A fundamental challenge in robot navigation lies in learning policies that generalize across diverse environments while conforming to the unique physical constraints and capabilities of a specific embodiment (e.g., quadrupeds can walk up stairs, but rovers cannot). We propose VAMOS, a hierarchical VLA that decouples semantic planning from embodiment grounding: a generalist planner learns from diverse, open-world data, while a specialist affordance model learns the robot's physical constraints and capabilities in safe, low-cost simulation. We enabled this separation by carefully designing an interface that lets a high-level planner propose candidate paths directly in image space that the affordance model then evaluates and re-ranks. Our real-world experiments show that VAMOS achieves higher success rates in both indoor and complex outdoor navigation than state-of-the-art model-based and end-to-end learning methods. We also show that our hierarchical design enables cross-embodied navigation across legged and wheeled robots and is easily steerable using natural language. Real-world ablations confirm that the specialist model is key to embodiment grounding, enabling a single high-level planner to be deployed across physically distinct wheeled and legged robots. Finally, this model significantly enhances single-robot reliability, achieving 3X higher success rates by rejecting physically infeasible plans. Website: https://vamos-vla.github.io/
Efficient analysis and processing of dental images are crucial for dentists to achieve accurate diagnosis and optimal treatment planning. However, dental imaging inherently poses several challenges, such as low contrast, metallic artifacts, and variations in projection angles. Combined with the subjectivity arising from differences in clinicians' expertise, manual interpretation often proves time-consuming and prone to inconsistency. Artificial intelligence (AI)-based automated dental image analysis (DIA) offers a promising solution to these issues and has become an integral part of computer-aided dental diagnosis and treatment. Among various AI technologies, deep learning (DL) stands out as the most widely applied and influential approach due to its superior feature extraction and representation capabilities. To comprehensively summarize recent progress in this field, we focus on the two fundamental aspects of DL research-datasets and models. In this paper, we systematically review 260 studies on DL applications in DIA, including 49 papers on publicly available dental datasets and 211 papers on DL-based algorithms. We first introduce the basic concepts of dental imaging and summarize the characteristics and acquisition methods of existing datasets. Then, we present the foundational techniques of DL and categorize relevant models and algorithms according to different DIA tasks, analyzing their network architectures, optimization strategies, training methods, and performance. Furthermore, we summarize commonly used training and evaluation metrics in the DIA domain. Finally, we discuss the current challenges of existing research and outline potential future directions. We hope that this work provides a valuable and systematic reference for researchers in this field. All supplementary materials and detailed comparison tables will be made publicly available on GitHub.
Rail is a cost-effective and relatively low-emission mode for transporting intermodal containers over long distances. This paper addresses tactical planning of intermodal railroad operations by introducing a new problem that simultaneously considers three consolidation processes and the management of a heterogeneous railcar fleet. We model the problem with a scheduled service network design with resource management (SSND-RM) formulation, expressed as an integer linear program. While such formulations are challenging to solve at scale, we demonstrate that our problem can be tackled with a general-purpose solver when provided with high-quality warm-start solutions. To this end, we design a construction heuristic inspired by a relax-and-fix procedure. We evaluate the methodology on realistic, large-scale instances from our industrial partner, the Canadian National Railway Company: a North American Class I railroad. The computational experiments show that the proposed approach efficiently solves practically relevant instances, and that solutions to the SSND-RM formulation yield substantially more accurate capacity estimations compared to those obtained from simpler baseline models. Managerial insights from our study highlight that ignoring railcar fleet management or container loading constraints can lead to a severe underestimation of required capacity, which may result in costly operational inefficiencies. Furthermore, our results show that the use of multi-platform railcars improves overall capacity utilization and benefits the network, even if they can locally lead to less efficient loading as measured by terminal-level slot utilization performance indicators.
