We present CivBench, an open-source benchmark for evaluating language model agents in long-horizon, tool-mediated environments through the Model Context Protocol (MCP). A single episode can span over 300 turns and produces thousands of tool calls over a large action space, requiring sustained planning, state monitoring, and execution under partial observability. The environment exposes 76 MCP tools and a deterministic narration layer that retrieves and formats game state as structured text. We use CivBench to characterise agent behaviour across four model families in 23 admissible runs. The sample is a pilot, not a model ranking: aggregate outcomes do not reliably discriminate models at this scale. Instead, we introduce two interface-level metrics that the environment makes measurable: Proactive Monitoring Rate (PMR), capturing whether agents actively query latent strategic state, and RAG@10, capturing whether commitments stated in structured planning reflections are executed within ten subsequent turns. Across runs we observe two consistent patterns under a shared playbook protocol. Agents under-monitor strategically relevant state that is available but requires explicit querying: despite playbook guidance to query victory progress every 20 turns, agents do so only every 30-75 turns, and in 8 of 20 detectable defeats they failed to query within the 20-turn warning window before game end. Agents also frequently fail to execute near-term commitments stated in their own planning reflections (RAG@10 between 48.2% and 65.8% across models). Both patterns arise despite tool access and explicit guidance, and we interpret them as deviations under instruction rather than absences of capability. We release the environment, scenarios, logs, metrics, and analysis pipeline at https://github.com/lmwilki/civ6-mcp.
SAR-to-EO image translation aims to generate electro-optical (EO) imagery from synthetic aperture radar (SAR) observations. Existing latent diffusion approaches typically inherit a predetermined autoencoder, although reconstruction fidelity can vary substantially across codecs and modalities. Because the latent codec affects the round-trip preservation of both SAR conditions and EO targets, codec selection constitutes a fundamental design choice; nevertheless, existing methods largely rely on codecs pretrained on natural images. To remedy this, we introduce ReFlowSET, a conditional latent flow-matching framework that selects its codec through a joint SAR--EO reconstruction audit. Rather than inheriting a heavyweight pretrained generator, ReFlowSET trains a substantially smaller conditional DiT from scratch in the selected latent space, using dual-stream SAR conditioning followed by joint feature refinement. To provide semantic guidance for this from-scratch training, intermediate noisy-EO features are aligned with clean target-EO representations extracted by a frozen vision foundation model. This alignment is used only during training and introduces no additional inference cost. Experiments on QXS-SAROPT and SAR2Opt demonstrate state-of-the-art performance across diverse perceptual fidelity and distributional metrics. Code and pretrained weights are publicly available at https://github.com/KAIST-VICLab/ReFlowSET.
LLMs have been rapidly adopted across writing tasks, prompting the development of tools for detecting LLM-generated text. Yet, these tools largely measure how much of a document's surface text was written by an LLM and aren't fundamentally designed to measure how much of the information content or ideas originated from the LLM itself rather than being supplied by the user in the prompt. In this work, we design a framework that measures how much value a person adds on top of what a language model could have easily produced by itself. The method requires no training or labeled data and never scores the document's surface text, insulating it from stylistic confounders. Instead, it extracts the document's content at increasing levels of granularity, uses an LLM to reconstruct the document from each partial representation, and compares these reconstructions with those produced from the task description alone. We call this framework Value Over Language Model (VOLM), which measures a document's contribution relative to a replacement-level document that an LLM could produce from the task description alone. We evaluate VOLM with a specific instantiation of this framework across three domains: news articles, ICLR peer reviews, and argumentative essays. VOLM separates human-authored documents from matched LLM-generated documents produced from generic task descriptions, while remaining substantially invariant to content-preserving transformations, including LLM-based reconstruction and round-trip translation. We further find that increasingly constrained content extractors reduce residual differences between LLM-generated and humanized text, demonstrating the importance of disentangling informational content from stylistic variation. We hope these results encourage further work on specialized instantiations of the framework and on assessing human contributions in LLM-assisted writing more generally.
