The effectiveness of collision-free trajectory planners depends on the quality and diversity of training data, especially for rare scenarios. A widely used approach to improve dataset diversity involves generating realistic synthetic traffic scenarios. However, producing such scenarios remains difficult due to the precision required when scripting them manually or generating them in a single pass. Natural language offers a flexible way to describe scenarios, but existing text-to-simulation pipelines often rely on static snippet retrieval, limited grammar, single-pass decoding, or lack robust executability checks. Moreover, they depend heavily on constrained LLM prompting with minimal post-processing. To address these limitations, we introduce ARISE - Adaptive Refinement and Iterative Scenario Engineering, a multi-stage tool that converts natural language prompts into executable Scenic scripts through iterative LLM-guided refinement. After each generation, ARISE tests script executability in simulation software, feeding structured diagnostics back to the LLM until both syntactic and functional requirements are met. This process significantly reduces the need for manual intervention. Through extensive evaluation, ARISE outperforms the baseline in generating semantically accurate and executable traffic scenarios with greater reliability and robustness.

In this work, we propose a new spatio-directional neural encoding that is compact and efficient, and supports all-frequency signals in both space and direction. Current learnable encodings focus on Cartesian orthonormal spaces, which have been shown to be useful for representing high-frequency signals in the spatial domain. However, directly applying these encodings in the directional domain results in distortions, singularities, and discontinuities. As a result, most related works have used more traditional encodings for the directional domain, which lack the expressivity of learnable neural encodings. We address this by proposing a new angular encoding that generalizes the hash-grid approach from proach from Müller et al. [2022] to the directional domain by encoding directions using a hierarchical geodesic grid. Each vertex in the geodesic grid stores a learnable latent parameter, which is used to feed a neural network. Armed with this directional encoding, we propose a five-dimensional encoding for spatio-directional signals. We demonstrate that both encodings significantly outperform other hash-based alternatives. We apply our five-dimensional encoding in the context of neural path guiding, outperforming the state of the art by up to a factor of 2 in terms of variance reduction for the same number of samples.

With the growing emphasis on multilingual and cultural evaluation benchmarks for large language models, language and culture are often treated as synonymous, and performance is commonly used as a proxy for a models understanding of a given language. In this work, we argue that such evaluations overlook meaningful cultural variation that exists within a single language. We address this gap by focusing on narratives from different regions of Ethiopia and demonstrate that, despite shared linguistic characteristics, region-specific and domain-specific content substantially influences language evaluation outcomes. To this end, we introduce AmharicStoryQA, a long-sequence story question answering benchmark grounded in culturally diverse narratives from Amharic-speaking regions. Using this benchmark, we reveal a significant narrative understanding gap in existing LLMs, highlight pronounced regional differences in evaluation results, and show that supervised fine-tuning yields uneven improvements across regions and evaluation settings. Our findings emphasize the need for culturally grounded benchmarks that go beyond language-level evaluation to more accurately assess and improve narrative understanding in low-resource languages.

The object-based nature of human visual attention is well-known in cognitive science, but has only played a minor role in computational visual attention models so far. This is mainly due to a lack of suitable datasets and evaluation metrics for object-based attention. To address these limitations, we present – a novel 120-participant dataset of spatial street-crossing navigation in virtual reality specifically geared to object-based attention evaluations. The uniqueness of the presented dataset lies in the ethical and safety affiliated challenges that make collecting comparable data in real-world environments highly difficult. not only features accurate gaze data and a complete state-space representation of objects in the virtual environment, but it also offers variable scenario complexities and rich annotations, including panoptic segmentation, depth information, and vehicle keypoints. We further propose object-based similarity (oSIM) as a novel metric to evaluate the performance of object-based visual attention models, a previously unexplored performance characteristic. Our evaluations show that explicitly optimising for object-based attention not only improves oSIM performance but also leads to an improved model performance on common metrics. In addition, we present SUMGraph, a Mamba U-Net-based model, which explicitly encodes critical scene objects (vehicles) in a graph representation, leading to further performance improvements over several state-of-the-art visual attention prediction methods. The dataset, code and models will be publicly released.

