Papers

The wrist plays a critical role in upper-limb function by enabling precise hand positioning, force regulation, and object manipulation. Continuous brain–muscle interfaces (BMIs) offer a promising approach for motor restoration by decoding neural activity into muscle activation signals. However, existing EEG-to-EMG models have mainly been developed for tasks with relatively stable muscle synergies and may be less effective for the heterogeneous and weakly coupled neuromuscular organisation involved in wrist movements. This paper proposes ST-Topo GAN, a Spatial–Temporal Topological Generative Adversarial Network for continuous EEG-to-EMG decoding of wrist movements. The framework integrates multi-band EEG representation, sensorimotor cortical topology modelling, and conditional adversarial learning to reconstruct multi-channel iEMG activation. The model was evaluated through cross-task comparison, wrist EEG-to-iEMG decoding, and ablation experiments. Compared with the WAY-EEG-GAL grasp-and-lift dataset, the wrist dataset exhibited lower inter-muscle activation similarity and greater decoding difficulty for conventional models. ST-Topo GAN achieved an average PCC of 0.4436 on the wrist dataset, outperforming all evaluated baselines, while the ablation study confirmed the contribution of its key components. These results support the effectiveness of ST-Topo GAN for continuous wrist EEG-to-iEMG decoding.

Deep learning registration methods routinely stack two kinds of enhancement on a base network: architectural additions such as affine pre-alignment stages, and training-objective additions such as regularization losses. Papers tend to adopt both at once, so it is unclear which is doing the work. I ran a controlled ablation to separate them. Using the OASIS brain MRI dataset (394 training subjects, 20 test subjects), I trained four variants of the same registration pipeline: a baseline 3D U-Net with basic similarity losses, the same U-Net with a full regularization suite, an affine-plus-deformable architecture with basic losses, and the affine architecture with the full suite. I evaluated registration accuracy (MSE, NCC, SSIM), deformation quality (Jacobian determinant preservation, displacement statistics, an anatomical plausibility score), and computational cost. Regularization alone accounted for most of the gain: a 21.3% relative gain on the MSE-improvement metric (1.78% to 2.16%, P<.001) and a 21.8% relative gain in NCC improvement, while cutting maximum deformation from 53.1 to 0.51 units, a 99.0% reduction, at essentially no computational cost (-0.06% inference time). The combined model produced the largest accuracy gain, 25.8% (1.78% to 2.24%), and raised anatomical plausibility from 0.596 to 0.930, at a moderate +9.8% inference-time cost. Gradient correlation rose from 0.742 at baseline to 0.980 for the fully enhanced model. All enhanced variants reached sub-voxel accuracy under plausible deformation constraints. Regularization losses are the primary driver in this setting, delivering the accuracy gains and almost all of the deformation control for free at inference time, while the affine architecture adds a smaller complementary benefit at acceptable cost. The 99% reduction in unrealistic deformations addresses a known barrier to clinical deployment.

In recent years, both academia and industry have focused on the development of computational architectures inspired by the distributed, adaptive, and event-driven characteristics of biological neural systems, with the aim of reducing the computational cost associated with conventional training approaches [1]. However, a major challenge is the lack of general models and design guidelines for emerging computational systems and hardware. This work introduces a general model based on an input-dependent stochastic weight network, referred to as a substrate. The substrate weights evolve through input-triggered stochastic updates, with correlations between weight coefficients described by a matrix-valued covariance kernel. The proposed framework is implemented using quadratic polynomial weight functions, where the input amplitude controls the magnitude of the stochastic perturbation and a substrate-dependent distance determines the correlation structure. Numerical simulations show that correlations in the stochastic weight evolution significantly affect the system response, suggesting a potential mechanism for neuromorphic-inspired computation without conventional weight training. The aim of this work is to provide a general formulation of the model and identify its main properties and characteristics. 1 H. Jaeger, Towards a generalized theory comprising digital, neuromorphic and unconventional computing, Neuromorphic Comput. Eng., vol. 1, no. 1, p. 012002, Sep. 2021

