Richard Green

Amelia Glaese, Nat McAleese, Maja Trębacz et al.

We present Sparrow, an information-seeking dialogue agent trained to be more helpful, correct, and harmless compared to prompted language model baselines. We use reinforcement learning from human feedback to train our models with two new additions to help human raters judge agent behaviour. First, to make our agent more helpful and harmless, we break down the requirements for good dialogue into natural language rules the agent should follow, and ask raters about each rule separately. We demonstrate that this breakdown enables us to collect more targeted human judgements of agent behaviour and allows for more efficient rule-conditional reward models. Second, our agent provides evidence from sources supporting factual claims when collecting preference judgements over model statements. For factual questions, evidence provided by Sparrow supports the sampled response 78% of the time. Sparrow is preferred more often than baselines while being more resilient to adversarial probing by humans, violating our rules only 8% of the time when probed. Finally, we conduct extensive analyses showing that though our model learns to follow our rules it can exhibit distributional biases.

Yida Lin, Bing Xue, Mengjie Zhang et al.

Manual pruning of radiata pine trees presents significant safety risks due to their substantial height and the challenging terrains in which they thrive. To address these risks, this research proposes the development of a drone-based pruning system equipped with specialized pruning tools and a stereo vision camera, enabling precise detection and trimming of branches. Deep learning algorithms, including YOLO and Mask R-CNN, are employed to ensure accurate branch detection, while the Semi-Global Matching algorithm is integrated to provide reliable distance estimation. The synergy between these techniques facilitates the precise identification of branch locations and enables efficient, targeted pruning. Experimental results demonstrate that the combined implementation of YOLO and SGBM enables the drone to accurately detect branches and measure their distances from the drone. This research not only improves the safety and efficiency of pruning operations but also makes a significant contribution to the advancement of drone technology in the automation of agricultural and forestry practices, laying a foundational framework for further innovations in environmental management.

Yida Lin, Bing Xue, Mengjie Zhang et al.

This paper presents a stereo-vision-based system mounted on a drone for detecting and localising radiata pine branches to support autonomous pruning. The proposed pipeline comprises two stages: branch segmentation and depth estimation. For segmentation, YOLOv8, YOLOv9, and Mask R-CNN variants are compared on a custom dataset of 71 stereo image pairs captured with a ZED Mini camera. For depth estimation, both a traditional method (SGBM with WLS filtering) and deep-learning-based methods (PSMNet, ACVNet, GWCNet, MobileStereoNet, RAFT-Stereo, and NeRF-Supervised Deep Stereo) are evaluated. A centroid-based triangulation algorithm with MAD outlier rejection is proposed to compute branch distance from the segmentation mask and disparity map. Qualitative evaluation at distances of 1-2 m indicates that the deep learning-based disparity maps produce more coherent depth estimates than SGBM, demonstrating the feasibility of low-cost stereo vision for automated branch positioning in forestry.

Yida Lin, Bing Xue, Mengjie Zhang et al.

Autonomous tree pruning with unmanned aerial vehicles (UAVs) is a safety-critical real-world task: the onboard perception system must estimate the metric distance from a cutting tool to thin tree branches in real time so that the UAV can approach, align, and actuate the pruner without collision. We address this problem by training five variants of DEFOM-Stereo - a recent foundation-model-based stereo matcher - on a task-specific synthetic dataset and deploying the checkpoints on an NVIDIA Jetson Orin Super 16 GB. The training corpus is built in Unreal Engine 5 with a simulated ZED Mini stereo camera capturing 5,520 stereo pairs across 115 tree instances from three viewpoints at 2m distance; dense EXR depth maps provide exact, spatially complete supervision for thin branches. On the synthetic test set, DEFOM-Stereo ViT-S achieves the best depth-domain accuracy (EPE 1.74 px, D1-all 5.81%, delta-1 95.90%, depth MAE 23.40 cm) but its Jetson inference speed of ~2.2 FPS (~450 ms per frame) remains too slow for responsive closed-loop tool control. A newly introduced balanced variant, DEFOM-PrunePlus (~21M backbone, ~3.3 FPS on Jetson), offers the best deployable accuracy-speed trade-off (EPE 5.87 px, depth MAE 64.26 cm, delta-1 87.59%): its frame rate is sufficient for real-time guidance and its depth accuracy supports safe branch approach planning at the 2m operating range. The lightweight DEFOM-PruneStereo (~6.9 FPS) and DEFOM-PruneNano (~8.5 FPS) run fast but sacrifice substantial accuracy (depth MAE > 57 cm), making estimates too unreliable for safe actuation. Zero-shot inference on real photographs confirms that full-capacity models preserve branch geometry, validating the sim-to-real transfer. We conclude that DEFOM-PrunePlus provides the most practical accuracy-latency balance for onboard distance estimation, while ViT-S serves as the reference for future hardware.