Reinforcement learning has recently shown promise in improving retrieval-augmented generation (RAG). Despite these advances, its effectiveness in multi-hop question answering (QA) remains limited by two fundamental limitations: (i) global planning absence to structure multi-step reasoning, and (ii) unfaithful execution, which hinders effective query formulation and consistent use of retrieved evidence. We propose GlobalRAG, a reinforcement learning framework designed to enhance global reasoning in multi-hop QA. GlobalRAG decomposes questions into subgoals, coordinates retrieval with reasoning, and refines evidence iteratively. To guide this process, we introduce Planning Quality Reward and SubGoal Completion Reward, which encourage coherent planning and reliable subgoal execution. In addition, a progressive weight annealing strategy balances process-oriented and outcome-based objectives. Extensive experiments on both in-domain and out-of-domain benchmarks demonstrate that GlobalRAG significantly outperforms strong baselines while using only 8k training data (42% of the training data used by strong baselines), achieving average improvements of 14.2% in both EM and F1.
Zero-shot reinforcement learning (RL) has emerged as a setting for developing general agents in an unsupervised manner, capable of solving downstream tasks without additional training or planning at test-time. Unlike conventional RL, which optimizes policies for a fixed reward, zero-shot RL requires agents to encode representations rich enough to support immediate adaptation to any objective, drawing parallels to vision and language foundation models. Despite growing interest, the field lacks a common analytical lens. We present the first unified framework for zero-shot RL. Our formulation introduces a consistent notation and taxonomy that organizes existing approaches and allows direct comparison between them. Central to our framework is the classification of algorithms into two families: direct representations, which learn end-to-end mappings from rewards to policies, and compositional representations, which decompose the representation leveraging the substructure of the value function. Within this framework, we highlight shared principles and key differences across methods, and we derive an extended bound for successor-feature methods, offering a new perspective on their performance in the zero-shot regime. By consolidating existing work under a common lens, our framework provides a principled foundation for future research in zero-shot RL and outlines a clear path toward developing more general agents.
In this paper the computational challenges of time-optimal path following are addressed. The standard approach is to minimize the travel time, which inevitably leads to singularities at zero path speed, when reformulating the optimization problem in terms of a path parameter. Thus, smooth trajectory generation while maintaining a low computational effort is quite challenging, since the singularities have to be taken into account. To this end, a different approach is presented in this paper. This approach is based on maximizing the path speed along a prescribed path. Furthermore, the approach is capable of planning smooth trajectories numerically efficient. Moreover, the discrete reformulation of the underlying problem is linear in optimization variables.
This paper presents a method for planning a trajectory in workspace coordinates using a spatially fixed tool center point (TCP), while taking into account the processing path on a part. This approach is beneficial if it is easier to move the part rather than moving the tool. Whether a mathematical description that defines the shape to be processed or single points from a design program are used, the robot path is finally represented using B-splines. The use of splines enables the path to be continuous with a desired degree, which finally leads to a smooth robot trajectory. While calculating the robot trajectory through prescribed orientation, additionally a given velocity at the TCP has to be considered. The procedure was validated on a real system using an industrial robot moving an arbitrary defined part.
Multicast allows sending a message to multiple recipients without having to create and send a separate message for each recipient. This preserves network bandwidth, which is particularly important in time-sensitive networks. These networks are commonly used to provide latency-bounded communication for real-time systems in domains like automotive, avionics, industrial internet of things, automated shop floors, and smart energy grids. The preserved bandwidth can be used to admit additional real-time messages with specific quality of service requirements or to reduce the end-to-end latencies for messages of any type. However, using multicast communication can complicate traffic planning, as it requires free queues or available downstream egress ports on all branches of the multicast tree. In this work, we present a novel multicast partitioning technique to split multicast trees into smaller multicast or unicast trees. This allows for a more fine-grained trade-off between bandwidth utilization and traffic scheduling difficulty. Thus, schedulability in dynamic systems can be improved, in terms the number of admitted streams and the accumulated network throughput. We evaluated the multicast partitioning on different network topologies and with three different scheduling algorithms. With the partitioning, 5-15\% fewer streams were rejected, while achieving 5-125\% more network throughput, depending on the scheduling algorithm.