Inter example relational distillation transfers a teacher's representation geometry by matching relations among examples within a mini batch. Computing all pairs has quadratic complexity in the batch size, whereas uniform subsampling may use a limited relation budget inefficiently. We introduce Reliability Aware Pair Importance Distillation, or RAPID, which separates a reliability gated relational target from a full support adaptive pair proposal. Reliability determines which teacher relations are emphasized, while calibrated teacher entropy and detached student-teacher residuals determine which relations are evaluated. Exact inverse proposal correction makes the loss and gradient estimators conditionally unbiased with respect to the gated mini batch target. We evaluate RAPID in two text classification settings: AG News with BERT-to-DistilBERT distillation using three paired seeds and a relation budget of 256, and SST-2 with DistilBERT to DistilBERT distillation using three paired seeds and a relation budget of 64. Reliability gated relational distillation achieves the highest observed mean student accuracy on both datasets: 94.285 plus or minus 0.054 percent on AG News and 88.800 plus or minus 0.532 percent on SST-2. RAPID ranks second, achieving 94.241 plus or minus 0.025 percent and 88.685 plus or minus 0.462 percent, respectively, compared with 94.154 plus or minus 0.124 percent and 87.271 plus or minus 0.162 percent for the cross entropy baseline. Pilot evaluations are counted toward the same total budget as the main relation evaluations. Across both settings, the gated target yields the highest mean accuracy, while the adaptive proposal remains within seed-level variation. These results support the modular view that target reliability and evaluation priority are separable design dimensions.

Clinical diagnosis relies on consistent multimodal data collected from the same patient throughout the disease course, yet such data are difficult to acquire at scale because of collection costs, missing modalities, fragmented systems, and longitudinal follow-ups. Existing synthetic-data approaches largely focus on individual modalities or vision-language dual modalities at report-level generation. Little work has been done to construct synthetic data with consistent patient backgrounds, coherent disease trajectories, and interrelated modality-specific evidence at a complete clinical case level. We introduce CaseWeaver, a multi-agent framework built around a timeline-anchored Latent Clinical Case Graph (LCCG). The LCCG organizes patient context, latent disease states, clinical events, and expected observations in a shared patient-level representation. Modality-agents use scoped observation subgraphs and clinical protocols to generate evidence including clinical records, laboratory results, physiological signals, and medical images. We evaluate clinical inferability using a calibrated AgentClinic protocol and case diversity using Virtual Case Diversity (VCD) score. CaseWeaver outperformed general-model and agentic-workflow baselines on both metrics, producing more diverse and coherent multimodal virtual clinical cases.

Geometric community detection seeks to recover latent communities in networks where connectivity depends jointly on community structure and continuous spatial geometry. Existing exact-recovery approaches typically discretize the underlying space, which can impose restrictive structural assumptions on the connectivity functions. We develop a polynomial-time, discretization-free algorithm for exact recovery in the Geometric Hidden Community Model (GHCM), operating directly on the continuous geometry. Our method succeeds even when connectivity functions coincide on a nontrivial portion of their visibility range and when two communities can be distinguished only through their connectivity to a third community. We prove exact recovery under these weaker conditions and provide experiments showing that the algorithm succeeds beyond the theoretically guaranteed regime.

Video-language models (VLMs) remain brittle on tasks that require tracking events over time and grounding answers in specific spatial regions. We propose that part of this limitation can be addressed through better organization of visual evidence at inference time. We introduce structured video prompting, a training-free inference-time method that augments the input video with lightweight spatial structure and temporal structure, providing explicit anchors for organizing evidence across space and time without changing model weights or decoding and without altering the question prompt in the main comparison. We evaluate this approach on two complementary video benchmarks and two open video-language models. Across these settings, structured inputs improve performance in several cases, with gains varying by model and task. Our findings suggest that some failures of VLMs arise not only from reasoning capacity, but also from how video evidence is presented at inference time. These results highlight structured video prompting as a simple and practical direction for improving video understanding.

Recent open-vocabulary segmentation models have significantly advanced semantic perception for UAVs, for aerial and remote-sensing imagery, yet predictions from moving platforms can remain temporally inconsistent across repeated observations of the same physical scene. While semantic segmentation is typically evaluated on single images, temporal reliability remains less explored despite its importance for long-horizon robotic perception. In this work, we investigate temporal semantic stability by associating frame-wise predictions with persistent world-space locations through metric 3D fusion. We introduce a voxel-level evaluation framework that jointly characterises final semantic agreement, Semantic Belief Drift (SBD), Observation Persistence (OP), and semantic uncertainty from accumulated evidence. Experiments on UAVid-3D-Scenes reveal substantial frame-wise semantic variation and show that high aggregate world-space agreement can overstate temporal stability when many locations have limited repeated-observation support. Persistence-stratified analysis shows that recurrent voxels expose substantially greater semantic disagreement, while higher-persistence voxels exhibit lower direct consecutive semantic-evidence divergence. This behaviour is observed across two segmentation backbones and remains consistent under variations in voxel resolution, geometric association, and temporal sampling density. Overall, conditions that reduce world-space recurrence can increase apparent aggregate stability, demonstrating that semantic consistency must be interpreted together with observation support. Our findings establish observation persistence as an essential conditioning variable for evaluating long-horizon semantic reliability and provide a framework for analysing temporal semantic behaviour beyond image-level prediction.