Understanding long-form video remains a fundamental challenge for multimodal large language models (MLLMs). Sparse frame sampling fails to capture fine-grained visual details, while dense sampling quickly exceeds context length limits. Retrieval-augmented generation (RAG) offers a promising middle ground by selectively retrieving relevant video segments for grounded generation, yet its effectiveness critically depends on the quality of the video segments used as retrieval units. In this paper, we investigate RAG for movie understanding, which demands story-level reasoning over characters, events, and narrative arcs spanning hours of content. Scene segmentation, a long-studied problem that partitions movies into semantically coherent units, is a natural candidate for defining such retrieval units. We reexamine whether existing methods actually serve this role through comprehensive evaluation on downstream movie understanding tasks, and find that they do not provide a consistent or dominant advantage over naive uniform temporal chunking. Our audit of the most standard scene segmentation benchmarks reveals why: current annotations prioritize visually salient transitions over narrative event structure. Motivated by this mismatch, we introduce NarraScene, a narrative-centric scene segmentation dataset annotated with a three-level cognitive taxonomy spanning physical, character, and narrative change, where every valid boundary requires a narrative-level shift. When used as retrieval units, these narrative-grounded segments outperform uniform chunking on downstream movie understanding tasks, suggesting that the central challenge for scene segmentation in movie RAG is not detecting boundaries, but identifying the narrative event units that matter for movie understanding. The dataset is available at https://huggingface.co/datasets/queez0405/NarraScene.
Parkinson disease PD is a progressive neurodegenerative disease that can have a significant impact on motor performance resulting in the appearance of symptoms such as tremors rigidity postural instabilities and bradykinesia. Timely clinical treatment disease management and quality life of the patients are closely linked to early and appropriate identification of PD. Over the past few years the growth of wearable sensor technology and artificial intelligence AI have made it possible to create noninvasive and data driven disease detection methods. This paper proposes a comparative system using artificial intelligence to detect Parkinsons disease by analyzing the motion and tremor data captured by an inertial measurement unit IMU. The data comprises the signals of the acceleration and gyroscope sensors measuring movement in three directions X Y and Z. The signs and symptoms provide helpful information about subtle motor deficits associated with PD. Several classification models like Support Vector Machine SVM Logistic Regression LR KNearest Neighbors KNN Decision Tree DT Extreme Gradient Boosting XGBoost and Light Gradient Boosting Machine LightGBM were used to compare their effectiveness. The Logistic Regression model had a performance around 75 percent in all evaluation metrics and KNearest Neighbours KNN around 90 percent. The support vector machine SVM performed almost 94 percent whereas the performance of classifiers such as Decision Tree and XGBoost was close to 96 percent and overall classification efficacy respectively. LightGBM model performs consistently at the best rank among all of the evaluated methods having Accuracy, Precision, Recall and F1score of around 97 percent. The results show that the proposed machine learning approach offers an accurate and effective predictive capability in the classification of PD severity.
Kolmogorov-Arnold Networks (KANs) learn univariate functions on edges, making the choice of edge response central to their design. Existing variants use basis expansions such as splines or polynomials, whose functional form is chosen before training. We introduce geometry-constrained KANs, which instead learn the shape of power-law edge responses through an exponent inspired by Banach duality maps. Banach-KAN combines this adaptive power-law response with a bounded hyperbolic tangent (tanh) component. We relate these responses to convex potentials and bound edge sensitivity on bounded input ranges. Across symbolic regression benchmarks, Banach-KAN is competitive with fixed-basis models, particularly with limited clean data. Learned power-law responses remain competitive under measurement noise, while splines are stronger with plentiful clean data. Fixed-exponent and branch-removal controls on a core subset support learning the exponent and retaining both Banach components. Learned exponents describe the shapes of the fitted edge responses.