The Surface Brightness Fluctuation method is one of the most reliable and efficient ways of measuring distances to galaxies within 100 Mpc. While recent implementations have increasingly relied on space-based observations, SBF remains effective when applied to ground-based data. In particular, deep, wide-field imaging surveys with sub-arcsecond seeing conditions allows us for accurate SBF measurements across large samples of galaxies. With the upcoming next generation wide-area imaging surveys, the thousands of galaxies suitable for SBF measurements will give us the opportunity to constrain the 3D structure of the local universe. We present FAST-SBF, a new Python-based pipeline for measuring SBF, developed to support the analysis of large datasets from upcoming wide-field imaging surveys such as LSST, Euclid, and Roman. The procedure, still in the testing and development stage, is designed for automation and minimal user intervention, offering a fast and flexible approach to SBF distance estimation. We validate the performance of the procedure on high-quality imaging data from the Hyper Suprime-Cam Subaru Strategic Program (HSC-SSP), a precursor to LSST, analyzing a sample of both luminous early-type galaxies and fainter dwarfs. Our measurements are also compared with recent results from the Next Generation Virgo Cluster Survey (NGVS) and with the SPoT stellar population synthesis models. The results show excellent agreement with published distances, with the capability of measuring the SBF signal also for faint dwarf galaxies. The pipeline allows the user to completely analyze a galaxy in relatively short time ($\approx$ minutes) and significantly reduces the need for user intervention. reduces at minimum the user intervention. The FAST-SBF tool is planned for public release to support the community in using SBF as a distance indicator in next-generation surveys.
Parking is a critical pillar of driving safety. While recent end-to-end (E2E) approaches have achieved promising in-domain results, robustness under domain shifts (e.g., weather and lighting changes) remains a key challenge. Rather than relying on additional data, in this paper, we propose Dino-Diffusion Parking (DDP), a domain-agnostic autonomous parking pipeline that integrates visual foundation models with diffusion-based planning to enable generalized perception and robust motion planning under distribution shifts. We train our pipeline in CARLA at regular setting and transfer it to more adversarial settings in a zero-shot fashion. Our model consistently achieves a parking success rate above 90% across all tested out-of-distribution (OOD) scenarios, with ablation studies confirming that both the network architecture and algorithmic design significantly enhance cross-domain performance over existing baselines. Furthermore, testing in a 3D Gaussian splatting (3DGS) environment reconstructed from a real-world parking lot demonstrates promising sim-to-real transfer.
Embodied Question Answering (EQA) requires agents to explore 3D environments to obtain observations and answer questions related to the scene. Existing methods leverage VLMs to directly explore the environment and answer questions without explicit thinking or planning, which limits their reasoning ability and results in excessive or inefficient exploration as well as ineffective responses. In this paper, we introduce ToolEQA, an agent that integrates external tools with multi-step reasoning, where external tools can provide more useful information for completing the task, helping the model derive better exploration directions in the next step of reasoning and thus obtaining additional effective information. This enables ToolEQA to generate more accurate responses with a shorter exploration distance. To enhance the model's ability for tool-usage and multi-step reasoning, we further design a novel EQA data generation pipeline that automatically constructs large-scale EQA tasks with reasoning trajectories and corresponding answers. Based on the pipeline, we collect the EQA-RT dataset that contains about 18K tasks, divided into a training set EQA-RT-Train, and two test sets EQA-RT-Seen (scenes overlapping with the training set) and EQA-RT-Unseen (novel scenes). Experiments on EQA-RT-Seen and EQA-RT-Unseen show that ToolEQA improves the success rate by 9.2~20.2% over state-of-the-art baselines, while outperforming the zero-shot ToolEQA by 10% in success rate. In addition, ToolEQA also achieves state-of-the-art performance on the HM-EQA, OpenEQA, and EXPRESS-Bench datasets, demonstrating its generality. Our homepage see https://tooleqa.github.io.