Despite the success of foundation models in language and vision, their expansion into embodied AI is bottlenecked by a lack of generalized touch sensing. This limitation is especially relevant to consumer electronics, where smartphones, wearables, VR controllers, home robots, and health monitoring devices require safe and adaptive physical interaction. Constrained by hardware heterogeneity and the necessity of active physical data collection, current haptic models remain rigidly task-specific. To overcome these limitations, this article explores the transformative potential and developmental trajectory of Haptic Foundation Models (HFMs). We detail the paradigm shift required to transition from passive Large Language Models and Vision Language Models into active HFMs across four core dimensions: action coupling, physical dynamical representation space, continuous time-series data granularity, and action-conditioned future state prediction. Furthermore, we synthesize existing large-scale tactile datasets and benchmark UniTouch, AnyTouch, T3, and Sparsh on TacBench for force estimation, slip detection, and relative pose estimation.

Knowledge-intensive multi-hop question answering requires systems to select evidence and compose dependent facts, yet multilingual benchmarks usually translate an entire example into one language. This hides failures at language boundaries inside the reasoning chain. We introduce XHotpotQA, a controlled benchmark for cross-lingual knowledge composition over mixed-language evidence. Each instance is modeled as an evidence-dependency graph whose question, bridge evidence, answer-bearing evidence, and distractors have explicit language assignments. The audited resource contains 15,661 training and 7,405 validation instances, with sentence-level support supervision and supplied distractors. In validation, 99.81% of items cross the question-to-gold-evidence language interface and 95.60% use gold paragraphs in different languages. Across three reader artifacts, full question-evidence mismatch is associated with 10.25 to 15.79 lower Unicode-aware answer F1 than partial alignment, and different-script evidence with deficits of 11.98 to 23.70 points; the corresponding adapted-selector contrasts are 1.71 and 1.78 points. Under this supplied-candidate design, the evaluated readers therefore show substantially larger condition-associated deficits than the selector. XHotpotQA provides role-aware diagnostics, modular evaluation, and an audited test bed for knowledge-based systems that must integrate evidence across languages.

World models enable agents to reason about future outcomes and learn policies from their knowledge of state transition, but existing approaches primarily focus on reconstructing future observations or features, which introduces unnecessary complexity and limits their effectiveness for decision making. In this work, we propose a compatibility prediction Latent World Model (LWM) for robot navigation that predicts action-conditioned latent feature compatibility rather than reconstructing observations. Our key insight is that spatial proximity correlates with latent feature similarity, enabling action consequences to be evaluated directly in latent space. To support counterfactual training, our model leverages action sequences sampled across trajectories and learns to predict which sequences lead closer to the goal. Furthermore, we demonstrate how the learned world model can supervise policy learning from unlabeled video data and further improve policies through reinforcement learning entirely within the world model. This imagination-driven framework eliminates the need for action annotations and additional environment interaction. Extensive experiments on multiple real-world robot navigation datasets show that our approach significantly outperforms prior world model and imitation learning methods in prediction accuracy, policy learning, and real-world navigation performance. The code, pretrained models, and additional materials are available at https://wzm206.github.io/latent-world-model-nav.

Tool-using agents fail two ways: choosing the wrong tool, or forming wrong arguments, and an early failure of either kind can silently corrupt everything downstream. We measure a correct-invocation rate that separates the two, under both a clean teacher-forced context and the model's own free-running context, on five open-weight models over contamination-free multi-step tasks (depths 1-8). By depth 6, roughly 70% of a model's own clean-context capability is lost to its own earlier mistakes (L6 = 0.686, 0.684). Our central finding concerns the measurement itself. Under exact-match scoring against a fixed gold trajectory, a propagation model's severity and recovery parameters are not merely hard to estimate - they are fixed by the scoring rule. Severity is forced to its boundary (0 of 869 poisoned steps correct); recovery is structurally unobservable (0 of 580 poisoned steps returned on-track, against an expected 0.0058 by chance). Both follow from one mechanism: post-divergence, the gold value is generated by tool constants the model never sees, so it is information the model cannot derive. A fit run anyway returns 0.92 and 0.73 for a quantity that is exactly 1.000 - confident numbers for a parameter the scoring rule already determined. We give the mechanism and a remedy, conditional-on-state scoring, applied retrospectively to cached completions at zero additional cost, which un-pins severity to interior estimates excluding zero (+0.149, +0.316).

