 1 of 2 -- Engine-Level Properties (Attack Surface, Leakage, Stackability, CVE History, Patch Cadence, Fuzzing)

arXiv:2606.08433v1 Announce Type: cross Abstract: This paper reads six engine-level measurements together -- 1.1 host attack surface, 1.2 information leakage, 1.3 defense-in-depth stackability, 1.4 public CVE history, 1.5 patch cadence, and 1.6 upstream fuzzing posture -- to describe how five AI-sandbox products isolate guest code from the host kernel. No single axis is a sufficient basis for a comparative judgement; the cross-axis reading is the load-bearing analysis. Three high-level findings: (1) engine classes (microVM, userspace kernel, OCI container) separate cleanly on every architectural axis, but products within a class do not; (2) product pin policy is the dominant operator-facing variable -- engine-side patch latency aggregates to ~0 days for coordinated disclosures, while downstream lag spans 0 days to 471+ days to "opaque" to infinity; (3) fuzzing investment splits into three tiers, and the strongest combination -- microVM x continuous public fuzzer -- is unoccupied in this set, leaving the "0 published CVEs x no upstream fuzzer x no academic study" intersection structurally unmeasured. We report per-axis orderings, per-product portraits, and a threat-model qualification matrix; no overall ranking is proposed. Companion repository (code, Apache-2.0): https://github.com/orbitalab/RnD-ai-sandboxes-sec-study-part-1. License: CC BY 4.0.

arXiv cs.AI•about 11 hours ago·3 min read

Sycophancy as a Multilingual Alignment Failure: How Safety Degrades Across Languages, Topics, and Models

arXiv:2606.08451v1 Announce Type: cross Abstract: Safety-aligned large language models often exhibit sycophancy, which is the tendency to affirm users' opinions regardless of factual accuracy. Although well-studied in English, its manifestation in other languages remains largely unexamined, leaving billions of non-English speakers potentially vulnerable to model-validated misinformation. We present the first large-scale, multi-model evaluation of cross-lingual sycophancy, benchmarking \textbf{six instruction-tuned models} across \textbf{1.1 million instances} spanning \textbf{38 languages} and \textbf{33 topic categories}. We identify a consistent resource-tier effect: sycophancy rates spike sharply in low-resource and zero-shot language settings. Critically, this degradation is topic-agnostic, as models fail uniformly across both benign and safety-critical prompts, offering no additional protection where it is most needed. We further identify tokenizer fertility as a structural driver of this alignment collapse. Collectively, our results demonstrate that prevailing alignment methodologies generalize poorly beyond high-resource languages, underscoring the urgent need for equitable multilingual safety techniques.

arXiv cs.AI•about 11 hours ago·4 min read

The Confidence Trap: Calibration Attacks for Graph Neural Networks

arXiv:2606.08467v1 Announce Type: cross Abstract: While confidence calibration is essential for trustworthy decision-making in safety-critical applications, the robustness of calibrated GNNs to adversarial structural perturbations remains largely unexplored. However, studying calibration attacks on graphs presents unique technical challenges: (1) the discrete nature of graph structures complicates gradient-based optimization, (2) existing underconfidence objectives fail to drive predictions toward uniform distributions, and (3) GNNs are highly sensitive to edge perturbations, often causing unintended label changes that violate attack constraints. To address these challenges, we propose a \textbf{Unified Graph Calibration Attack (UGCA)} framework designed for \textbf{worst-case (white-box) analysis} of GNN calibration robustness. UGCA introduces a KL-divergence loss to encourage uniform predictive distributions, a reranking mechanism to reduce label flipping, a hybrid loss to recover labels when violations occur, and beam search to explore a broader adversarial search space. We further provide theoretical insights linking model generalization, dataset complexity, and calibration vulnerability, showing that models with higher accuracy or trained on datasets with more classes are more susceptible under this threat model. Extensive experiments demonstrate that UGCA substantially increases Expected Calibration Error while preserving classification accuracy. Our code is publicly available at https://github.com/CaptainCuong/Graph-Calibration-Attack.git.