Finding optimal join orders is among the most crucial steps to be performed by query optimisers. Though extensively studied in data management research, the problem remains far from solved: While query optimisers rely on exhaustive search methods to determine ideal solutions for small problems, such methods reach their limits once queries grow in size. Yet, large queries become increasingly common in real-world scenarios, and require suitable methods to generate efficient execution plans. While a variety of heuristics have been proposed for large-scale query optimisation, they suffer from degrading solution quality as queries grow in size, or feature highly sub-optimal worst-case behavior, as we will show. We propose a novel method based on the paradigm of mixed integer linear programming (MILP): By deriving a novel MILP model capable of optimising arbitrary bushy tree structures, we address the limitations of existing MILP methods for join ordering, and can rely on highly optimised MILP solvers to derive efficient tree structures that elude competing methods. To ensure optimisation efficiency, we embed our MILP method into a hybrid framework, which applies MILP solvers precisely where they provide the greatest advantage over competitors, while relying on more efficient methods for less complex optimisation steps. Thereby, our approach gracefully scales to extremely large query sizes joining up to 100 relations, and consistently achieves the most robust plan quality among a large variety of competing join ordering methods.
Human mobility forecasting is crucial for disaster relief, city planning, and public health. However, existing models either only model location sequences or include time information merely as auxiliary input, thereby failing to leverage the rich semantic context provided by points of interest (POIs). To address this, we enrich a BERT-based mobility model with derived temporal descriptors and POI embeddings to better capture the semantics underlying human movement. We propose STaBERT (Semantic-Temporal aware BERT), which integrates both POI and temporal information at each location to construct a unified, semantically enriched representation of mobility. Experimental results show that STaBERT significantly improves prediction accuracy: for single-city prediction, the GEO-BLEU score improved from 0.34 to 0.75; for multi-city prediction, from 0.34 to 0.56.
Robots often face manipulation tasks in environments where vision is inadequate due to clutter, occlusions, or poor lighting--for example, reaching a shutoff valve at the back of a sink cabinet or locating a light switch above a crowded shelf. In such settings, robots, much like humans, must rely on contact feedback to distinguish free from occupied space and navigate around obstacles. Many of these environments often exhibit strong structural priors--for instance, pipes often span across sink cabinets--that can be exploited to anticipate unseen structure and avoid unnecessary collisions. We present a theoretically complete and empirically efficient framework for manipulation in the blind that integrates contact feedback with structural priors to enable robust operation in unknown environments. The framework comprises three tightly coupled components: (i) a contact detection and localization module that utilizes joint torque sensing with a contact particle filter to detect and localize contacts, (ii) an occupancy estimation module that uses the history of contact observations to build a partial occupancy map of the workspace and extrapolate it into unexplored regions with learned predictors, and (iii) a planning module that accounts for the fact that contact localization estimates and occupancy predictions can be noisy, computing paths that avoid collisions and complete tasks efficiently without eliminating feasible solutions. We evaluate the system in simulation and in the real world on a UR10e manipulator across two domestic tasks--(i) manipulating a valve under a kitchen sink surrounded by pipes and (ii) retrieving a target object from a cluttered shelf. Results show that the framework reliably solves these tasks, achieving up to a 2x reduction in task completion time compared to baselines, with ablations confirming the contribution of each module.
Robotic arms require precise, task-aware trajectory planning, yet sequence models that ignore motion structure often yield invalid or inefficient executions. We present a Path-based Transformer that encodes robot motion with a 3-grid (where/what/when) representation and constraint-masked decoding, enforcing lattice-adjacent moves and workspace bounds while reasoning over task graphs and action order. Trained on 53,755 trajectories (80% train / 20% validation), the model aligns closely with ground truth -- 89.44% stepwise accuracy, 93.32% precision, 89.44% recall, and 90.40% F1 -- with 99.99% of paths legal by construction. Compiled to motor primitives on an xArm Lite 6 with a depth-camera digital twin, it attains up to 97.5% reach and 92.5% pick success in controlled tests, and 86.7% end-to-end success across 60 language-specified tasks in cluttered scenes, absorbing slips and occlusions via local re-grounding without global re-planning. These results show that path-structured representations enable Transformers to generate accurate, reliable, and interpretable robot trajectories, bridging graph-based planning and sequence-based learning and providing a practical foundation for general-purpose manipulation and sim-to-real transfer.