Organisations deploying AI face a scoping problem: which tasks can be automated, which should remain with humans, and which are best shared between the two. Aggregate benchmark scores provide little insight into where systems will succeed or fail in practice, while human judgements of model capabilities quickly become outdated. We introduce a pipeline that profiles agents and tasks using a shared set of core cognitive capabilities. Cognitive capability profiling infers an agent's capabilities from performance on a benchmark battery annotated for the cognitive demands of each item. Task requirements weighting elicits from domain experts the relative importance of these same capabilities for their work. As both use a common set of cognitive dimensions, they can be updated independently as models and roles change, and combined to estimate AI suitability at the level of a domain, organisation, role, or individual duty. We validate capability recovery on synthetic agents, profile six AI systems, and elicit task requirements from 410 employees across six occupational domains. AI systems differed more across cognitive dimensions than across model families, while workplace activities converged on a shared cognitive core. The resulting scores provide a comparative scoping tool for identifying promising candidates for piloting and areas where current systems are unlikely to be well suited. We discuss extending the framework to profile human workers alongside AI systems, moving from AI suitability towards human-machine task allocation.
Every deployed KV-cache-eviction rule and every top-$k$ sparse-attention rule keeps a subset of the keys, discards the rest, and renormalizes the attention weights over the kept set. Identifying the exact best subset under that constraint, we find that keeping the largest weights is already near-optimal, since the best subset we find closes only a median $2$ to $5\%$ of the remaining gap to full attention. We evaluate deployed eviction methods in KVPress and find that methods with non-uniform per-head budgets apply attention masks over the full cache because dense cache tensors require equal length across heads. Truncating each head to the target budget lowers benchmark scores by $14$ to $62$ points, and computing token importance with the question in context accounts for $87.6$ points in retrieval accuracy. We propose ContourKV, a training-free rule that applies KVzip's score under a physically enforced budget. It beats KVzip in $93$ of $160$ benchmark runs across RULER and LongBench while using $28$ to $1078$ times less memory, and ties Compactor, the strongest baseline at equal memory.
The advancement of Embodied AI necessitates high-quality simulation assets that faithfully mirror the real world. However, transforming raw visual observations into simulation-ready scenes remains challenging due to the lack of physical grounding and scene-level interactivity in current image-to-URDF methods. We propose NeoWorld-Pro, a framework that reformulates monocular scene reconstruction as procedural programming for interactive 3D environments. Leveraging the zero-shot reasoning and code synthesis capabilities of MLLMs, NeoWorld-Pro converts a single RGB image into executable programs specifying object geometry, articulation, and physical properties. A physics-in-the-loop mechanism then iteratively refines the generated programs by validating their execution in a physics engine, enforcing physically plausible articulations, valid object compositions and interactions, and accurate spatial relationships. Experiments show that NeoWorld-Pro outperforms open-loop and prior monocular reconstruction methods, while enabling complex downstream tasks such as stable stacking and fine-grained manipulation.
Researchers must often choose between Bayesian or neural network models of behavior, two paradigms with complementary strengths and weaknesses. An ideal paradigm would facilitate testing many kinds of representations and inductive biases; Bayesian models make this easy, while neural networks do not. Similarly, an ideal paradigm would avoid over-simplifications; neural networks make this easy, while Bayesian models do not. Here, we introduce Bayesian distillation with Behavioral Tuning (BBT) as an approach to getting the best of both traditions. BBT offers a simple recipe for model building: first, a neural network is trained to mimic a Bayesian model through synthetic data, and second, the network is fine-tuned on human behavior to capture additional structure and nuance. Across four case studies in human concept learning, we find that BBT outperforms traditional approaches at predicting human behavior while also revealing psychological insights, resulting in models that can both mimic Bayesian priors and capture heuristics and biases that violate simple modeling assumptions.
