This paper proposes a novel inverse reinforcement learning framework using a diffusion-based adaptive lookahead planner (IRL-DAL) for autonomous vehicles. Training begins with imitation from an expert finite state machine (FSM) controller to provide a stable initialization. Environment terms are combined with an IRL discriminator signal to align with expert goals. Reinforcement learning (RL) is then performed with a hybrid reward that combines diffuse environmental feedback and targeted IRL rewards. A conditional diffusion model, which acts as a safety supervisor, plans safe paths. It stays in its lane, avoids obstacles, and moves smoothly. Then, a learnable adaptive mask (LAM) improves perception. It shifts visual attention based on vehicle speed and nearby hazards. After FSM-based imitation, the policy is fine-tuned with Proximal Policy Optimization (PPO). Training is run in the Webots simulator with a two-stage curriculum. A 96\% success rate is reached, and collisions are reduced to 0.05 per 1k steps, marking a new benchmark for safe navigation. By applying the proposed approach, the agent not only drives in lane but also handles unsafe conditions at an expert level, increasing robustness.We make our code publicly available.
Despite rapid advances in autonomous AI scientists powered by language models, generating publication-ready illustrations remains a labor-intensive bottleneck in the research workflow. To lift this burden, we introduce PaperBanana, an agentic framework for automated generation of publication-ready academic illustrations. Powered by state-of-the-art VLMs and image generation models, PaperBanana orchestrates specialized agents to retrieve references, plan content and style, render images, and iteratively refine via self-critique. To rigorously evaluate our framework, we introduce PaperBananaBench, comprising 292 test cases for methodology diagrams curated from NeurIPS 2025 publications, covering diverse research domains and illustration styles. Comprehensive experiments demonstrate that PaperBanana consistently outperforms leading baselines in faithfulness, conciseness, readability, and aesthetics. We further show that our method effectively extends to the generation of high-quality statistical plots. Collectively, PaperBanana paves the way for the automated generation of publication-ready illustrations.
Gravitational wave (GW) astronomy opens a new venue to explore the universe. Future observatories such as LISA, the Laser Interferometer Space Antenna, are expected to observe previously undetectable fundamental physics effects in signals predicted by General Relativity (GR).One particularly interesting such signal is associated to the displacement memory effect, which corresponds to a permanent deformation of spacetime due to the passage of gravitational radiation. In this work, we explore the ability of LISA to observe and characterize this effect. In order to do this, we use state-of-the-art simulations of the LISA instrument, and we perform a Bayesian analysis to assess the detectability and establish general conditions to claim detection of the displacement memory effect from individual massive black hole binary (MBHB) merger events in LISA. We perform parameter estimation both to explore the impact of the displacement memory effect and to reconstruct its amplitude. We discuss the precision at which such a reconstruction can be obtained thus opening the way to tests of GR and alternative theories. To provide astrophysical context, we apply our analysis to black hole binary populations models and estimate the rates at which the displacement memory effect could be observed within the LISA planned lifetime.
Home-based care (HBC) delivers medical and care services in patients' living environments, offering unique opportunities for patient-centered care. However, patient agency is often inadequately represented in shared HBC planning processes. Through 23 multi-stakeholder interviews with HBC patients, healthcare professionals, and care workers, alongside 60 hours of ethnographic observations, we examined how patient agency manifests in HBC and why this representation gap occurs. Our findings reveal that patient agency is not a static individual attribute but a relational capacity shaped through maintaining everyday continuity, mutual recognition from care providers, and engagement with material home environments. Furthermore, we identified that structured documentation systems filter out contextual knowledge, informal communication channels fragment patient voices, and doctor-centered hierarchies position patients as passive recipients. Drawing on these insights, we propose design considerations to bridge this representation gap and to integrate patient agency into shared HBC plans.