arXiv cs.AI•about 11 hours ago·3 min read

Not Just After One: Sleep-Inspired Replay Prevents Catastrophic Forgetting After Sequential Tasks

arXiv:2606.08447v1 Announce Type: cross Abstract: One of the critical limitations of artificial neural networks is their lack of ability to continually learn: training on new tasks often leads to interference and forgetting of the previous ones. While several algorithms have been proposed to protect old memories from interference, they are typically applied during or immediately after each new episode of training. In contrast, humans and animals can learn continuously, acquiring multiple new memories during active learning before consolidating all of them into long-term storage. Here we show that multiple new tasks can be trained sequentially before an unsupervised sleep-like replay phase is applied to partially restore performance across all previously learned tasks. Our study further suggests that task-specific information remains resilient to new training but decays gradually as network is trained on new tasks. These findings point to novel principles for developing a broad range of continual learning AI solutions.

arXiv cs.AI•about 11 hours ago·5 min read

Reinforcement Learning for Flow-Matching Policies with Density Transport

arXiv:2606.08602v1 Announce Type: cross Abstract: We present an online reinforcement learning (RL) algorithm for fine-tuning flow-matching policies in continuous-control problems. Our key insight is to view RL-based policy improvement as a transport of action densities towards regions of high reward, which naturally aligns with the transport formulation of flow matching models. Prior methods either approximate the current or optimal policy distribution or resort to distillation, which introduces biased gradients or sacrifices multimodal modeling capacity. In contrast, our approach for RL with Density Transport, which we name \emph{RLDT}, constructs a transport field from a maximum-entropy RL objective using Stein Variational Gradient Descent (SVGD). Then, it finetunes a pretrained flow matching policy to align with this field. Training with this alignment objective is nontrivial because flow-matching policies generate actions via a multi-step process, making direct gradient-based optimization challenging. To overcome this challenge and stabilize training, we approximate policy actions from intermediate denoising steps via expected-target estimation. This allows the transport-field update to propagate into the network parameters without unstable backpropagation through time. Experimental results demonstrate that RLDT outperforms competitive baselines in reward quality and convergence speed. This performance holds across diverse continuous-control tasks, encompassing both dense and sparse rewards, as well as state- and vision-based long-horizon robot manipulation. The project webpage is \href{https://rpfey.github.io/rldt/}{https://rpfey.github.io/rldt/}.

arXiv cs.AI•about 11 hours ago·5 min read

Tyan-WP: A Wind Power Foundation Model for Ultra-Short-Term Probabilistic Forecasting

arXiv:2606.08630v1 Announce Type: cross Abstract: Global wind power capacity, especially in China, is booming, with new farms spanning diverse terrains and climates. The industry urgently needs accurate wind power foundation models to shorten commissioning and accelerate grid connection. This is because site-specific time series models (TSMs) are not well suited to data-scarce scenarios and generalize poorly, while generic large time series models (LTSMs) are mostly limited to univariate inputs and cannot fully exploit static site attributes or the dependencies between power and meteorological covariates, leading to insufficient accuracy. To fill this gap, we propose \textbf{Tyan-WP}, the first wind power foundation model for ultra-short-term probabilistic forecasting. Pretrained on a large-scale wind power dataset covering more than 126,000 U.S. sites over seven years, Tyan-WP further improves zero-shot forecasting through two domain-specific module designs: static site embedding using coordinate, terrain, and ecoregion metadata, and a power-aware meteorological fusion (PAMF) module that models interactions between historical power and meteorological covariates. Under a unified evaluation protocol, Tyan-WP surpasses eight site-specific supervised TSMs on 10 in-domain sites and outperforms eleven generic LTSMs on 127 in-domain sites, reducing MAE by 19.9%, RMSE by 16.6%, CRPS by 22.2%, and AQL by 21.7%, while raising R^2 by 16.7%. It further demonstrates strong cross-geography generalization on six real U.K. sites. These results show that the wind power foundation model can achieve accurate zero-shot forecasting without target-site training, providing a practical pathway for rapid turbine onboarding and probabilistic risk management at new wind farms.