Retrieval-Augmented Generation (RAG) grounds Large Language Model (LLM) outputs in external knowledge, but RAG systems usually trust whatever they retrieve, creating a Security-Reliability Gap: high semantic relevance does not guarantee factual truth. Adversaries exploit this through knowledge poisoning, inserting malicious documents to cause targeted misinformation. We propose an Evaluation Agent, middleware that combines Natural Language Inference (NLI) factual verification, a five-signal poison detector with relevance-weighted aggregation, and a Trust Index T = 0.4 F + 0.35 C + 0.25 (1 - P ) with a non-linear dampener for high-contamination contexts. On TruthfulQA with Llama 3.3 70B, the agent reaches 91% accuracy and 100% precision, with 100% recall on instruction injection, while in-place edits, such as entity swaps, remain hard to detect. Across three LLMs the Trust Index stays discriminative, with a Receiver Operating Characteristic Area Under the Curve (ROC-AUC) of 0.73 to 0.81; generation style matters more than model size, and per-LLM threshold calibration restores baseline competitive accuracy, whereas a weaker FEVER result shows that cross-dataset generalization requires domain-specific calibration. In a software-engineering use case, a secure-coding assistant over guidance from the Open Worldwide Application Security Project (OWASP) Top 10 and the Common Weakness Enumeration (CWE), the agent reliably blocks instruction injection of unsafe advice (F1 92%), while contradiction and subtle semantic weakening remain hard. Throughout, the agent measures detection of poisoned context before generation, not whether the LLM adopts the injected misinformation. We release the proposed approach, attack generator, and experimental artifacts at the link: https://github.com/GPT-Laboratory/TrustworthyRAG.
Cyclic peptides are emerging as promising molecular scaffolds in drug discovery due to their high binding affinity and structural stability. However, extending generative models from linear to cyclic peptide design remains challenging, as cyclization sharply restricts the feasible design space through coupled geometric and biophysical constraints. Moreover, limited training data has led existing approaches to rely largely on zero-shot generation or post hoc filtering, resulting in low yields of feasible designs and limited control over multi-objective trade-offs. To address these limitations, we propose FAR-DPO (Feasibility-Aware and Robust Direct Preference Optimization), an architecture-agnostic framework that steers generative models toward structurally and biophysically feasible cyclic peptide designs, particularly for challenging targets. FAR-DPO integrates feasibility-aware preference construction with difficulty-aware group-robust optimization. Specifically, it constructs within-target preference pairs through feasibility-gated multi-objective dominance and adaptively reweights predefined difficulty groups according to their current preference losses. Under a fixed generation budget, FAR-DPO increases overall success rate on the CPSea benchmark from 46.9% to 57.8% on PepGLAD and from 48.0% to 49.6% on PepFlow. These gains also extend to the hardest target quartile and are accompanied by more favorable best-per-target binding scores. Together, these results demonstrate FAR-DPO's effectiveness in improving feasibility and target-wise robustness.
Watermarking is a central tool for provenance in generative models, yet its application to multivariate time series remains hindered by reliability failures under post-editing attacks. We show that existing detectors, which rely on globally coupled re-encoding, suffer from bidirectional drift of the null distribution: post-editing attacks can shift the $z$-score of non-watermarked samples in either direction, invalidating clean-calibrated thresholds. We argue that this instability is a property of the re-encoding, and that reliable detection requires each recovered unit to depend only on a bounded temporal neighborhood. Guided by this principle, we propose L-VQVAE, a generative model in which each discrete token is produced from a short contiguous window, and LVQMark, a watermarking method over this token space that combines logit-bias injection with robust re-encoding for attack-time detection. Experiments on four benchmarks spanning finance, energy, and neuroimaging show that our approach preserves generation quality while stabilizing both detection power and false-positive behavior under post-editing attacks.