Sidewalk sheds are a common feature of the streetscape in New York City, reflecting ongoing construction and maintenance activities. However, policymakers and local business owners have raised concerns about reduced storefront visibility and altered pedestrian navigation. Although sidewalk sheds are widely used for safety, their effects on pedestrian visibility and movement are not directly measured in current planning practices. To address this, we developed an AI-based chatbot survey that collects image-based annotations and route choices from pedestrians, linking these responses to specific shed design features, including clearance height, post spacing, and color. This AI chatbot survey integrates a large language model (e.g., Google's Gemini-1.5-flash-001 model) with an image-annotation interface, allowing users to interact with street images, mark visual elements, and provide structured feedback through guided dialogue. To explore pedestrian perceptions and behaviors, this paper conducts a grid-based analysis of entrance annotations and applies logistic mixed-effects modeling to assess sidewalk choice patterns. Analysis of the dataset (n = 25) shows that: (1) the presence of scaffolding significantly reduces pedestrians' ability to identify ground-floor retail entrances, and (2) variations in weather conditions and shed design features significantly influence sidewalk selection behavior. By integrating generative AI into urban research, this study demonstrates a novel method for evaluating sidewalk shed designs and provides empirical evidence to support adjustments to shed guidelines that improve the pedestrian experience without compromising safety.
Diagnosing the failure mechanisms of Deep Research Agents (DRAs) remains a critical challenge. Existing benchmarks predominantly rely on end-to-end evaluation, obscuring critical intermediate hallucinations, such as flawed planning, that accumulate throughout the research trajectory. To bridge this gap, we propose a shift from outcome-based to process-aware evaluation by auditing the full research trajectory. We introduce the PIES Taxonomy to categorize hallucinations along functional components (Planning vs. Summarization) and error properties (Explicit vs. Implicit). We instantiate this taxonomy into a fine-grained evaluation framework that decomposes the trajectory to rigorously quantify these hallucinations. Leveraging this framework to isolate 100 distinctively hallucination-prone tasks including adversarial scenarios, we curate DeepHalluBench. Experiments on six state-of-theart DRAs reveal that no system achieves robust reliability. Furthermore, our diagnostic analysis traces the etiology of these failures to systemic deficits, specifically hallucination propagation and cognitive biases, providing foundational insights to guide future architectural optimization. Data and code are available at https://github.com/yuhao-zhan/DeepHalluBench.
Diffusion policies (DP) have demonstrated significant potential in visual navigation by capturing diverse multi-modal trajectory distributions. However, standard imitation learning (IL), which most DP methods rely on for training, often inherits sub-optimality and redundancy from expert demonstrations, thereby necessitating a computationally intensive "generate-then-filter" pipeline that relies on auxiliary selectors during inference. To address these challenges, we propose Self-Imitated Diffusion Policy (SIDP), a novel framework that learns improved planning by selectively imitating a set of trajectories sampled from itself. Specifically, SIDP introduces a reward-guided self-imitation mechanism that encourages the policy to consistently produce high-quality trajectories efficiently, rather than outputs of inconsistent quality, thereby reducing reliance on extensive sampling and post-filtering. During training, we employ a reward-driven curriculum learning paradigm to mitigate inefficient data utility, and goal-agnostic exploration for trajectory augmentation to improve planning robustness. Extensive evaluations on a comprehensive simulation benchmark show that SIDP significantly outperforms previous methods, with real-world experiments confirming its effectiveness across multiple robotic platforms. On Jetson Orin Nano, SIDP delivers a 2.5$\times$ faster inference than the baseline NavDP, i.e., 110ms VS 273ms, enabling efficient real-time deployment.
Trajectory planning is a core task in autonomous driving, requiring the prediction of safe and comfortable paths across diverse scenarios. Integrating Multi-modal Large Language Models (MLLMs) with Reinforcement Learning (RL) has shown promise in addressing "long-tail" scenarios. However, existing methods are constrained to single-turn reasoning, limiting their ability to handle complex tasks requiring iterative refinement. To overcome this limitation, we present MTDrive, a multi-turn framework that enables MLLMs to iteratively refine trajectories based on environmental feedback. MTDrive introduces Multi-Turn Group Relative Policy Optimization (mtGRPO), which mitigates reward sparsity by computing relative advantages across turns. We further construct an interactive trajectory understanding dataset from closed-loop simulation to support multi-turn training. Experiments on the NAVSIM benchmark demonstrate superior performance compared to existing methods, validating the effectiveness of our multi-turn reasoning paradigm. Additionally, we implement system-level optimizations to reduce data transfer overhead caused by high-resolution images and multi-turn sequences, achieving 2.5x training throughput. Our data, models, and code will be made available soon.