Large language models are increasingly used to inform safety decisions in cities, such as where it is safe to walk, rent, or travel. We ask whether such judgments track measured risk or the patterns attached to an urban neighborhood's name. We probe seven instruct-tuned models under three conditions that dissociate name from geography: coordinates-only, name-only, and name+coordinates, across 186 neighborhoods in Los Angeles and Chicago, joined to violent crime and American Community Survey data. First, ratings are nearly flat under coordinates for six of seven models, while names carry most between neighborhood variation and are moderately calibrated to violent crime; only at frontier scale does the coordinate channel show appreciable variation. Second, names lower safety ratings more for neighborhoods with higher shares of the locally dominant marginalized group (percent Black in Chicago, percent Hispanic in Los Angeles), and this name effect tracks demographic share in all seven models and both cities. In Los Angeles, where demographic share and crime are more separable, the effect survives controls for crime and income and is confirmed by crime-matched pairs. An enforcement-elasticity analysis further shows that over-caution tracks near-fully-reported homicide rather than discretionary, deployment-driven offenses. Third, the effect scales with geographic knowledge: models that better distinguish real neighborhoods apply more demographic stereotype to them. Because neighborhood names carry both genuine crime signal and demographic stereotype, removing names reduces both bias and accuracy. We discuss implications for deploying LLMs in advice and decision-support settings.

We present GreenLeaf Law Embed Tiny, a 0.6B parameter embedding model for legal domain retrieval. GreenLeaf-Tiny achieves 75.11% on the Massive Legal Embedding Benchmark (MLEB) and 64.38% on MTEB(Law, v1),demonstrating competitive performance among models under 1B parameters. Our approach combines a two-stage training pipeline that first distills knowledge from a larger teacher model into a compact student architecture, then applies domain-specific fine-tuning with hard negative mining; a carefully curated dataset of 3.4 million query-passage pairs, including 150,000 human-curated samples across diverse legal jurisdictions; and an efficient inference architecture supporting multiple quantization levels (BF16, INT8, binary) enabling deployment in resource-constrained environments. We provide detailed analysis of our training methodology, architectural choices, and comprehensive evaluation across legal retrieval tasks. Our results demonstrate that domain-specific training with high-quality data can improve performance for specialized domain applications

Accurate identification of early-stage apple fruitlet anatomical structures, including the calyx, fruitlet body, and peduncle, is essential for robotic thinning, crop-load management, and other precision orchard operations. This study presents a lightweight multimodal vision-language framework that adapts TinyCLIP for fine-grained fruitlet anatomy classification in complex orchard environments. A dataset of 600 high-resolution RGB images collected from Scilate and Scifresh apple orchards was converted into 224 x 224 image patches and annotated for three anatomical classes. Domain-specific language prompts, such as “a photo of a class,” were used to guide multimodal alignment between orchard imagery and horticultural structures. A sliding-window inference strategy with a stride of 112 pixels aggregates patch-level predictions into spatial heatmaps, enabling interpretable whole-image localization of fruitlet components relevant to robotic thinning. Patch-level evaluation on an NVIDIA T4 GPU achieved F1-scores of 0.95 for calyx, 0.98 for fruitlet, and 0.85 for peduncle, with a macro-F1 score of 0.93. Deployment-oriented optimization using ONNX and TensorRT enabled efficient inference on NVIDIA Jetson hardware, preserved accuracy under INT8 quantization, and supported model sizes of approximately 127-137 MB with millisecond-level patch inference. These results demonstrate that lightweight vision-language models can provide interpretable and edge-deployable perception for automated fruitlet analysis and future robotic thinning systems. The source code and implementation details are publicly available at https://github.com/WilliamBu1/A-Lightweight-Vision-Language-Model-for-Early-Stage-Fruitlet-Classification-in-Apple-Orchards.