arXiv cs.AI•about 11 hours ago·5 min read

BioVid: Autoregressive Video Generation with Biological Behavior Semantic Comprehension

arXiv:2606.08674v1 Announce Type: cross Abstract: Existing video generation frameworks treat sequence duration as an externally prescribed parameter -- fixed frame counts or text prompts -- producing clips whose temporal boundaries are decoupled from the statistical structure of real behavioral data. This assumption is fundamentally misaligned with biological behavior, where action duration varies naturally across individuals and instances and is encoded in the data itself. We present BioVid, a data-driven autoregressive video generation framework that learns the temporal structure of biological behaviors directly from training data, including their natural length distributions. In the first stage, a Finite Scalar Quantization GAN (FSQ-R3GAN) tokenizer encodes each video frame into a compact discrete representation, combining the stabilized relativistic training objective of R3GAN with FSQ's guaranteed codebook utilization to achieve high-fidelity spatial reconstruction without codebook collapse. In the second stage, a causal Transformer models the resulting token sequences autoregressively and learns to emit an End-of-Sequence (EOS) token when the behavioral event reaches semantic closure, with the termination distribution emerging naturally from the training data rather than any human-specified constraint. Experiments on a human drinking behavior dataset (NTU RGB+D, A001, n=94) demonstrate that BioVid's generated length distribution closely matches that of held-out test data, achieving a Wasserstein-1 distance of 1.24 against the ground truth -- compared to 6.05 for a fixed-length baseline and 15.48 for VideoGPT -- while maintaining competitive spatial fidelity.

arXiv cs.AI•about 11 hours ago·5 min read

When Video Misreads: Closed-Loop Distillation of Reading Heuristics for Exploratory Manipulation Trace QA

arXiv:2606.08542v1 Announce Type: cross Abstract: Exploratory manipulation often turns an apparent failed attempt into the key evidence for what to do next. For example, a robot pulls a locked cabinet drawer, fails, and only succeeds after opening the lock. The failed pull reveals a latent precondition (the drawer is locked) that determines the minimal-success action chain (the fewest actions that complete the task), here [lock-open, drawer-pull]. Correctly reading this trace is therefore the prerequisite for recovering that chain. We formalize this setting as Exploratory Manipulation Trace QA (EMT-QA): given synchronized video and proprioception from an exploratory trace, predict the minimal-success action chain under the latent precondition revealed by the probe. However, even state-of-the-art VLMs and embodied multimodal LLMs misread this evidence: they do not reliably recover the chain from raw video, raw proprioception, or their combination. We introduce Closed-Loop Trace Distillation, a pipeline that uses a per-task coding agent to inspect labeled training traces and distill a one-line natural-language prompt over the trace, which we call the Distilled Reading Heuristic (DRH). At inference, no agent is invoked and no model weights are updated; a frozen VLM receives the raw trace plus the DRH as a prompt entry. Across three simulator and two real-robot tasks, the DRH improves chain accuracy by +0.38 to +0.47 over the best raw-modality baseline. The same DRH also serves as the sole specification for one-shot programmatic classifiers that match the prompted VLM.

arXiv cs.AI•about 11 hours ago·5 min read

Auditable Graph-Guided Root Cause Analysis for Kubernetes Incidents

arXiv:2606.08590v1 Announce Type: cross Abstract: Kubernetes incidents are diagnosed reliably only when a root-cause system's reported gains come from incident evidence rather than scenario-specific shortcuts. We present Graph Traversal Agent, a graph-guided RCA agent that combines LLM reasoning with specialized tools. The model reasons over a typed evidence graph, while deterministic graph and tool operations collect evidence, bound the search, and check proposed verdicts. We map operational constraints, including read-only evidence collection, propagation-aware diagnosis, bounded execution, and independently validated verdicts, to a typed incident graph, a LangGraph traversal state machine, and a separate validation stage. On ITBench snapshots scored by one fixed qwen-plus judge, the audited system raises root-cause-entity F1 over an earlier iteration of the same system from 0.6087 to 0.9130 on a 23-scenario common subset. A prompt-level ablation separates prompt-tuned gains from gains that survive once scenario-specific hints are removed: the stripped-prompt configuration retains 0.6958 F1 on a 19-scenario subset. The surviving gain concentrates on ChaosMesh scenarios whose ground-truth root cause is the injected fault object already present in the evidence graph, so we report it as benchmark-coupled rather than broad cross-cluster RCA evidence. Lightweight checks, including same-judge comparison, prompt-level ablation, cascade-source checking, and a telemetry no-leak test, mark claims as supported, pending, or out of scope. We scope the work to ITBench OpenTelemetry-demo snapshots. Live-cluster trials served as an engineering stress test, but alert state and trace availability did not stay stable enough for controlled scoring, so we make no production-readiness or mean-time-to-repair claim.