Yida Lin, Bing Xue, Mengjie Zhang et al.

Dense ground-truth disparity maps are practically unobtainable in forestry environments, where thin overlapping branches and complex canopy geometry defeat conventional depth sensors -- a critical bottleneck for training supervised stereo matching networks for autonomous UAV-based pruning. We present UE5-Forest, a photorealistic synthetic stereo dataset built entirely in Unreal Engine 5 (UE5). One hundred and fifteen photogrammetry-scanned trees from the Quixel Megascans library are placed in virtual scenes and captured by a simulated stereo rig whose intrinsics -- 63 mm baseline, 2.8 mm focal length, 3.84 mm sensor width -- replicate the ZED Mini camera mounted on our drone. Orbiting each tree at up to 2 m across three elevation bands (horizontal, +45 degrees, -45 degrees) yields 5,520 rectified 1920 x 1080 stereo pairs with pixel-perfect disparity labels. We provide a statistical characterisation of the dataset -- covering disparity distributions, scene diversity, and visual fidelity -- and a qualitative comparison with real-world Canterbury Tree Branches imagery that confirms the photorealistic quality and geometric plausibility of the rendered data. The dataset will be publicly released to provide the community with a ready-to-use benchmark and training resource for stereo-based forestry depth estimation.

Yida Lin, Bing Xue, Mengjie Zhang et al.

Accurate per-branch 3D reconstruction is a prerequisite for autonomous UAV-based tree pruning; however, dense disparity maps from modern stereo matchers often remain too noisy for individual branch analysis in complex forest canopies. This paper introduces a progressive pipeline integrating DEFOM-Stereo foundation-model disparity estimation, SAM3 instance segmentation, and multi-stage depth optimization to deliver robust per-branch point clouds. Starting from a naive baseline, we systematically identify and resolve three error families through successive refinements. Mask boundary contamination is first addressed through morphological erosion and subsequently refined via a skeleton-preserving variant to safeguard thin-branch topology. Segmentation inaccuracy is then mitigated using LAB-space Mahalanobis color validation coupled with cross-branch overlap arbitration. Finally, depth noise - the most persistent error source - is initially reduced by outlier removal and median filtering, before being superseded by a robust five-stage scheme comprising MAD global detection, spatial density consensus, local MAD filtering, RGB-guided filtering, and adaptive bilateral filtering. Evaluated on 1920x1080 stereo imagery of Radiata pine (Pinus radiata) acquired with a ZED Mini camera (63 mm baseline) from a UAV in Canterbury, New Zealand, the proposed pipeline reduces the average per-branch depth standard deviation by 82% while retaining edge fidelity. The result is geometrically coherent 3D point clouds suitable for autonomous pruning tool positioning. All code and processed data are publicly released to facilitate further UAV forestry research.

Yida Lin, Bing Xue, Mengjie Zhang et al.