Radio Frequency Identification (RFID) tracking may be a viable solution for defense assets that must be stored in accordance with security guidelines. However, poor sensor specificity (vulnerabilities include long range detection, spoofing, and counterfeiting) can lead to erroneous detection and operational security events. We present a supervised learning simulation with realistic Received Signal Strength Indicator (RSSI) data and Decision Tree classification in a Computer Assisted Design (CAD)-modeled floor plan that encapsulates some of the challenges encountered in defense storage. In this work, we focused on classifying 12 lab zones (LabZoneA-L) to perform location inference. The raw dataset had approximately 980,000 reads. Class frequencies were imbalanced, and class weights were calculated to account for class imbalance in this multi-class setting. The model, trained on stratified subsamples to 5,000 balanced observations, yielded an overall accuracy of 34.2% and F1-scores greater than 0.40 for multiple zones (Zones F, G, H, etc.). However, rare classes (most notably LabZoneC) were often misclassified, even with the use of class weights. An adjacency-aware confusion matrix was calculated to allow better interpretation of physically adjacent zones. These results suggest that RSSI-based decision trees can be applied in realistic simulations to enable zone-level anomaly detection or misplacement monitoring for defense supply logistics. Reliable classification performance in low-coverage and low-signal zones could be improved with better antenna placement or additional sensors and sensor fusion with other modalities.
This paper tackles the challenge of learning a generalizable minimum-time flight policy for UAVs, capable of navigating between arbitrary start and goal states while balancing agile flight and stable hovering. Traditional approaches, particularly in autonomous drone racing, achieve impressive speeds and agility but are constrained to predefined track layouts, limiting real-world applicability. To address this, we propose a reinforcement learning-based framework that simultaneously learns state-to-state minimum-time planning and control and generalizes to arbitrary state-to-state flights. Our approach leverages Point Mass Model (PMM) trajectories as proxy rewards to approximate the true optimal flight objective and employs curriculum learning to scale the training process efficiently and to achieve generalization. We validate our method through simulation experiments, comparing it against Nonlinear Model Predictive Control (NMPC) tracking PMM-generated trajectories and conducting ablation studies to assess the impact of curriculum learning. Finally, real-world experiments confirm the robustness of our learned policy in outdoor environments, demonstrating its ability to generalize and operate on a small ARM-based single-board computer.
Generative Artificial Intelligence (GenAI) presents transformative opportunities for organizations, yet both midsize organizations and larger enterprises face distinctive adoption challenges. Midsize organizations encounter resource constraints and limited AI expertise, while enterprises struggle with organizational complexity and coordination challenges. Existing technology adoption frameworks, including TAM (Technology Acceptance Model), TOE (Technology Organization Environment), and DOI (Diffusion of Innovations) theory, lack the specificity required for GenAI implementation across these diverse contexts, creating a critical gap in adoption literature. This paper introduces FAIGMOE (Framework for the Adoption and Integration of Generative AI in Midsize Organizations and Enterprises), a conceptual framework addressing the unique needs of both organizational types. FAIGMOE synthesizes technology adoption theory, organizational change management, and innovation diffusion perspectives into four interconnected phases: Strategic Assessment, Planning and Use Case Development, Implementation and Integration, and Operationalization and Optimization. Each phase provides scalable guidance on readiness assessment, strategic alignment, risk governance, technical architecture, and change management adaptable to organizational scale and complexity. The framework incorporates GenAI specific considerations including prompt engineering, model orchestration, and hallucination management that distinguish it from generic technology adoption frameworks. As a perspective contribution, FAIGMOE provides the first comprehensive conceptual framework explicitly addressing GenAI adoption across midsize and enterprise organizations, offering actionable implementation protocols, assessment instruments, and governance templates requiring empirical validation through future research.
Building agents that generalize across web, desktop, and mobile environments remains an open challenge, as prior systems rely on environment-specific interfaces that limit cross-platform deployment. We introduce Surfer 2, a unified architecture operating purely from visual observations that achieves state-of-the-art performance across all three environments. Surfer 2 integrates hierarchical context management, decoupled planning and execution, and self-verification with adaptive recovery, enabling reliable operation over long task horizons. Our system achieves 97.1% accuracy on WebVoyager, 69.6% on WebArena, 60.1% on OSWorld, and 87.1% on AndroidWorld, outperforming all prior systems without task-specific fine-tuning. With multiple attempts, Surfer 2 exceeds human performance on all benchmarks. These results demonstrate that systematic orchestration amplifies foundation model capabilities and enables general-purpose computer control through visual interaction alone, while calling for a next-generation vision language model to achieve Pareto-optimal cost-efficiency.