Efficient planning of assembly motions is a long standing challenge in the field of robotics that has been primarily tackled with reinforcement learning and sampling-based methods by using extensive physics simulations. This paper proposes a sample-efficient robust optimal control approach for the determination of assembly motions, which requires significantly less physics simulation steps during planning through the efficient use of derivative information. To this end, a differentiable physics simulation is constructed that provides second-order analytic derivatives to the numerical solver and allows one to traverse seamlessly from informative derivatives to accurate contact simulation. The solution of the physics simulation problem is made differentiable by using smoothing inspired by interior-point methods applied to both the collision detection as well as the contact resolution problem. We propose a modified variant of an optimization-based formulation of collision detection formulated as a linear program and present an efficient implementation for the nominal evaluation and corresponding first- and second-order derivatives. Moreover, a multi-scenario-based trajectory optimization problem that ensures robustness with respect to sim-to-real mismatches is derived. The capability of the considered formulation is illustrated by results where over 99\% successful executions are achieved in real-world experiments. Thereby, we carefully investigate the effect of smooth approximations of the contact dynamics and robust modeling on the success rates. Furthermore, the method's capability is tested on different peg-in-hole problems in simulation to show the benefit of using exact Hessians over commonly used Hessian approximations.
In observational studies of treatment effects, estimates may be biased by unmeasured confounders, which can potentially affect the validity of the results. Understanding sensitivity to such biases helps assess how unmeasured confounding impacts credibility. The design of an observational study strongly influences its sensitivity to bias. Previous work has shown that the sensitivity to bias can be reduced by dividing a dataset into a planning sample and a larger analysis sample, where the planning sample guides design decisions. But the choice of what fraction of the data to put in the planning sample vs. the analysis sample was ad hoc. Here, we develop an approach to find the optimal fraction using plasmode datasets. We show that our method works well in high-dimensional outcome spaces. We apply our method to study the effects of exposure to second-hand smoke in children. The OptimalSampling R package implementing our method is available at GitHub.
Urban region profiling, the task of characterizing geographical areas, is crucial for urban planning and resource allocation. However, existing research in this domain faces two significant limitations. First, most methods are confined to single-task prediction, failing to capture the interconnected, multi-faceted nature of urban environments where numerous indicators are deeply correlated. Second, the field lacks a standardized experimental benchmark, which severely impedes fair comparison and reproducible progress. To address these challenges, we first establish a comprehensive benchmark for multi-task urban region profiling, featuring multi-modal features and a diverse set of strong baselines to ensure a fair and rigorous evaluation environment. Concurrently, we propose UrbanMoE, the first sparse multi-modal, multi-expert framework specifically architected to solve the multi-task challenge. Leveraging a sparse Mixture-of-Experts architecture, it dynamically routes multi-modal features to specialized sub-networks, enabling the simultaneous prediction of diverse urban indicators. We conduct extensive experiments on three real-world datasets within our benchmark, where UrbanMoE consistently demonstrates superior performance over all baselines. Further in-depth analysis validates the efficacy and efficiency of our approach, setting a new state-of-the-art and providing the community with a valuable tool for future research in urban analytics
This work proposes Bonnet, an ultra-fast sparse-volume pipeline for whole-body bone segmentation from CT scans. Accurate bone segmentation is important for surgical planning and anatomical analysis, but existing 3D voxel-based models such as nnU-Net and STU-Net require heavy computation and often take several minutes per scan, which limits time-critical use. The proposed Bonnet addresses this by integrating a series of novel framework components including HU-based bone thresholding, patch-wise inference with a sparse spconv-based U-Net, and multi-window fusion into a full-volume prediction. Trained on TotalSegmentator and evaluated without additional tuning on RibSeg, CT-Pelvic1K, and CT-Spine1K, Bonnet achieves high Dice across ribs, pelvis, and spine while running in only 2.69 seconds per scan on an RTX A6000. Compared to strong voxel baselines, Bonnet attains a similar accuracy but reduces inference time by roughly 25x on the same hardware and tiling setup. The toolkit and pre-trained models will be released at https://github.com/HINTLab/Bonnet.