Autonomous UAV forestry operations require robust depth estimation methods with strong cross-domain generalization. However, existing evaluations focus on urban and indoor scenarios, leaving a critical gap for specialized vegetation-dense environments. We present the first systematic zero-shot evaluation of eight state-of-the-art stereo methods--RAFT-Stereo, IGEV, IGEV++, BridgeDepth, StereoAnywhere, DEFOM (plus baseline methods ACVNet, PSMNet, TCstereo)--spanning iterative refinement, foundation model, and zero-shot adaptation paradigms. All methods are trained exclusively on Scene Flow and evaluated without fine-tuning on four standard benchmarks (ETH3D, KITTI 2012/2015, Middlebury) plus a novel 5,313-pair Canterbury forestry dataset captured with ZED Mini camera (1920x1080). Performance reveals scene-dependent patterns: foundation models excel on structured scenes (BridgeDepth: 0.23 px on ETH3D, 0.83-1.07 px on KITTI; DEFOM: 0.35-4.65 px across benchmarks), while iterative methods maintain cross-domain robustness (IGEV++: 0.36-6.77 px; IGEV: 0.33-21.91 px). Critical finding: RAFT-Stereo exhibits catastrophic ETH3D failure (26.23 px EPE, 98 percent error rate) due to negative disparity predictions, while performing normally on KITTI (0.90-1.11 px). Qualitative evaluation on Canterbury forestry dataset identifies DEFOM as the optimal gold-standard baseline for vegetation depth estimation, exhibiting superior depth smoothness, occlusion handling, and cross-domain consistency compared to IGEV++, despite IGEV++'s finer detail preservation.

Yida Lin, Bing Xue, Mengjie Zhang et al.

Autonomous drone-based tree pruning needs accurate, real-time depth estimation from stereo cameras. Depth is computed from disparity maps using $Z = f B/d$, so even small disparity errors cause noticeable depth mistakes at working distances. Building on our earlier work that identified DEFOM-Stereo as the best reference disparity generator for vegetation scenes, we present the first study to train and test ten deep stereo matching networks on real tree branch images. We use the Canterbury Tree Branches dataset -- 5,313 stereo pairs from a ZED Mini camera at 1080P and 720P -- with DEFOM-generated disparity maps as training targets. The ten methods cover step-by-step refinement, 3D convolution, edge-aware attention, and lightweight designs. Using perceptual metrics (SSIM, LPIPS, ViTScore) and structural metrics (SIFT/ORB feature matching), we find that BANet-3D produces the best overall quality (SSIM = 0.883, LPIPS = 0.157), while RAFT-Stereo scores highest on scene-level understanding (ViTScore = 0.799). Testing on an NVIDIA Jetson Orin Super (16 GB, independently powered) mounted on our drone shows that AnyNet reaches 6.99 FPS at 1080P -- the only near-real-time option -- while BANet-2D gives the best quality-speed balance at 1.21 FPS. We also compare 720P and 1080P processing times to guide resolution choices for forestry drone systems.

Yida Lin, Bing Xue, Mengjie Zhang et al.

Autonomous UAV forestry operations require robust depth estimation with strong cross-domain generalization, yet existing evaluations focus on urban and indoor scenarios, leaving a critical gap for vegetation-dense environments. We present the first systematic zero-shot evaluation of eight stereo methods spanning iterative refinement, foundation model, diffusion-based, and 3D CNN paradigms. All methods use officially released pretrained weights (trained on Scene Flow) and are evaluated on four standard benchmarks (ETH3D, KITTI 2012/2015, Middlebury) plus a novel 5,313-pair Canterbury Tree Branches dataset ($1920 \times 1080$). Results reveal scene-dependent patterns: foundation models excel on structured scenes (BridgeDepth: 0.23 px on ETH3D; DEFOM: 4.65 px on Middlebury), while iterative methods show variable cross-benchmark performance (IGEV++: 0.36 px on ETH3D but 6.77 px on Middlebury; IGEV: 0.33 px on ETH3D but 4.99 px on Middlebury). Qualitative evaluation on the Tree Branches dataset establishes DEFOM as the gold-standard baseline for vegetation depth estimation, with superior cross-domain consistency (consistently ranking 1st-2nd across benchmarks, average rank 1.75). DEFOM predictions will serve as pseudo-ground-truth for future benchmarking.

Ans Qureshi, Neville Loh, Young Min Kwon et al.