Synthesizing text-driven 3D human motion within realistic scenes requires learning both semantic intent (“walk to the couch”) and physical feasibility (e.g., avoiding collisions). Current methods use generative frameworks that simultaneously learn high-level planning and low-level contact reasoning, and rely on computationally expensive 3D scene data such as point clouds or voxel occupancy grids. We propose SceMoS, a scene-aware motion synthesis framework that shows that structured 2D scene representations can serve as a powerful alternative to full 3D supervision in physically grounded motion synthesis. SceMoS disentangles global planning from local execution using lightweight 2D cues and relying on (1) a text-conditioned autoregressive global motion planner that operates on a top-down bird’s-eye-view (BEV) image of the scene, encoded with DINOv2 features, as the scene representation, and (2) geometry-grounded motion tokenizer trained via a conditional VQ-VAE, that uses 2D local scene heightmap, thus embedding surface physics directly into a discrete vocabulary. This 2D factorization reaches an efficiency-fidelity trade-off: BEV semantics capture spatial layout and affordance for global reasoning, while local heightmaps enforce fine-grained physical adherence without full 3D volumetric reasoning. SceMoS achieves state-of-the-art motion realism and contact accuracy on the TRUMANS benchmark, reducing the number of trainable parameters for scene encoding by over 50%, showing that 2D scene cues can effectively ground 3D human-scene interaction.

Large language models (LLMs) have demonstrated significant capabilities in solving mathematical problems expressed in natural language. However, multilingual and culturally-grounded mathematical reasoning in low-resource languages lags behind English due to the scarcity of socio-cultural task datasets that reflect accurate native entities such as person names, organization names, and currencies. Existing multilingual benchmarks are predominantly produced via translation and typically retain English-centric entities, owing to the high cost associated with human annotater-based localization. Moreover, automated localization tools are limited, and hence, truly localized datasets remain scarce. To bridge this gap, we introduce a framework for LLM-driven cultural localization of math word problems that automatically constructs datasets with native names, organizations, and currencies from existing sources. We find that translated benchmarks can obscure true multilingual math ability under appropriate socio-cultural contexts. Through extensive experiments, we also show that our framework can help mitigate English-centric entity bias and improves robustness when native entities are introduced across various languages.

Video depth estimation is crucial in various applications, such as scene reconstruction and augmented reality. In contrast to the naive method of estimating depths from images, a more sophisticated approach uses temporal information, thereby eliminating flickering and geometrical inconsistencies. We propose a consistent method for dense video depth estimation; however, unlike the existing monocular methods, ours relates to stereo videos. This technique overcomes the limitations arising from the monocular input. As a benefit of using stereo inputs, a left-right consistency loss is introduced to improve the performance. Besides, we use SLAM-based camera pose estimation in the process. To address the problem of depth blurriness during test-time training (TTT), we present an edge-preserving loss function that improves the visibility of fine details while preserving geometrical consistency. We show that our edge-aware stereo video model can accurately estimate the dense depth maps.

Animation of virtual characters in a crowd is one of the unavoidable hurdles artists and developers encounter throughout their work in different fields, such as cinematography, simulations, and game development. Depending on the requirements for the quality of motions, various techniques are used to overcome this constantly rising problem, each with its own shortcomings. Ignoring the interaction between individuals by using simple hitboxes to avoid collisions produces unrealistic movement patterns, while the manual animation of characters is a tedious and costly endeavor. This work discusses the implementation of an automated system capable of realistic human motion synthesis for multiple avatars, focusing on the interaction between individuals. The method is based on an autoregressive, data-driven conditional variational autoencoder and an additional control policy providing the latent input vector for the motion network trained with reinforcement learning. Different virtual sensor types that perceive a variable number of individuals and obstacles around the agent are proposed and evaluated. The system provides the simulation of the body movements in a multi-agent environment. It can operate at runtime with an adequate frame rate in environments containing more than 20 characters.

Inverse rendering has emerged as a standard tool to reconstruct the parameters of appearance models from images (e.g., textured BSDFs). In this work, we present several novel contributions motivated by the practical challenges of recovering high-resolution surface appearance textures, including spatially-varying subsurface scattering parameters.First, we propose Laplacian mipmapping, which combines differentiable mipmapping and a Laplacian pyramid representation into an effective preconditioner. This seemingly simple technique significantly improves the quality of recovered surface textures on a set of challenging inverse rendering problems. Our method automatically adapts to the render and texture resolutions, only incurs moderate computational cost and achieves better quality than prior work while using fewer hyperparameters. Second, we introduce a specialized gradient computation algorithm for textured, path-traced subsurface scattering, which facilitates faithful reconstruction of translucent materials. By using path tracing, we enable the recovery of complex appearance while avoiding the approximations of the previously used diffusion dipole methods. Third, we demonstrate the application of both these techniques to reconstructing the textured appearance of human faces from sparse captures. Our method recovers high-quality relightable appearance parameters that are compatible with current production renderers.