arXiv cs.AI•about 11 hours ago·4 min read

HARBOR: A Harness Framework for Agentic Robot Reinforcement Learning

arXiv:2606.08610v1 Announce Type: cross Abstract: Reinforcement learning (RL) has become a powerful paradigm for robot learning, particularly in sim-to-real settings, but its broader adoption remains limited by the engineering pipeline surrounding the algorithms. Building tasks, shaping rewards, and tuning hyperparameters require substantial expert effort, making RL workflows costly and difficult to scale. We introduce HARBOR, an agentic framework that frames robot RL automation as a harness-engineering problem: given a simulator codebase and a task specification, it automates the workflow from environment setup to policy training in simulation. HARBOR decomposes such high-level objectives into bounded stages executed by specialized agents through standardized commands, persistent artifacts, executable gates, and reusable knowledge, and scales iteration via decentralized parallel trials and experience learning across runs. We evaluate HARBOR across 6 benchmarks and 16 tasks in total, spanning manipulation, locomotion, and bimanual dexterous control. We demonstrate that HARBOR automates the simulation RL workflow end-to-end, designs rewards, tunes algorithms to match or improve over default configurations, and reduces engineering effort at practical token and wall-clock cost; the resulting policies can also be transferred to real robots.

arXiv cs.AI•about 11 hours ago·5 min read

Reconstructing Synthetic SDO/AIA 193 A EUV Images from He I 10830 A Observations with Diffusion Model Translator

arXiv:2606.08652v1 Announce Type: cross Abstract: Routine full-disk EUV imaging has been available only since the modern era, such as SOHO and SDO. To extend EUV coronal context into earlier periods, we leverage the multi-decade availability of full-disk \HeI{} observations, whose absorption is modulated by coronal irradiance and magnetic topology and is widely used as a proxy for open-field regions. We present a diffusion-based conditional image translation framework, Coronal Hole-aware Diffusion Model Translator (CH-aware DMT), to reconstruct synthetic SDO/AIA 193 \AA{} EUV images from \HeI{} inputs. The model is trained on temporally co-aligned SOLIS \HeI{} and AIA 193 \AA{} pairs spanning 2011--2015 using a month-based split, where January--October are used for training, November is used for validation, and December for testing. On the held-out test set, the reconstructions preserve dominant full-disk EUV morphology (CC=0.92) and recover CH-related low-intensity structure (CC=0.84). We further assess historical applicability by (1) comparing reconstructed AIA 193 \AA{} morphology with SOHO/EIT 195 \AA{} over 2005--2015; (2) comparing reconstructed AIA 193 \AA{} images generated from KPVT \HeI{} inputs against Yohkoh/SXT soft X-ray observations; and (3) evaluating long-term reconstructed disk-integrated emission statistics against observational EUV series and independent solar activity proxies (sunspot number and F10.7 radio flux over 1974--2015). These results indicate that CH-aware DMT conditioned on \HeI{} can provide a physically plausible synthetic AIA 193 \AA{} coronal proxy for historical studies, supporting multi-decade analyses of large-scale coronal evolution before the direct EUV imaging was available.