Accurate field plant disease diagnosis requires reliable fusion of uncertain and conflicting perceptual evidence. We present the Hybrid Hierarchical Multi-Agent Framework (H$^{2}$MAF), combining decision-level fusion of EfficientNet-B3 and ConvNeXt-Tiny with semantic arbitration by open-weight multimodal large language models (MLLMs), Gemma 4 E4B and Qwen3.5 4B, using structured JSON evidence to generate explainable diagnoses, risk levels, treatment urgency, and financial exposure. (H$^{2}$MAF) is evaluated on 14,364 images (1,370 test images) across PlantDoc (2,922 images, 27 classes) and two non-public, continuously captured Cornell robot-acquired field datasets: Stage 2 (20 GB; 4,215 images) and Stage 4 (40 GB; 7,227 images), covering Early Blight, Late Blight, and Septoria Leaf Spot under uncontrolled field conditions. On PlantDoc, Gemma improves accuracy from 63.9% to 68.5%, achieving +7.6 points on the 41.7% CNN-conflict subset. Cornell accuracies reach 99.3% and 98.9%, with only 1.7-4.1% disagreement, demonstrating conflict-dependent MLLM utility. The critical-risk error of gemma is 0.14-0.5 points, whereas Qwen overflags by 3.5-14.4 points. These results establish MLLM arbitration as a promising, yet calibration-dependent, approach for explainable agricultural AI and robotic field decision support. Github Link: https://github.com/Applied-AI-Research-Lab/Explainable-AI-Plant-Disease-Detection

Multimodal neuroimaging combining structural MRI and positron emission tomography (PET) captures complementary structure-function relationships across the Alzheimer's disease (AD) continuum, yet existing artificial intelligence systems produce a single diagnostic label without quantifying which imaging modality drove that decision for a specific patient. We introduce the Modality Contribution Network (MCNet) and the Modality Contribution Score (MCS), the first per-patient attribution framework quantifying the shift in modality dominance from structural atrophy to amyloid and metabolic dysfunction across the cognitively normal to MCI to AD continuum. MCS is normalised to unity per subject via modality ablation (MCS_MRI_i + MCS_PET_i = 1.0 for every subject i), providing an interpretable, clinically actionable score that fluid biomarkers cannot supply. Applied to 327 ADNI-3 participants balanced across cognitively normal, mild cognitive impairment, and AD groups, MCNet achieved competitive three-class staging performance (AUC=0.881). The MCS revealed a statistically significant monotonic gradient (Kruskal-Wallis p<0.0001), with increasing PET dominance from cognitively normal (MCS_PET 0.412+/-0.229) through MCI (0.489+/-0.289) to AD (0.671+/-0.426), validated against amyloid SUVR (r=0.172, p=0.006) and FDG metabolic biomarkers (r=-0.287, p=0.0005) from separate imaging pipelines. External replication in 1,073 independent OASIS-3 subjects confirmed cross-cohort generalisability (H=166.99, p<0.0001, eta^2=0.156). A mechanistic comparison with SHAP demonstrated that ablation-based MCS captures clinically meaningful modality dependence that deviation-based methods cannot. These findings position MCNet as a foundation for personalised imaging decisions, clinical trial stratification, and trustworthy AI in dementia care.

When a large language model (LLM) is asked to write a person's life, how much of what it writes actually happened? We present a scene-level case-study audit - the first quantified audit of LLM-generated autobiography against a subject-specific ground-truth corpus that we are aware of, based on an unsystematic literature search. The subject and the author of this paper are the same person: a 366-day "page-a-day" book of first-person anecdotal entries was drafted with a conversational LLM whose documented inputs were a template, two exemplar days, and each day's quote - not her corpus - and every day was subsequently audited at the anecdote-scene level against an independent verification corpus using a four-level rubric fixed before analysis. We define the verification-failure rate as the share of days not rated VERIFIED (scene positively corroborated): 354 of 366 days fail, 96.7% (Wilson 95% CI 94.4-98.1%). Only 12 days contain a corroborated scene; 19 days (5.2%) assert claims actively contradicted by the record; the dominant failure mode is grounded drift - real people, employers, and settings inside invented scenes - though its measured share varies across raters. Independent re-rating replicates the headline (no evidence the original rate was inflated) while showing that the four-way taxonomy has only fair-to-moderate reliability. Regenerating the same days with current named models reproduces 100% verification failure under the same inputs; grounding generation in the subject's corpus significantly improves the verification rate while leaving substantial residual failure (83.3%). We contribute the measurement, a reusable audit instrument whose WEAK/UNVERIFIED boundary we show to be unreliable, and a grounding remedy with quantified effect.

Two analysts who calibrate the same predictive model on independent samples will deploy different prediction sets every time, because the calibration threshold inherits the randomness of the data. Wherever deployments must be audited, cached, or approved across sites, this instability is costly: no one can verify that two calibrations produced the same object. We ask two questions: when can independent calibrations yield the identical classifier, and what must that agreement cost? Perfect agreement is impossible, since a procedure that almost always returns one fixed answer cannot remain valid for every distribution, and exact agreement through shared randomness forces the procedure to ignore its data. Sharing a single random seed and rounding the calibrated threshold up to a coarse shared grid resolves the tension: the deployed classifier becomes identical across analysts with any desired probability, coverage guarantees survive, and the price is a quantified increase in set size and calibration data. Matching lower bounds show that no threshold method can pay less, and the method's one tuning constant vanishes asymptotically. Without any shared seed, a fixed grid still confines all analysts to two adjacent classifiers, and no method does better. Replicability also blocks gaming: selecting the most favorable of many recalibrations barely moves a replicable classifier, while the same selection silently undercovers standard conformal prediction. Experiments on real ImageNet outputs, a four-hospital site split, and four language-model families match the theory, including the measured sample-cost frontier.