Real-time path planning in constrained environments remains a fundamental challenge for autonomous systems. Traditional classical planners, while effective under perfect perception assumptions, are often sensitive to real-world perception constraints and rely on online search procedures that incur high computational costs. In complex surroundings, this renders real-time deployment prohibitive. To overcome these limitations, we introduce a Deep Reinforcement Learning (DRL) framework for real-time path planning in parking scenarios. In particular, we focus on challenging scenes with tight spaces that require a high number of reversal maneuvers and adjustments. Unlike classical planners, our solution does not require ideal and structured perception, and in principle, could avoid the need for additional modules such as localization and tracking, resulting in a simpler and more practical implementation. Also, at test time, the policy generates actions through a single forward pass at each step, which is lightweight enough for real-time deployment. The task is formulated as a sequential decision-making problem grounded in a bicycle model dynamics, enabling the agent to directly learn navigation policies that respect vehicle kinematics and environmental constraints in the closed-loop setting. A new benchmark is developed to support both training and evaluation, capturing diverse and challenging scenarios. Our approach achieves state-of-the-art success rates and efficiency, surpassing classical planner baselines by +96% in success rate and +52% in efficiency. Furthermore, we release our benchmark as an open-source resource for the community to foster future research in autonomous systems. The benchmark and accompanying tools are available at https://github.com/dqm5rtfg9b-collab/Constrained_Parking_Scenarios.
Humans learn efficiently from their environment by engaging multiple interacting neural systems that support distinct yet complementary forms of control, including model-based (goal-directed) planning, model-free (habitual) responding, and episodic memory-based learning. Model-based mechanisms compute prospective action values using an internal model of the environment, supporting flexible but computationally costly planning; model-free mechanisms cache value estimates and build heuristics that enable fast, efficient habitual responding; and memory-based mechanisms allow rapid adaptation from individual experience. In this work, we aim to elucidate the computational principles underlying this biological efficiency and translate them into a sampling algorithm for scalable Bayesian inference through effective exploration of the posterior distribution. More specifically, our proposed algorithm comprises three components: a model-based module that uses the target distribution for guided but computationally slow sampling; a model-free module that uses previous samples to learn patterns in the parameter space, enabling fast, reflexive sampling without directly evaluating the expensive target distribution; and an episodic-control module that supports rapid sampling by recalling specific past events (i.e., samples). We show that this approach advances Bayesian methods and facilitates their application to large-scale statistical machine learning problems. In particular, we apply our proposed framework to Bayesian deep learning, with an emphasis on proper and principled uncertainty quantification.
Multimodal Large Language Model (MLLM) agents facilitate Graphical User Interface (GUI) automation but struggle with long-horizon, cross-application tasks due to limited context windows. While memory systems provide a viable solution, existing paradigms struggle to adapt to dynamic GUI environments, suffering from a granularity mismatch between high-level intent and low-level execution, and context pollution where the static accumulation of outdated experiences drives agents into hallucination. To address these bottlenecks, we propose the Darwinian Memory System (DMS), a self-evolving architecture that constructs memory as a dynamic ecosystem governed by the law of survival of the fittest. DMS decomposes complex trajectories into independent, reusable units for compositional flexibility, and implements Utility-driven Natural Selection to track survival value, actively pruning suboptimal paths and inhibiting high-risk plans. This evolutionary pressure compels the agent to derive superior strategies. Extensive experiments on real-world multi-app benchmarks validate that DMS boosts general-purpose MLLMs without training costs or architectural overhead, achieving average gains of 18.0% in success rate and 33.9% in execution stability, while reducing task latency, establishing it as an effective self-evolving memory system for GUI tasks.