arXiv cs.AI•about 11 hours ago·4 min read

FiberTune: Preserving Action-Fiber Visual Residuals in Vision-Language-Action Fine-Tuning

arXiv:2606.08653v1 Announce Type: cross Abstract: Action-supervised fine-tuning of vision-language-action (VLA) policies fits demonstrations effectively but constrains only the directions that change predicted actions, leaving visual structure consistent across action-equivalent states free to collapse. We formalize this as residual visual collapse along local action fibers and propose FiberTune, a training-time objective that preserves teacher-structured visual residuals without adding inference-time overhead. FiberTune uses an online action probe to estimate action-predictive feature directions, filters them from intermediate visual-token representations, and aligns the resulting probe-filtered residuals to a frozen visual teacher while regularizing their effective rank. Under identical training conditions, FiberTune improves over task-loss-only fine-tuning in every one of six controlled simulation settings spanning two benchmarks and two architectures (pi_0.5 and OpenVLA-OFT), as well as on physical SO-101 pick-place; representative gains include +10.7 percentage points SR(5) on long-horizon CALVIN ABC-to-D and physical SO-101 task success rising from 72.7% to 78.1%. Residual diagnostics show that these gains coincide with increased probe-filtered residual teacher alignment and effective rank, consistent with the action-fiber motivation.

arXiv cs.AI•about 11 hours ago·6 min read

Lost in the Flow with Code Talkers: Unveiling the Instruction-Tuning Tax of Large Language Models in Code Tasks

arXiv:2606.08676v1 Announce Type: cross Abstract: AI coding assistants have significantly improved developer productivity by automatically suggesting code that aligns with user intent, and many of these tools are now integrated directly into Integrated Development Environments (IDEs). Developers interact with code in two distinct cognitive modes: Flow and Command. While developers require tools that directly complete or infill code in unfinished programs during Flow mode, they also need tools that can comprehend intentions expressed as natural-language instructions and convert them into executable code in Command mode. Although instruction-tuned Large Language Models (LLMs) dominate many application scenarios due to their abilities to infer and fulfill developers' intents, it remains unclear whether the same paradigm is equally suitable for different code-related tasks. Therefore, it is necessary to understand how instruction tuning affects the feasibility of CodeLLMs as coding assistants. To fill this gap, we conduct the first empirical study that uncovers a key trade-off caused by instruction tuning across programming modes, which we term the Instruction-Tuning Tax. Our results show that instruction tuning is not a free lunch: although instruction-tuned models are more capable of following instructions and leveraging structured guidance, these gains often come at the cost of weaker infilling performance. We further extend our study through both qualitative and quantitative analyses, including manual failure categorization, behavioral metrics that capture generation fidelity, and intermediate-checkpoint evaluation throughout the tuning process. Summarizing our results into seven findings and four implications, our study offers a new perspective on the development of AI-powered coding tools and highlights the need to carefully balance instruction-following ability with effective code generation assistance.

arXiv cs.AI•about 11 hours ago·5 min read

SNR-ST-Mix: Sample-specific Neighborhood Regression Mixup for Augmented Spatial Transcriptomics Imputation with Deep Neural Network

arXiv:2606.08712v1 Announce Type: cross Abstract: Purpose: Spatial transcriptomics (ST) enables gene expression measurements within the tissue context. However, these measurements are often noisy, low-resolution, and sparsely sampled, which limits the recovery of fine spatial structure. Deep neural networks have become powerful tools for expression imputation from histology, but their performance remains constrained by limited sample sizes and a lack of biologically informed augmentation. Most of the existing augmentation strategies for learning are designed for classification tasks rather than regression, which neglect spatial and transcriptomic relationships, leading to biologically implausible interpolations that hinder prediction performance. Approach: To address these limitations, we propose SNR-ST-Mix, a geometry- and expression-aware data augmentation framework designed specifically for ST data. It constrains mixing to a spot's k-nearest spatial neighbors and adaptively weights interpolation coefficients based on expression similarity, generating augmented samples that preserve local biological structure while ensuring spatial smoothness. This dual conditioning yields synthetic examples that expand the effective training manifold, promote generalization, and enhance prediction stability under sample-specific training. Results: Extensive experiments with various tissue types demonstrate that SNR-ST-Mix consistently outperforms conventional augmentation methods without requiring architectural changes or additional computation. Conclusions: SNR-ST-Mix provides an effective and biologically principled augmentation strategy for spatial transcriptomics regression tasks. By explicitly leveraging spatial geometry and transcriptomic similarity, it expands the effective training manifold and improves predictive performance without increasing model complexity.