Shape-controlled curve design plays a fundamental role in computer-aided geometric design, computer graphics, and engineering applications. In this paper, we present a novel three-parameter 9-point binary approximating subdivision scheme constructed by a weighted combination of the refinement rules of the 7-point Lagrange and 7-point B-spline subdivision schemes. The proposed construction employs displacement vectors between the corresponding refinement points of the parent schemes and constructs resultant vectors by combining neighbouring displacement vectors through three independent design-control parameters. These resultant vectors are subsequently used to derive the refinement rules, thereby yielding a unified family of binary subdivision schemes with adjustable geometric characteristics while preserving the approximating nature of the refinement process. Two representative sub-schemes are derived as special cases by imposing suitable constraints on the design-control parameters. Theoretical investigations establish the support, continuity, endpoint rules for open polygons, and Gibbs oscillation behavior of the proposed family, while graphical examples demonstrate the influence of the design-control parameters on the generated limit curves. Owing to its flexible geometric construction, intuitive parameterization, and favorable mathematical properties, the proposed subdivision framework provides a flexible and effective framework for design-controlled curve design and a valuable tool for computer-aided geometric design and related engineering applications.

This study presents an architectural and experimental cross-generation analysis of Ultralytics YOLO27 (YOLOv27), YOLO26 (YOLOv26), YOLO11 (YOLOv11), and YOLOv8 for fine-grained robotic perception in complex orchard environments. Fine-grained detection and instance segmentation of early-stage fruit anatomy remain challenging in complex orchard environments because of limited pixel footprints, green-on-green similarity, occlusion, and substantial scale variation. Because YOLO27 has been announced but its public implementation and trainable segmentation models are not yet available, the present study provides an architectural analysis of YOLO27, while controlled experiments benchmark YOLOv8, YOLO11, and YOLO26; YOLO27 experiments will be incorporated following public model availability. Five model scales-nano (n), small (s), medium (m), large (l), and extra-large (x)-were evaluated for fruitlet, calyx, and peduncle detection and segmentation using conventional 640 x 640 and small-object-focused 960 x 960 configurations, yielding 30 experiments. YOLO11s-960 achieved the highest observed mask mAP@50:95(0.402) and box mAP@50:95(0.426), whereas YOLO26s-960 achieved comparable values of 0.397 and 0.425 with only 10.37~M parameters and 34.1~GFLOPs. Peduncle remained the most challenging class, and increasing model capacity did not consistently improve accuracy. Overall, compact-to-moderate YOLO models combined with small-object-focused training provided favorable accuracy-efficiency trade-offs, establishing a reproducible benchmark for fine-grained agricultural robotic perception. Code, trained models, and experimental configurations are publicly available, and will be updated through our Github Link: https://github.com/rnjnspkt/Optimizing-and-Comparing-Ultralytics-YOLOv26-YOLOv11-and-YOLOv8-for-Small-Object-Detection-and-Seg

Large language models (LLMs) rely on tool calling as a fundamental agent capability, enabling them to invoke external systems and complete tasks beyond text generation. However, clean end-to-end (E2E) success cannot identify where a tool-use failure originates or how it propagates through a call. We introduce ToolRobustBench, a stage-wise diagnostic benchmark for tool-calling agents, where a tool-calling agent is an LLM system that selects a tool, supplies structured arguments, and interprets its returned feedback. ToolRobustBench aligns four perturbation families with the tool-use pipeline: tool-interface, user-intent, tool-output/observation, and runtime-environment perturbations. It attributes failures to tool selection, schema grounding, argument binding, tool-output/runtime-feedback handling, and E2E task success. Experiments on 15,456 single-family instances across 7 models, 16 sampled local tools, 4 perturbation families, and 14 subtypes show high but non-uniform clean performance and substantial robustness degradation, with tool-output/observation perturbation the dominant bottleneck. Mixed-family experiments reveal non-additive failure patterns that are not explained by isolated single-family results. Thus, ToolRobustBench provides a deterministic and cascade-aware benchmark for diagnosing robustness beyond clean tool-calling accuracy;