Training fast generators on new domains with limited data remains challenging for two reasons. First, adapting a pretrained diffusion or flow model leaves its costly multi-step sampling intact, and existing acceleration methods are tied to the source parameterization (x, epsilon, v, or u), so heterogeneous pretrained models share no common acceleration target. Second, while adversarial post-training sharpens generation quality, existing criteria each cover only one end of the step range. Discrete criteria compare jump endpoints and lose their signal as the jump shrinks, while continuous criteria score the instantaneous velocity and misjudge the large jumps of few-step sampling. Neither fits MeanFlow (MF) models, which predict average velocities over intervals of any length. We propose MeanFlow-Transfer (MF-T), which adapts a source domain model of any parameterization and compresses it into a few-step MF generator through a common velocity map. We then introduce Continuous Adversarial MeanFlow (CAMF), which scores the change of a learned potential along each predicted interval jump of MF. CAMF needs no gradient penalty or Jacobian-vector product (JVP). We also prove it stays well-conditioned for every interval length and recovers the instantaneous criterion as the interval vanishes. Adapting four ImageNet-based source models (DiT, SiT, JiT, iMF) to five target domains, MF-T + CAMF approaches or exceeds the fine-tuned many-step teacher in FID and FDD at up to 125x fewer Neural Function Evaluations (NFEs). Further, CAMF improves MF-T's few-step FID by about 30% on average, outperforming prior adversarial post-training on iMF at every NFE.
We address two open questions in streaming PCA via Oja's algorithm: sharp operator-norm convergence for general rank under sub-Gaussian data, and distributional inference for the resulting subspace estimator. Existing convergence analyses, even in the rank-one case, either assume bounded data or leave non-vanishing remainder terms that prevent adaptation to a polynomially vanishing tail spectrum, while existing distributional results are confined to the rank-one case. Our convergence theory removes these remainder terms and yields a sharp rate. In the dense-tail spiked covariance regime, this rate matches the minimax rate up to logarithmic factors. More generally, we prove a matching lower bound, up to logarithmic factors, across both dense-tail and sparse-tail regimes under a mild nondegeneracy condition. The analysis yields a linearization of Oja's iterates, which in turn enables a high-dimensional Gaussian approximation for the general-rank subspace estimation error with an explicit limiting covariance. We also establish a row-wise Gaussian approximation over convex sets for the aligned difference, recovering prior rank-one results as special cases. For practical inference, we develop an online multiplier bootstrap algorithm and prove its consistency. Beyond streaming PCA, our techniques contribute to Gaussian approximation and bootstrap inference for nonconvex stochastic approximation.
General-purpose motion trackers enable humanoid robots to follow diverse whole-body motions while maintaining balance, but are trained only on flat ground, failing to exploit bipedal mobility over complex terrain. Cross-terrain controllers, meanwhile, are task-specific or accept only low-dimensional locomotion commands. We introduce InterTrack, the first behavior world model (BWM) for robust whole-body tracking with environment interaction. Its Transformer jointly predicts the next action, state, and behavior distribution, learning environment-conditioned dynamics. To scale interaction training data, an automatic annotation pipeline reconstructs 3D support geometry from retargeted motions. At deployment, the policy handles commands implausible in the current environment in a "best-effort" manner. Quantitatively, InterTrack achieves an 81.3% success rate on terrain interaction (4.3 times the best evaluated baseline) and a 99.3% fall-recovery rate, while also improving free-space tracking and outperforming three leading tracking baselines across all of these regimes. To our knowledge, we provide the first demonstration of real-time cross-terrain whole-body teleoperation on a humanoid robot, alongside object interaction, stable responses to missing supports, and robust recovery from falls.
Reasoning has become a standard technique and feature for contemporary LLMs; however, its application and quality in the context of demanding legal-oriented tasks, such as legal case forecasting, remain under explored. We investigate how LLMs reason in the context of legal case forecasting, using legal cases from the European Court of Human Rights (ECtHR) as a testbed. We evaluate OpenAI GPT 5.4, a recent top-tier LLM, by exploring alternative prompting strategies that are more or less suggestive of what counts as legally meaningful reasoning in the context of ECtHR jurisprudence. We present our findings derived from assessing the model's responses with both human and LLM evaluation. We find that the examined model scores far from ideal in legal reasoning, the model produces structurally complete but substantively shallow analyses, and that LLM-as-a-Judge evaluators are internally consistent yet align only weakly with our trained annotators, i.e., reliable but not a valid substitute for human evaluation. Overall, the expert-curated prompt leads to more comprehensive reasoning, which does not result in more accurate predictions compared to the other examined settings. Based on our findings, we urge the community not to rely solely on automated LLM-based evaluation and to avoid using task accuracy as an appropriate proxy for reasoning quality.