Planning with world models offers a powerful paradigm for robotic control. Conventional approaches train a model to predict future frames conditioned on current frames and actions, which can then be used for planning. However, the objective of predicting future pixels is often at odds with the actual planning objective; strong pixel reconstruction does not always correlate with good planning decisions. This paper posits that instead of reconstructing future frames as pixels, world models only need to predict task-relevant semantic information about the future. For such prediction the paper poses world modeling as a visual question answering problem about semantic information in future frames. This perspective allows world modeling to be approached with the same tools underlying vision language models. Thus vision language models can be trained as "semantic" world models through a supervised finetuning process on image-action-text data, enabling planning for decision-making while inheriting many of the generalization and robustness properties from the pretrained vision-language models. The paper demonstrates how such a semantic world model can be used for policy improvement on open-ended robotics tasks, leading to significant generalization improvements over typical paradigms of reconstruction-based action-conditional world modeling. Website available at https://weirdlabuw.github.io/swm.
Recent advances in multimodal models have demonstrated remarkable text-guided image editing capabilities, with systems like GPT-4o and Nano-Banana setting new benchmarks. However, the research community's progress remains constrained by the absence of large-scale, high-quality, and openly accessible datasets built from real images. We introduce Pico-Banana-400K, a comprehensive 400K-image dataset for instruction-based image editing. Our dataset is constructed by leveraging Nano-Banana to generate diverse edit pairs from real photographs in the OpenImages collection. What distinguishes Pico-Banana-400K from previous synthetic datasets is our systematic approach to quality and diversity. We employ a fine-grained image editing taxonomy to ensure comprehensive coverage of edit types while maintaining precise content preservation and instruction faithfulness through MLLM-based quality scoring and careful curation. Beyond single turn editing, Pico-Banana-400K enables research into complex editing scenarios. The dataset includes three specialized subsets: (1) a 72K-example multi-turn collection for studying sequential editing, reasoning, and planning across consecutive modifications; (2) a 56K-example preference subset for alignment research and reward model training; and (3) paired long-short editing instructions for developing instruction rewriting and summarization capabilities. By providing this large-scale, high-quality, and task-rich resource, Pico-Banana-400K establishes a robust foundation for training and benchmarking the next generation of text-guided image editing models.
Model-learning agents should gather information to learn world models that support many downstream tasks and inferences, such as predicting unobserved states, estimating near- and far-term consequences of actions, planning action sequences, and detecting changes in dynamics. Current methods for learning and evaluating world models diverge from this goal: training and evaluation are anchored to next-frame prediction, and success is scored by reward maximization in the same environment. We propose WorldTest, a protocol to evaluate model-learning agents that separates reward-free interaction from a scored test phase in a different but related environment. WorldTest is open-ended$\unicode{x2014}$models should support many different tasks unknown ahead of time$\unicode{x2014}$and agnostic to model representation, allowing comparison across approaches. We instantiated WorldTest with AutumnBench, a suite of 43 interactive grid-world environments and 129 tasks across three families: masked-frame prediction, planning, and predicting changes to the causal dynamics. We compared 517 human participants and three frontier models on AutumnBench. We found that humans outperform the models, and scaling compute improves performance only in some environments but not others. WorldTest provides a novel template$\unicode{x2014}$reward-free exploration, derived tests, and behavior-based scoring$\unicode{x2014}$to evaluate what agents learn about environment dynamics, and AutumnBench exposes significant headroom in world-model learning.
We propose explicitly incorporating large-scale load siting into a stochastic nodal power system capacity expansion planning model that concurrently co-optimizes generation, transmission and storage expansion. The potential operational flexibility of some of these large loads is also taken into account by considering them as consisting of a set of tranches with different reliability requirements, which are modeled as a constraint on expected served energy across operational scenarios. We implement our model as a two-stage stochastic mixed-integer optimization problem with cross-scenario expectation constraints. To overcome the challenge of scalability, we build upon existing work to implement this model on a high performance computing platform and exploit scenario parallelization using an augmented Progressive Hedging Algorithm. The algorithm is implemented using the bounding features of mpisppy, which have shown to provide satisfactory provable optimality gaps despite the absence of theoretical guarantees of convergence. We test our approach to assess the value of this proactive planning framework on total system cost and reliability metrics using realistic testcases geographically assigned to San Diego and South Carolina, with datacenter and direct air capture facilities as large loads.