Following a global trend, the lack of reliable access to skilled labour is causing critical issues for the effective management of apple orchards. One of the primary challenges is maintaining skilled human operators capable of making precise fruitlet thinning decisions. Thinning requires accurately measuring the true crop load for individual apple trees to provide optimal thinning decisions on an individual basis. A challenging task due to the dense foliage obscuring the fruitlets within the tree structure. This paper presents the initial design, implementation, and evaluation details of the vision system for an automatic apple fruitlet thinning robot to meet this need. The platform consists of a UR5 robotic arm and stereo cameras which enable it to look around the leaves to map the precise number and size of the fruitlets on the apple branches. We show that this platform can measure the fruitlet load on the apple tree to with 84% accuracy in a real-world commercial apple orchard while being 87% precise.

Harry Dobbs, Oliver Batchelor, Richard Green et al.

This paper introduces Smart-Tree, a supervised method for approximating the medial axes of branch skeletons from a tree point cloud. Smart-Tree uses a sparse voxel convolutional neural network to extract the radius and direction towards the medial axis of each input point. A greedy algorithm performs robust skeletonization using the estimated medial axis. Our proposed method provides robustness to complex tree structures and improves fidelity when dealing with self-occlusions, complex geometry, touching branches, and varying point densities. We evaluate Smart-Tree using a multi-species synthetic tree dataset and perform qualitative analysis on a real-world tree point cloud. Our experimentation with synthetic and real-world datasets demonstrates the robustness of our approach over the current state-of-the-art method. The dataset and source code are publicly available.

Casey Peat, Oliver Batchelor, Richard Green et al.

We present Zero-NeRF, a projective surface registration method that, to the best of our knowledge, offers the first general solution capable of alignment between scene representations with minimal or zero visual correspondence. To do this, we enforce consistency between visible surfaces of partial and complete reconstructions, which allows us to constrain occluded geometry. We use a NeRF as our surface representation and the NeRF rendering pipeline to perform this alignment. To demonstrate the efficacy of our method, we register real-world scenes from opposite sides with infinitesimal overlaps that cannot be accurately registered using prior methods, and we compare these results against widely used registration methods.

Gheorghe Comanici, Eric Bieber, Mike Schaekermann et al. · amazon-science, baidu

In this report, we introduce the Gemini 2.X model family: Gemini 2.5 Pro and Gemini 2.5 Flash, as well as our earlier Gemini 2.0 Flash and Flash-Lite models. Gemini 2.5 Pro is our most capable model yet, achieving SoTA performance on frontier coding and reasoning benchmarks. In addition to its incredible coding and reasoning skills, Gemini 2.5 Pro is a thinking model that excels at multimodal understanding and it is now able to process up to 3 hours of video content. Its unique combination of long context, multimodal and reasoning capabilities can be combined to unlock new agentic workflows. Gemini 2.5 Flash provides excellent reasoning abilities at a fraction of the compute and latency requirements and Gemini 2.0 Flash and Flash-Lite provide high performance at low latency and cost. Taken together, the Gemini 2.X model generation spans the full Pareto frontier of model capability vs cost, allowing users to explore the boundaries of what is possible with complex agentic problem solving.

Adam Suski, Elina Spyrou, Jacob Mays et al.

Long-duration energy storage (LDES) faces significant revenue volatility that impedes investment. This paper evaluates four contract-based support mechanisms using an equilibrium model with risk-averse investors and incomplete risk markets. Applied to a stylized 2035 Great Britain case, we find that all mechanisms can achieve the targeted LDES capacity but differ substantially in cost-effectiveness and risk-aversion sensitivity. Contracts that eliminate revenue volatility achieve the lowest costs but may weaken operational incentives, while contracts that preserve market exposure maintain incentives at higher costs.

Yida Lin, Bing Xue, Mengjie Zhang et al.