Many few-shot adaptation methods for vision-language models classify with a convex combination of the zero-shot text prototype and the mean of the K labelled image features, with a single blending ratio routinely tuned on held-out labels, often on the test set itself. We ask what the family's own bias-variance justification invites: what is the right ratio, can it be estimated without validation data, and is finding it where the performance is? First, the ratio minimising prototype mean-squared error has a closed form whose support-set plug-in is exactly a positive-part James-Stein coefficient shrinking towards the text prototype. Across 4,800 cells (ten datasets, five backbones including SigLIP, five shot counts, five seeds, four prompt tiers) this theoretically optimal ratio is a reliable estimate of the wrong quantity: on the 950 primary-tier cells where it is defined it trails a test-set-oracle ratio by 8.5 points. It saturates near 1, discarding the text prior for a nearest-class-mean classifier, because 78% of the text-image prototype distance it treats as bias is a class-independent offset that the arg max largely cancels. We prove the mechanism and bound its share of the damage at 26% by a counterfactual. Second, leave-one-out on the support set alone sets a ratio landing within 0.9 points of the oracle blend, so it is estimable without validation data. Third, validation-free linear probes beat even the oracle-tuned blend: CLAP by +1.9 points and LP++ by +1.5 on average, and at K >= 4 all four validation-free baselines sit above the oracle, the linear probes by margins excluding zero. These results locate the ceiling in the model class, not the hyperparameter: the ratio can be set near-optimally for free, and it is still not where the performance is. Code, cached features, per-cell records: https://huggingface.co/datasets/Liangzhi-Li/clipbench-blending

Process supervision has improved mathematical reasoning, where intermediate steps are naturally expressed as chains of thought. In code generation, however, process supervision remains underexplored because there is no standard notion of a step. Supervision can target lines, reasoning traces, or program states, making it unclear what to label and optimize. We propose STEP-KTODER, a framework for code preference optimization that defines steps as module-level functions in decomposed multi-function programs and assigns binary correctness labels via automatically generated unit tests. Our method provides a code-specific instantiation of stepwise KTO, combining function-level process supervision with outcome-level feedback on the full program. We evaluate on HumanEval(+), MBPP(+), BigCodeBench, and LiveCodeBench, showing that STEP-KTODER improves over outcome-only KTO and DPO. Further analysis shows that execution-based labels are essential: LLM-as-a-judge annotations systematically over-predict function failures, corrupt positive step labels, and degrade downstream preference optimization. Code is available at: https://github.com/inechnech/STEP-KTODER.

Multi-objective materials discovery with LLM agents is often limited not only by how many candidates can be proposed, but by how effectively each costly property evaluation informs the next search step. Existing agents mainly store evaluated candidates and their scores, so they know which materials succeeded but not which executable edits caused useful property changes. This makes local refinement difficult when objectives compete and an edit that improves one property may damage another. We propose TRACE, a transition-aware residual control framework that treats evaluated edits as the basic unit of feedback. TRACE records each local refinement as a parent-edit-child transition with observed property deltas, aggregates transition evidence to estimate reusable edit effects, and ranks future edits by their predicted ability to reduce the current candidate's remaining constraint violations while avoiding damage to already satisfied objectives. In a controlled same-backbone comparison, TRACE improves over LLEMA, the state-of-the-art LLM-agent baseline, raising macro-average hit rate from 18.13% to 25.96%.

Large language models (LLMs) have shown substantial promise for high-level synthesis (HLS) code generation, but most existing approaches validate only simulation or synthesis results. Because of timing and place-and-route constraints, HLS code that passes simulation and synthesis may still fail to produce deployable, runnable designs on real FPGA platforms. Moreover, the lack of public benchmarks has limited many evaluations to small, self-curated test suites. We propose FPGAgent, a multi-agent framework tailored to real FPGA environments for autonomous HLS coding with end-to-end executability validation. To the best of our knowledge, FPGAgent is the first task-specification-to-executable HLS generation framework experimentally validated on a well-established benchmark. Given a natural-language task specification, FPGAgent injects HLS-specific knowledge and employs evolutionary search to iteratively derive reliable HLS kernel implementations. It then generates a C++ validation program to verify functional correctness, diagnoses potential defects, and guides targeted repairs. Finally, it synthesizes host code for compilation and board-level execution on FPGA hardware. We comprehensively evaluate FPGAgent with five established LLMs on HLS-Eval, a benchmark containing 78 tasks across multiple domains, and verify board-level executability on a real FPGA platform. Compared with existing baselines, FPGAgent improves the synthesizable rate by 16.9% on average, executability by 26.7%, and functional correctness by 30.6%. These results show that FPGAgent substantially improves the practical usability of LLM-based HLS generation and demonstrates the value of end-to-end validation.