Large Language Model (LLM) agents deployed in production environments face a fundamental tension: the agent's behavior is frozen at deployment time, while the business rules and edge cases it must handle continue to evolve. Existing approaches address agent construction and one-time evaluation but provide no structured mechanism for continuous post-deployment behavioral correction without modifying the agent's source code. Most of the approaches offered in the market, require intense collection of logs and traces, and re-examining the agent design by the engineering team, a process which is heavy, long and negates the economical value of agentic transformation. We introduce Agent Gym, a modular, domain-agnostic framework that wraps any existing LLM-based agent in a continuous evaluation-and-evolution loop. The framework provides six composable capabilities --- Act, Evaluate, Investigate, Correct, Learn, and Observe --- organized across three architectural zones: a constitution layer that codifies domain knowledge in configuration artifacts, a runtime inference pipeline that chains acting, investigation, and adaptive correction, and a learning loop that enables subject matter experts to discover and validate new correction rules through natural language interaction. The key technical contributions include a hybrid deterministic-LLM correction engine with 21 condition operators and three-tier actions, a three-layer investigation architecture for ground-truth-free compliance validation, and a programmatic safety loop that guarantees rule correctness before human approval. We further introduce the Spec-to-Note Gap, an autoencoder-inspired view of agentic system transparency. An open-source reference implementation for invoice processing demonstrates that the framework is fully operational and ready for adoption.
Reinforcement learning (RL) post-training provides a direct way to align diffusion models with human preferences and task-specific rewards. However, current RL algorithms for diffusion models remain fragmented: reverse-trajectory methods rely on discretized likelihood ratios, whereas forward-matching methods train on reward-labeled noising versions of the rollout samples. This paper shows that these seemingly different losses arise from a single path-space principle. Starting from the regularized diffusion-RL objective, we use importance sampling between sampling SDEs to obtain an explicit policy-gradient estimator on trajectory space. The estimator contains the stochastic It\^o integral underlying Flow-GRPO-type updates; we derive an equivalent variance-reduced value-gradient form that recovers the forward-matching structure of AWM and DiffusionNFT. This identifies the empirical gap between these method families as a variance-reduction effect rather than a difference in RL principle. The derivation yields a unified design space organized by value-gradient estimation, weight functions, and sampling choices. Within this space, we propose a multi-sample KDE value-gradient estimator that reuses rollout groups, together with scale-bounded weight families that retain stable existing recipes while excluding singular ones. Experiments on SD3.5-M and Qwen-Image models validate the variance-reduction explanation and show that the resulting recipe improves over prior diffusion-RL baselines.
Cytometry measures the complex characteristics of single cells (e.g., counts and protein expression of immune cells) and is widely used across immunological research and clinical settings. However, cytometry data is highly heterogeneous and unstandardized due to experimental protocols and the choice of measured features. While machine learning methods hold the potential to gain deeper insights into cell biology, these challenges make them difficult to apply and transfer across studies. Recent advances in foundation models can alleviate these issues, but corresponding approaches are still scarce in this field. To address this, we provide CytoBERT, a publicly available, open-source, open-weight foundation model for single-cell cytometry data with variable marker panels. CytoBERT is pretrained in a self-supervised manner on a large-scale cytometry corpus (15 human datasets with heterogeneous marker panels and more than 50 million cells) curated through marker standardization, enabling it to learn transferable inter-marker relationships within cells. Fine-tuning CytoBERT for sample-level classification demonstrates that transfer learning across heterogeneous cytometry datasets is feasible, providing a starting point for scalable, generalizable cytometry analysis. Code is available at GitHub.