The traffic assignment problem is essential for traffic flow analysis, traditionally solved using mathematical programs under the Equilibrium principle. These methods become computationally prohibitive for large-scale networks due to non-linear growth in complexity with the number of OD pairs. This study introduces a novel data-driven approach using deep neural networks, specifically leveraging the Transformer architecture, to predict equilibrium path flows directly. By focusing on path-level traffic distribution, the proposed model captures intricate correlations between OD pairs, offering a more detailed and flexible analysis compared to traditional link-level approaches. The Transformer-based model drastically reduces computation time, while adapting to changes in demand and network structure without the need for recalculation. Numerical experiments are conducted on the Manhattan-like synthetic network, the Sioux Falls network, and the Eastern-Massachusetts network. The results demonstrate that the proposed model is orders of magnitude faster than conventional optimization. It efficiently estimates path-level traffic flows in multi-class networks, reducing computational costs and improving prediction accuracy by capturing detailed trip and flow information. The model also adapts flexibly to varying demand and network conditions, supporting traffic management and enabling rapid `what-if' analyses for enhanced transportation planning and policy-making.
Solving complex real-world control tasks often takes multiple tries: if we fail at first, we reflect on what went wrong, and change our strategy accordingly to avoid making the same mistake. In robotics, Vision-Language-Action models (VLAs) offer a promising path towards solving complex control tasks, but lack the ability to contextually and dynamically readjust behavior when they fail to accomplish a task. In this work, we introduce Learning from Inference-Time Execution (LITEN), which connects a VLA low-level policy to a high-level VLM that conditions on past experiences by including them in-context, allowing it to learn the affordances and capabilities of the low-level VLA. Our approach iterates between a reasoning phase that generates and executes plans for the low-level VLA, and an assessment phase that reflects on the resulting execution and draws useful conclusions to be included in future reasoning contexts. Unlike similar approaches to self-refinement in non-robotics domains, LITEN must reflect on unstructured real-world robot trajectories (e.g., raw videos), which requires structured guiderails during assessment. Our experimental results demonstrate LITEN is able to effectively learn from past experience to generate plans that use high-affordance instructions to accomplish long-horizon tasks.
Zero-shot Vision-and-Language Navigation in Continuous Environments (VLN-CE) requires an agent to navigate unseen environments based on natural language instructions without any prior training. Current methods face a critical trade-off: either rely on environment-specific waypoint predictors that limit scene generalization, or underutilize the reasoning capabilities of large models during navigation. We introduce LaViRA, a simple yet effective zero-shot framework that addresses this dilemma by decomposing action into a coarse-to-fine hierarchy: Language Action for high-level planning, Vision Action for perceptual grounding, and Robot Action for robust navigation. This modular decomposition allows us to leverage the distinct strengths of different scales of Multimodal Large Language Models (MLLMs) at each stage, creating a system that is powerful in its reasoning, grounding and practical control. LaViRA significantly outperforms existing state-of-the-art methods on the VLN-CE benchmark, demonstrating superior generalization capabilities in unseen environments, while maintaining transparency and efficiency for real-world deployment.
Despite remarkable progress in driving world models, their potential for autonomous systems remains largely untapped: the world models are mostly learned for world simulation and decoupled from trajectory planning. While recent efforts aim to unify world modeling and planning in a single framework, the synergistic facilitation mechanism of world modeling for planning still requires further exploration. In this work, we introduce a new driving paradigm named Policy World Model (PWM), which not only integrates world modeling and trajectory planning within a unified architecture, but is also able to benefit planning using the learned world knowledge through the proposed action-free future state forecasting scheme. Through collaborative state-action prediction, PWM can mimic the human-like anticipatory perception, yielding more reliable planning performance. To facilitate the efficiency of video forecasting, we further introduce a dynamically enhanced parallel token generation mechanism, equipped with a context-guided tokenizer and an adaptive dynamic focal loss. Despite utilizing only front camera input, our method matches or exceeds state-of-the-art approaches that rely on multi-view and multi-modal inputs. Code and model weights will be released at https://github.com/6550Zhao/Policy-World-Model.