Recently, the integration of unmanned aerial vehicle (UAV) and visible light communication (VLC) technologies has emerged as a promising solution to offer flexible communication and efficient lighting. This letter investigates the three-dimensional trajectory planning in a UAV-assisted VLC system, where a UAV is dispatched to collect data from ground users (GUs). The core objective is to develop a trajectory planning framework that minimizes UAV flight distance, which is equivalent to maximizing the data collection efficiency. This issue is formulated as a challenging mixed-integer non-convex optimization problem. To tackle it, we first derive a closed-form optimal flight altitude under specific VLC channel gain threshold. Subsequently, we optimize the UAV horizontal trajectory by integrating a novel pheromone-driven reward mechanism with the twin delayed deep deterministic policy gradient algorithm, which enables adaptive UAV motion strategy in complex environments. Simulation results validate that the derived optimal altitude effectively reduces the flight distance by up to 35% compared to baseline methods. Additionally, the proposed reward mechanism significantly shortens the convergence steps by approximately 50%, demonstrating notable efficiency gains in the context of UAV-assisted VLC data collection.
Hierarchical policies in offline goal-conditioned reinforcement learning (GCRL) addresses long-horizon tasks by decomposing control into high-level subgoal planning and low-level action execution. A critical design choice in such architectures is the goal representation-the compressed encoding of goals that serves as the interface between these levels. Existing approaches commonly derive goal representations while learning value functions, implicitly assuming that preserving information sufficient for value estimation is adequate for optimal control. We show that this assumption can fail, even when the value estimation is exact, as such representations may collapse goal states that need to be differentiated for action learning. To address this, we introduce an information-theoretic framework that defines action sufficiency, a condition on goal representations necessary for optimal action selection. We prove that value sufficiency does not imply action sufficiency and empirically verify that the latter is more strongly associated with control success in a discrete environment. We further demonstrate that standard log-loss training of low-level policies naturally induces action-sufficient representations. Our experimental results a popular benchmark demonstrate that our actor-derived representations consistently outperform representations learned via value estimation.
Reinforcement learning with verifiable rewards has emerged as a powerful paradigm for training intelligent agents. However, existing methods typically employ binary rewards that fail to capture quality differences among trajectories achieving identical outcomes, thereby overlooking potential diversity within the solution space. Inspired by the ``sweet spot'' concept in tennis-the racket's core region that produces optimal hitting effects, we introduce \textbf{S}weet \textbf{S}pot \textbf{L}earning (\textbf{SSL}), a novel framework that provides differentiated guidance for agent optimization. SSL follows a simple yet effective principle: progressively amplified, tiered rewards guide policies toward the sweet-spot region of the solution space. This principle naturally adapts across diverse tasks: visual perception tasks leverage distance-tiered modeling to reward proximity, while complex reasoning tasks reward incremental progress toward promising solutions. We theoretically demonstrate that SSL preserves optimal solution ordering and enhances the gradient signal-to-noise ratio, thereby fostering more directed optimization. Extensive experiments across GUI perception, short/long-term planning, and complex reasoning tasks show consistent improvements over strong baselines on 12 benchmarks, achieving up to 2.5X sample efficiency gains and effective cross-task transferability. Our work establishes SSL as a general principle for training capable and robust agents.
The Spin Physics Detector (SPD) is a universal detector in the one of two interaction points of the NICA collider under construction at JINR, Dubna. SPD plans to study the spin structure of the proton and deuteron and other spin-related phenomena using a unique possibility to operate with polarized proton and deuteron beams at a collision energy up to 27 GeV and a luminosity up to $10^{32}$ cm$^{-2}$ s$^{-1}$. As the main goal, the experiment aims to provide access to the gluon TMD PDFs in the proton and deuteron, as well as the gluon transversity distribution and tensor PDFs in the deuteron, via the measurements of specific single and double spin asymmetries using different complementary probes such as charmonia, open charm, and prompt photon production processes. Other polarized and unpolarized physics is possible, especially at the first stage of NICA operation with reduced luminosity and collision energy of the proton and ion beams. Construction of the first stage of the SPD facility is included in the JINR seven-year development plan for 2024-2030. The physics program of the SPD project and the design of the SPD setup are presented.