UAV-based autonomous forestry operations require rapid and precise tree branch segmentation for safe navigation and automated pruning across varying pixel resolutions and operational conditions. We evaluate different deep learning methods at three resolutions (256x256, 512x512, 1024x1024) using the Urban Street Tree Dataset, employing standard metrics (IoU, Dice) and specialized measures including Thin Structure IoU (TS-IoU) and Connectivity Preservation Rate (CPR). Among 22 configurations tested, U-Net with MiT-B4 backbone achieves strong performance at 256x256. At 512x512, MiT-B4 leads in IoU, Dice, TS-IoU, and Boundary-F1. At 1024x1024, U-Net+MiT-B3 shows the best validation performance for IoU/Dice and precision, while U-Net++ excels in boundary quality. PSPNet provides the most efficient option (2.36/9.43/37.74 GFLOPs) with 25.7/19.6/11.8 percentage point IoU reductions compared to top performers at respective resolutions. These results establish multi-resolution benchmarks for accuracy-efficiency trade-offs in embedded forestry systems. Implementation is available at https://github.com/BennyLinntu/PerformanceTreeBranchSegmentation.

Meet Shah, Jason Gusdorf, Anil Palepu et al.

The practice of medicine relies not only upon skillful dialogue but also on the nuanced exchange and interpretation of rich auditory and visual cues between doctors and patients. Building on the low-latency voice and video processing capabilities of Gemini, we introduce AI co-clinician, a first-of-its-kind conversational AI system utilizing continuous streams of audio-visual data from live patient conversations to inform real-time clinical decisions. Its dual-agent architecture balances deep clinical reasoning with the low latency required for natural dialogue. To assess this system, we implemented a video-based interface emulating telemedicine consultations. We crafted 20 standardized outpatient scenarios requiring proactive real-time auditory and visual reasoning and designed "TelePACES" evaluation criteria alongside case-specific rubrics. In a randomized, interface-blinded, crossover simulation study (n = 120 encounters) with 10 internal medicine residents as patient actors, we compared AI co-clinician with primary care physicians (PCPs), GPT-Realtime, and a baseline agent. AI co-clinician approached PCPs in key TelePACES dimensions, including management plans and differential diagnosis, while significantly outperforming GPT-Realtime across all general criteria. While our agent demonstrated parity with PCPs in case-specific triage measures, physicians maintained superior overall performance in case-specific assessments. Although AI co-clinician marks a significant advance in real-time telemedical AI, gaps remain in physical examination and disease-specific reasoning. Our work shows that text-only approaches fail to capture the true challenges of medical consultation and suggests that high-stakes real-time diagnostic AI is most safely advanced in collaborative, triadic models where AI can be a supportive co-clinician for doctors and patients.

Yida Lin, Bing Xue, Mengjie Zhang et al.

Manual pruning of radiata pine, a species of major economic importance to New Zealand forestry, is hazardous, labour-intensive, and increasingly constrained by workforce shortages. Existing autonomous pruning platforms typically rely on expensive sensors such as LiDAR and are limited to thick branches, which restricts their wider adoption. This paper investigates whether a single low-cost stereo camera mounted on a drone can provide sufficiently accurate branch detection and three-dimensional positioning to support autonomous pruning of branches as thin as 10 mm, thereby removing the need for auxiliary depth sensors. The proposed pipeline comprises two stages: branch segmentation and depth estimation. For segmentation, Mask R-CNN variants and the YOLOv8 and YOLOv9 families are compared on a custom dataset of 71 stereo image pairs captured with a ZED Mini camera; YOLOv8 and YOLOv9 are selected as representative state-of-the-art real-time segmentors at the time of data collection, and the framework is designed to remain compatible with newer YOLO releases. For depth estimation, a traditional method (SGBM with WLS filtering) and deep-learning-based methods (PSMNet, ACVNet, GWCNet, MobileStereoNet, RAFT-Stereo, and NeRF-Supervised Deep Stereo) are evaluated, including cross-dataset fine-tuning experiments that expose the domain gap between urban driving benchmarks and natural forestry scenes. The main novelty of this work lies in coupling stereo segmentation with a centroid-based triangulation algorithm and Median-Absolute-Deviation outlier rejection that converts a segmentation mask and disparity map into a single robust branch-to-camera distance, addressing the challenges of sparse texture, thin structures, and noisy disparity values typical of forest scenes. Qualitative evaluations at distances of 1-2 m show that the learning-based stereo methods produce more coherent depth es...