Evacuation simulation is essential for building safety design, ensuring properly planned evacuation routes. However, traditional evacuation simulation relies heavily on refined modeling with extensive parameters, making it challenging to adopt such methods in a rapid iteration process in early design stages. Thus, this study proposes DiffEvac, a novel method to learn building evacuation patterns based on Generative Models (GMs), for efficient evacuation simulation and enhanced safety design. Initially, a dataset of 399 diverse functional layouts and corresponding evacuation heatmaps of buildings was established. Then, a decoupled feature representation is proposed to embed physical features like layouts and occupant density for GMs. Finally, a diffusion model based on image prompts is proposed to learn evacuation patterns from simulated evacuation heatmaps. Compared to existing research using Conditional GANs with RGB representation, DiffEvac achieves up to a 37.6% improvement in SSIM, 142% in PSNR, and delivers results 16 times faster, thereby cutting simulation time to 2 minutes. Case studies further demonstrate that the proposed method not only significantly enhances the rapid design iteration and adjustment process with efficient evacuation simulation but also offers new insights and technical pathways for future safety optimization in intelligent building design. The research implication is that the approach lowers the modeling burden, enables large-scale what-if exploration, and facilitates coupling with multi-objective design tools.
In the quest for scientific progress, communicating research is as vital as the discovery itself. Yet, researchers are often sidetracked by the manual, repetitive chore of building project webpages to make their dense papers accessible. While automation has tackled static slides and posters, the dynamic, interactive nature of webpages has remained an unaddressed challenge. To bridge this gap, we reframe the problem, arguing that the solution lies not in a single command, but in a collaborative, hierarchical process. We introduce $\textbf{AutoPage}$, a novel multi-agent system that embodies this philosophy. AutoPage deconstructs paper-to-page creation into a coarse-to-fine pipeline from narrative planning to multimodal content generation and interactive rendering. To combat AI hallucination, dedicated "Checker" agents verify each step against the source paper, while optional human checkpoints ensure the final product aligns perfectly with the author's vision, transforming the system from a mere tool into a powerful collaborative assistant. To rigorously validate our approach, we also construct $\textbf{PageBench}$, the first benchmark for this new task. Experiments show AutoPage not only generates high-quality, visually appealing pages but does so with remarkable efficiency in under 15 minutes for less than \$0.1. Code and dataset will be released at $\href{https://mqleet.github.io/AutoPage_ProjectPage/}{Webpage}$.
Launching in 2027 and 2029, respectively, Twinkle and Ariel will conduct the first large-scale homogeneous spectroscopic surveys of the atmospheres of hundreds of diverse exoplanets. This will fundamentally transition the field to an era of population-level characterisation. In this pilot study, we aim to explore possible synergies between Twinkle and Ariel to determine for instance whether prior Twinkle observations can substantially inform the target selection and observing strategy of Ariel. This study primarily aims to encourage further investigation by both consortium communities by showing what a potential scientific synergy would look like on a promising scientific case that requires further exploration. For this aim, we select a small subset of "cool" planets that are also particularly well-suited to be observed by Twinkle and therefore Ariel. By using representative noise estimates for both missions, we compute the number of visits required for an observation. Then, we simulate and retrieve transmission spectra of each target, assuming gaseous, H2/He-dominated atmospheres and various atmospheric models. For all candidates, we find that atmospheric parameters are generally retrieved well within 1-sigma to input values, with Ariel typically achieving tighter constraints. We demonstrate that for a small subset of cool gaseous planets, exploitable synergies exist between Twinkle and Ariel observations and Twinkle may very well provide a vantage point to plan Ariel observations. The true extent of the potential synergies, far beyond our considered sample, will be determined by the final target lists. Once Twinkle is operational and its performance is known, it could reliably inform Ariel's target prioritization and Ariel's capabilities which are already well-established can help define optimal targets and observational approaches for Twinkle.