We focus on multi-object rearrangement planning in densely cluttered environments using a car-like robot pusher. The combination of kinematic, geometric and physics constraints underlying this domain results in challenging nonmonotone problem instances which demand breaking each manipulation action into multiple parts to achieve a desired object rearrangement. Prior work tackles such instances by planning prerelocations, temporary object displacements that enable constraint satisfaction, but deciding where to prerelocate remains difficult due to local minima leading to infeasible or high-cost paths. Our key insight is that these minima can be avoided by steering a prerelocation optimization toward low-cost regions informed by Dubins path classification. These optimized prerelocations are integrated into an object traversability graph that encodes kinematic, geometric, and pushing constraints. Searching this graph in a depth-first fashion results in efficient, feasible rearrangement sequences. Across a series of densely cluttered scenarios with up to 13 objects, our framework, ReloPush-BOSS, exhibits consistently highest success rates and shortest pushing paths compared to state-of-the-art baselines. Hardware experiments on a 1/10 car-like pusher demonstrate the robustness of our approach. Code and footage from our experiments can be found at: https://fluentrobotics.com/relopushboss.
Accurate and timely population data are essential for disaster response and humanitarian planning, but traditional censuses often cannot capture rapid demographic changes. Social media data offer a promising alternative for dynamic population monitoring, but their representativeness remains poorly understood and stringent privacy requirements limit their reliability. Here, we address these limitations in the context of the Philippines by calibrating Facebook user counts with the country's 2020 census figures. First, we find that differential privacy techniques commonly applied to social media-based population datasets disproportionately mask low-population areas. To address this, we propose a Bayesian imputation approach to recover missing values, restoring data coverage for $5.5\%$ of rural areas. Further, using the imputed social media data and leveraging predictors such as urbanisation level, demographic composition, and socio-economic status, we develop a statistical model for the proportion of Facebook users in each municipality, which links observed Facebook user numbers to the true population levels. Out-of-sample validation demonstrates strong result generalisability, with errors as low as ${\approx}18\%$ and ${\approx}24\%$ for urban and rural Facebook user proportions, respectively. We further demonstrate that accounting for overdispersion and spatial correlations in the data is crucial to obtain accurate estimates and appropriate credible intervals. Crucially, as predictors change over time, the models can be used to regularly update the population predictions, providing a dynamic complement to census-based estimates. These results have direct implications for humanitarian response in disaster-prone regions and offer a general framework for using biased social media signals to generate reliable and timely population data.
To enhance the efficiency of capacity markets, many electricity markets in the U.S. are adopting or planning to implement marginal capacity accreditation reforms. This paper provides new insights into energy storage capacity accreditation using Marginal Reliability Impact (MRI). We reformulate the commonly used reliability-based storage dispatch model as an optimization problem, enabling direct calculation of the MRI from the Lagrange multipliers, rather than using brute-force perturbation analysis. The analysis demonstrates that the EUE is a piecewise linear function and the storage MRI retains a non-negative property across various system scenarios. We further explore the influence of qualified capacity (QC), storage dispatch rules, and other key factors on storage accreditation, providing practical insights for system operators. Additionally, comparisons of storage capacity accreditation under different reliability criteria offer valuable guidance for policymakers in setting future standards. Numerical results from a modified California system validate our findings and highlight several important phenomena associated with the MRI-based accreditation scheme.