Creating symbolic operators by hand is one of the main bottlenecks in deploying Task and Motion Planning systems (TAMP). Recent works show that these operators can instead be learned from demonstration data. Existing methods, however, typically learn each action in isolation and cannot capture the recurring multi-step structure of manipulation tasks, so the search becomes intractable on long sequential tasks. A further inefficiency arises in the symbolic state: every provided predicate is evaluated at every search node, even when it never appears in any learned operator. We present a system that addresses both problems. Its central component is the automatic generation of macro-operators, composite actions that compress a recurring sequence of individual actions into a single planning step. Our system discovers causally linked action pairs from the training data, where one action produces exactly the condition that the next one requires, and turns each pair into a new operator. Alongside this, our system prunes every predicate that no learned operator references, which shrinks the symbolic state evaluated at each search node. Together, these changes shorten the effective planning horizon, and the benefit they bring grows with the length of the task. Across four TAMP domains, our method reaches up to a ~4.6x planning speedup compared to the baseline method, namely Learning Operators for TAMP. More importantly, it solves a long sequential task that the baseline cannot solve. Macro-operator discovery thus not only accelerates planning but, in our long-horizon domain, determines solvability in practice. Our code is available at https://github.com/canemirbora4/macro-tamp

A tool call can be perfectly valid yet still leave an autonomous agent unable to determine what to do next. For example, if an external effect commits but its response is lost, committed and uncommitted states may become indistinguishable to the agent even though they require different continuation actions. We study this gap between callability and operability: whether a tool interface exposes the action-relevant state and semantics needed for an agent to continue safely under operational uncertainty. We operationalize tool operability through Agent-First Tooling (AFT), a set of interface mechanisms spanning selective capability discovery, execution lifecycle and recovery, explicit external-effect semantics, machine-readable results, and postcondition verification. We introduce AFT-Bench, a controlled interface-intervention framework that holds the task, backend, initial state, injected failure, agent, and language model fixed while varying the interface exposed to the agent.

Emergency departments (EDs) operate under time pressure, generating multimodal data such as clinical conversations, triage notes, and discharge documents. Recent advances in natural language processing (NLP), particularly pretrained transformers and large language models, have created new opportunities to support language and time-intensive stages of emergency care. Yet existing surveys map clinical NLP across the broader hospital workflow or focus on specific tasks. This survey analyses 46 papers spanning the three phases of ED: triage, diagnosis, and disposition, covering tasks such as triage classification, clinical summarisation, automatic diagnosis, report generation, and discharge documentation. We examine modelling paradigms, evaluation practices, and emerging benchmarks and shared tasks. Across tasks, we identify common trends, including a shift from task-specific neural architectures to pretrained language models, growing interest in interactive clinical systems, and increasing attention to clinically grounded evaluation. Finally, we detail open challenges such as limited generalisability, noisy clinical inputs, and workflow constraints that inform future ED-NLP research.

Foundation models for astronomy are trained on survey pixels together with the catalogue products derived from those pixels. Those catalogues are incomplete at a measurable rate, and a model trained on both inherits that incompleteness as a systematic. We audit AION-1, a 39-modality transformer trained on more than 200 million objects, using causal interventions on its inputs. Holding the image tokens byte-identical and editing only the survey segmentation map changes every quantity the model reports – flux, size, ellipticity, redshift – by 110-4400 times a matched placebo. The mechanism is detection gating, presence at the field centre (r = 0.47), not the light the mask encloses (r = 0.30); across 322 real blends the model ignores how the pipeline partitioned the light (R = -0.006). Nor is the preference specific to that channel: contradicted catalogue photometry leaves the model nine times worse than supplying no metadata at all. The Legacy Survey pipeline leaves 3.68% of targets with no segment covering their position. Propagating that rate, with a miss represented by the fields the pipeline actually returns, shifts tomographic mean redshifts by a median 0.71 times the LSST DESC requirement over 40 assignments and exceeds it in 12; observed positional errors take the worst bin to 8.3 times. Drawing the misses by their measured magnitude dependence rather than uniformly does not change it. Spectroscopy removes the effect, withholding the detection channel removes it at no measurable cost, and the effect grows with model scale. Two further limits lie in the tokeniser: its image codec resolves 28 effective states on source patches against 934 for the spectrum codec, and the redshift readout is quantisation-limited. Sparse dictionaries are unreliable causal handles: across 15, recovery spans 26-75% and moves up to 18 points on the seed alone.