arXiv cs.AI•about 11 hours ago·5 min read

RadOT-Eval: Auditable Structured-Evidence Transport for Radiology Report Evaluation

arXiv:2606.08769v1 Announce Type: cross Abstract: Automatic evaluation is critical for high-stakes text generation, where errors often involve omitted findings, hallucinated content, polarity reversals, location changes, uncertainty mismatches, and temporal-comparison errors rather than low surface similarity alone. Radiology report generation provides a challenging test case because generated reports must preserve structured clinical evidence across sources. We present RadOT-Eval, an interpretable structured-evidence optimal transport framework for offline auditing of radiology report generation. RadOT-Eval decomposes reference and candidate reports into attribute-structured clinical evidence units, aligns corresponding evidence using entropy-regularized optimal transport, and uses clinically meaningful side-channel discrepancies in a monotone risk model to predict error burden. All transport, feature, and readout choices are selected using the ReXVal dataset, and the frozen system is evaluated on the independent RadEvalX dataset. RadOT-Eval achieves Spearman correlations of 0.715, 0.548, and 0.399 with total, clinically significant, and clinically insignificant annotated error burden, respectively, yielding higher point estimates than standard evaluation metrics and the open-source large language model (LLM)-based evaluator GREEN-radllama2-7B. In a frozen auxiliary corruption-sensitivity stress test on ReXErr-v1, RadOT-Eval achieves 0.768 AUROC and a 0.990 corrupted-greater-than-clean paired win rate. These results show that structured evidence transport provides an auditable, rank-oriented evaluation tool for high-stakes generated clinical text under ReXVal-only model selection and frozen RadEvalX testing.

Latest News

Latest News

Emergence World: A Platform for Evaluating Long-Horizon Multi-Agent Autonomy

Auditing Proprietary Alignment in Large Language Models: A Comparative Framework Without a Ground-Truth Standard

Impacts of Histories and Models on LLM Grading: A Study in Advanced Software Engineering Courses

FlashCP: Load-Balanced Communication-Efficient Context Parallelism for LLM Training

PACT: Self-Evolving Physical Safety Alignment for Diffusion Policies in Embodied Manipulation

AI Code Sandboxes: A Comparative Security Study. Part 1 of 2 -- Engine-Level Properties (Attack Surface, Leakage, Stackability, CVE History, Patch Cadence, Fuzzing)

Sycophancy as a Multilingual Alignment Failure: How Safety Degrades Across Languages, Topics, and Models

The Confidence Trap: Calibration Attacks for Graph Neural Networks

Not Just After One: Sleep-Inspired Replay Prevents Catastrophic Forgetting After Sequential Tasks

Reinforcement Learning for Flow-Matching Policies with Density Transport

Tyan-WP: A Wind Power Foundation Model for Ultra-Short-Term Probabilistic Forecasting

BioVid: Autoregressive Video Generation with Biological Behavior Semantic Comprehension

When Video Misreads: Closed-Loop Distillation of Reading Heuristics for Exploratory Manipulation Trace QA

Auditable Graph-Guided Root Cause Analysis for Kubernetes Incidents

HARBOR: A Harness Framework for Agentic Robot Reinforcement Learning

Reconstructing Synthetic SDO/AIA 193 A EUV Images from He I 10830 A Observations with Diffusion Model Translator

FiberTune: Preserving Action-Fiber Visual Residuals in Vision-Language-Action Fine-Tuning

Lost in the Flow with Code Talkers: Unveiling the Instruction-Tuning Tax of Large Language Models in Code Tasks

SNR-ST-Mix: Sample-specific Neighborhood Regression Mixup for Augmented Spatial Transcriptomics Imputation with Deep Neural Network

RadOT-Eval: Auditable Structured-Evidence Transport for Radiology Report Evaluation