Retrieval-augmented generation is a practical paradigm for question answering over long documents, but it remains brittle for multimodal reading where text, tables, and figures are interleaved across many pages. First, flat chunking breaks document-native structure and cross-modal alignment, yielding semantic fragments that are hard to interpret in isolation. Second, even iterative retrieval can fail in long contexts by looping on partial evidence or drifting into irrelevant sections as noise accumulates, since each step is guided only by the current snippet without a persistent global search state. We introduce $G^2$-Reader, a dual-graph system, to address both issues. It evolves a Content Graph to preserve document-native structure and cross-modal semantics, and maintains a Planning Graph, an agentic directed acyclic graph of sub-questions, to track intermediate findings and guide stepwise navigation for evidence completion. On VisDoMBench across five multimodal domains, $G^2$-Reader with Qwen3-VL-32B-Instruct reaches 66.21\% average accuracy, outperforming strong baselines and a standalone GPT-5 (53.08\%).
Scaling has powered recent advances in vision foundation models, yet extending this paradigm to metric depth estimation remains challenging due to heterogeneous sensor noise, camera-dependent biases, and metric ambiguity in noisy cross-source 3D data. We introduce Metric Anything, a simple and scalable pretraining framework that learns metric depth from noisy, diverse 3D sources without manually engineered prompts, camera-specific modeling, or task-specific architectures. Central to our approach is the Sparse Metric Prompt, created by randomly masking depth maps, which serves as a universal interface that decouples spatial reasoning from sensor and camera biases. Using about 20M image-depth pairs spanning reconstructed, captured, and rendered 3D data across 10000 camera models, we demonstrate-for the first time-a clear scaling trend in the metric depth track. The pretrained model excels at prompt-driven tasks such as depth completion, super-resolution and Radar-camera fusion, while its distilled prompt-free student achieves state-of-the-art results on monocular depth estimation, camera intrinsics recovery, single/multi-view metric 3D reconstruction, and VLA planning. We also show that using pretrained ViT of Metric Anything as a visual encoder significantly boosts Multimodal Large Language Model capabilities in spatial intelligence. These results show that metric depth estimation can benefit from the same scaling laws that drive modern foundation models, establishing a new path toward scalable and efficient real-world metric perception. We open-source MetricAnything at http://metric-anything.github.io/metric-anything-io/ to support community research.
End-to-end autonomous driving increasingly leverages self-supervised video pretraining to learn transferable planning representations. However, pretraining video world models for scene understanding has so far brought only limited improvements. This limitation is compounded by the inherent ambiguity of driving: each scene typically provides only a single human trajectory, making it difficult to learn multimodal behaviors. In this work, we propose Drive-JEPA, a framework that integrates Video Joint-Embedding Predictive Architecture (V-JEPA) with multimodal trajectory distillation for end-to-end driving. First, we adapt V-JEPA for end-to-end driving, pretraining a ViT encoder on large-scale driving videos to produce predictive representations aligned with trajectory planning. Second, we introduce a proposal-centric planner that distills diverse simulator-generated trajectories alongside human trajectories, with a momentum-aware selection mechanism to promote stable and safe behavior. When evaluated on NAVSIM, the V-JEPA representation combined with a simple transformer-based decoder outperforms prior methods by 3 PDMS in the perception-free setting. The complete Drive-JEPA framework achieves 93.3 PDMS on v1 and 87.8 EPDMS on v2, setting a new state-of-the-art.
Volumetric medical imaging offers great potential for understanding complex pathologies. Yet, traditional 2D slices provide little support for interpreting spatial relationships, forcing users to mentally reconstruct anatomy into three dimensions. Direct volumetric path tracing and VR rendering can improve perception but are computationally expensive, while precomputed representations, like Gaussian Splatting, require planning ahead. Both approaches limit interactive use. We propose a hybrid rendering approach for high-quality, interactive, and immersive anatomical visualization. Our method combines streamed foveated path tracing with a lightweight Gaussian Splatting approximation of the periphery. The peripheral model generation is optimized with volume data and continuously refined using foveal renderings, enabling interactive updates. Depth-guided reprojection further improves robustness to latency and allows users to balance fidelity with refresh rate. We compare our method against direct path tracing and Gaussian Splatting. Our results highlight how their combination can preserve strengths in visual quality while re-generating the peripheral model in under a second, eliminating extensive preprocessing and approximations. This opens new options for interactive medical visualization.