Reilly Haskins, Richard Green

This paper proposes a novel method to diagnose sensory ataxia via an automated Romberg Test - the current de facto medical procedure used to diagnose this condition. It utilizes a convolutional neural network to predict joint locations, used for the calculation of various bio-mechanical markers such as the center of mass of the subject and various joint angles. This information is used in combination with data filtering techniques such as Kalman Filters, and center of mass analysis which helped make accurate inferences about the relative weight distribution in the lateral and anterior-posterior axes, and provide an objective, mathematically based diagnosis of this condition. In order to evaluate the performance of this method, testing was performed using dual weight scales and pre-annotated diagnosis videos taken from medical settings. These two methods both quantified the veritable weight distribution upon the ground surface with a ground truth and provided a real-world estimate of accuracy for the proposed method. A mean absolute error of 0.2912 percent was found for the calculated relative weight distribution difference, and an accuracy of 83.33 percent was achieved on diagnoses.

Yida Lin, Bing Xue, Mengjie Zhang et al.

Manual pruning of radiata pine trees poses significant safety risks due to extreme working heights and challenging terrain. This paper presents a computer vision framework that integrates YOLO object detection with Semi-Global Block Matching (SGBM) stereo vision for autonomous drone-based pruning operations. Our system achieves precise branch detection and depth estimation using only stereo camera input, eliminating the need for expensive LiDAR sensors. Experimental evaluation demonstrates YOLO's superior performance over Mask R-CNN, achieving 82.0% mAPmask50-95 for branch segmentation. The integrated system accurately localizes branches within a 2 m operational range, with processing times under one second per frame. These results establish the feasibility of cost-effective autonomous pruning systems that enhance worker safety and operational efficiency in commercial forestry.

Yida Lin, Bing Xue, Mengjie Zhang et al.

Traditional stereo matching algorithms like Semi-Global Block Matching (SGBM) with Weighted Least Squares (WLS) filtering offer speed advantages over neural networks for UAV applications, generating disparity maps in approximately 0.5 seconds per frame. However, these algorithms require meticulous parameter tuning. We propose a Genetic Algorithm (GA) based parameter optimization framework that systematically searches for optimal parameter configurations for SGBM and WLS, enabling UAVs to measure distances to tree branches with enhanced precision while maintaining processing efficiency. Our contributions include: (1) a novel GA-based parameter optimization framework that eliminates manual tuning; (2) a comprehensive evaluation methodology using multiple image quality metrics; and (3) a practical solution for resource-constrained UAV systems. Experimental results demonstrate that our GA-optimized approach reduces Mean Squared Error by 42.86% while increasing Peak Signal-to-Noise Ratio and Structural Similarity by 8.47% and 28.52%, respectively, compared with baseline configurations. Furthermore, our approach demonstrates superior generalization performance across varied imaging conditions, which is critcal for real-world forestry applications.

Scott Paulin, Tom Botterill, XiaoQi Chen et al.

Many robot arms can accomplish one task using many different joint configurations. Often only one of these configurations is used as a goal by the path planner. Ideally the robot's path planner would be able to use the extra configurations to find higher quality paths. In this paper we use the extra goal configurations to find significantly shorter paths that are faster to execute compared to a planner that chooses one goal configuration arbitrarily. In a grape vine pruning robot arm experiment our proposed approach reduced execution times by 58%.

Scott Paulin, Tom Botterill, XiaoQi Chen et al.

Many robots operating in unpredictable environments require an online path planning algorithm that can quickly compute high quality paths. Asymptotically optimal planners are capable of finding the optimal path, but can be slow to converge. Local optimisation algorithms are capable of quickly improving a solution, but are not guaranteed to converge to the optimal solution. In this paper we develop a new way to integrate an asymptotically optimal planners with a local optimiser. We test our approach using RRTConnect* with a short-cutting local optimiser. Our approach results in a significant performance improvement when compared with the state-of-the-art RRTConnect* asymptotically optimal planner and computes paths that are 31\% faster to execute when both are given 3 seconds of planning time.