Real-world reinforcement learning is often \emph{nonstationary}: rewards and dynamics drift, accelerate, oscillate, and trigger abrupt switches in the optimal action. Existing theory often represents nonstationarity with coarse-scale models that measure \emph{how much} the environment changes, not \emph{how} it changes locally -- even though acceleration and near-ties drive tracking error and policy chattering. We take a geometric view of nonstationary discounted Markov Decision Processes (MDPs) by modeling the environment as a differentiable homotopy path and tracking the induced motion of the optimal Bellman fixed point. This yields a length-curvature-kink signature of intrinsic complexity: cumulative drift, acceleration/oscillation, and action-gap-induced nonsmoothness. We prove a solver-agnostic path-integral stability bound and derive gap-safe feasible regions that certify local stability away from switch regimes. Building on these results, we introduce \textit{Homotopy-Tracking RL (HT-RL)} and \textit{HT-MCTS}, lightweight wrappers that estimate replay-based proxies of length, curvature, and near-tie proximity online and adapt learning or planning intensity accordingly. Experiments show improved tracking and dynamic regret over matched static baselines, with the largest gains in oscillatory and switch-prone regimes.
AI research has traditionally prioritised algorithmic performance, such as optimising accuracy in machine learning or runtime in automated planning. The emerging paradigm of Green AI challenges this by recognising energy consumption as a critical performance dimension. Despite the high computational demands of automated planning, its energy efficiency has received little attention. This gap is particularly salient given the modular planning structure, in which domain models are specified independently of algorithms. On the other hand, this separation also enables systematic analysis of energy usage through domain model design. We empirically investigate how domain model characteristics affect the energy consumption of classical planners. We introduce a domain model configuration framework that enables controlled variation of features, such as element ordering, action arity, and dead-end states. Using five benchmark domains and five state-of-the-art planners, we analyse energy and runtime impacts across 32 domain variants per benchmark. Results demonstrate that domain-level modifications produce measurable energy differences across planners, with energy consumption not always correlating with runtime.
The performance of speaker verification systems degrades significantly under language mismatch, a critical challenge exacerbated by the field's reliance on English-centric data. To address this, we propose the TidyVoice Challenge for cross-lingual speaker verification. The challenge leverages the TidyVoiceX dataset from the novel TidyVoice benchmark, a large-scale, multilingual corpus derived from Mozilla Common Voice, and specifically curated to isolate the effect of language switching across approximately 40 languages. Participants will be tasked with building systems robust to this mismatch, with performance primarily evaluated using the Equal Error Rate on cross-language trials. By providing standardized data, open-source baselines, and a rigorous evaluation protocol, this challenge aims to drive research towards fairer, more inclusive, and language-independent speaker recognition technologies, directly aligning with the Interspeech 2026 theme, "Speaking Together."
The evolution of Retrieval-Augmented Generation (RAG) has shifted from static retrieval pipelines to dynamic, agentic workflows where a central planner orchestrates multi-turn reasoning. However, existing paradigms face a critical dichotomy: they either optimize modules jointly within rigid, fixed-graph architectures, or empower dynamic planning while treating executors as frozen, black-box tools. We identify that this \textit{decoupled optimization} creates a ``strategic-operational mismatch,'' where sophisticated planning strategies fail to materialize due to unadapted local executors, often leading to negative performance gains despite increased system complexity. In this paper, we propose \textbf{JADE} (\textbf{J}oint \textbf{A}gentic \textbf{D}ynamic \textbf{E}xecution), a unified framework for the joint optimization of planning and execution within dynamic, multi-turn workflows. By modeling the system as a cooperative multi-agent team unified under a single shared backbone, JADE enables end-to-end learning driven by outcome-based rewards. This approach facilitates \textit{co-adaptation}: the planner learns to operate within the capability boundaries of the executors, while the executors evolve to align with high-level strategic intent. Empirical results demonstrate that JADE transforms disjoint modules into a synergistic system, yielding remarkable performance improvements via joint optimization and enabling a flexible balance between efficiency and effectiveness through dynamic workflow orchestration.