Large Language Models (LLMs) are increasingly employed in high-stakes decision-making tasks, such as loan approvals. While their applications expand across domains, LLMs struggle to process tabular data, ensuring fairness and delivering reliable predictions. In this work, we assess the performance and fairness of LLMs on serialized loan approval datasets from three geographically distinct regions: Ghana, Germany, and the United States. Our evaluation focuses on the model’s zero-shot and in-context learning (ICL) capabilities. Our results reveal that the choice of serialization format significantly affects both performance and fairness in LLMs, with certain formats such as GReaT and LIFT yielding higher F1 scores but exacerbating fairness disparities. Notably, while ICL improved model performance by 4.9-59.6% relative to zero-shot baselines, its effect on fairness varied considerably across datasets. Our work underscores the importance of effective tabular data representation methods and fairness-aware models to improve the reliability of LLMs in financial decision-making.

Image processing workflows typically consist of a series of different functions, each working with “image” inputs and outputs in an abstract sense. However, the specific in-memory representation of images differs between and sometimes within libraries. Conversion is therefore necessary when integrating functions from several sources into a single program. The conversion process forces users to consider low-level implementation details, including data types, color channels, channel order, minibatch layout, memory locations, and pixel intensity ranges. Specifically in the case of visual programming languages (VPLs), this distracts from high-level operations. We introduce im2im, a Python library that automates the conversion of in-memory image representations. The central concept of this library is a knowledge graph that describes image representations and how to convert between them. The system queries this knowledge graph to generate the conversion code and execute it, converting an image to the desired representation. The effectiveness of the approach is evaluated through two case studies in VPLs. In each case, we compared a workflow created in a basic block-based VPL with the same workflow enhanced using im2im. These evaluations show that im2im automates type conversions and eliminates the need for manual intervention. Additionally, we compared the overhead of using explicit intermediate representations versus im2im, both of which avoid manual type conversions in VPLs. The results indicate that im2im generates only the necessary conversions, avoiding the runtime overhead associated with converting to and from intermediate formats. A performance comparison between the step-by-step approach used by im2im and a single-function approach demonstrates that the overhead introduced by im2im does not impact practical usability. While focused on block-based VPLs, im2im can be generalized to other VPLs and textual programming environments. Its principles are also applicable to domains other than images. The source code and analyses are available via GitHub.

We present DuetGen, a novel framework for generating interactive two-person dances from music. The key challenge of this task lies in the inherent complexities of two-person dance interactions, where the partners need to synchronize both with each other and with the music. Inspired by the recent advances in motion synthesis, we propose a two-stage solution: encoding two-person motions into discrete tokens and then generating these tokens from music. To effectively capture intricate interactions, we represent both dancers' motions as a unified whole to learn the necessary motion tokens, and adopt a coarse-to-fine learning strategy in both the stages. Our first stage utilizes a VQ-VAE that hierarchically separates high-level semantic features at a coarse temporal resolution from low-level details at a finer resolution, producing two discrete token sequences at different abstraction levels. Subsequently, in the second stage, two generative masked transformers learn to map music signals to these dance tokens: the first producing high-level semantic tokens, and the second, conditioned on music and these semantic tokens, producing the low-level tokens. We train both transformers to learn to predict randomly masked tokens within the sequence, enabling them to iteratively generate motion tokens by filling an empty token sequence during inference. Through the hierarchical masked modeling and dedicated interaction representation, DuetGen achieves the generation of synchronized and interactive two-person dances across various genres. Extensive experiments and user studies on a benchmark duet dance dataset demonstrate state-of-the-art performance of DuetGen in motion realism, music-dance alignment, and partner coordination. Code and model weights are available at https://github.com/anindita127/DuetGen.

This white paper discusses and explores the various points of intersection between quantum computing and artificial intelligence (AI). It describes how quantum computing could support the development of innovative AI solutions. It also examines use cases of classical AI that can empower research and development in quantum technologies, with a focus on quantum computing and quantum sensing. The purpose of this white paper is to provide a long-term research agenda aimed at addressing foundational questions about how AI and quantum computing interact and benefit one another. It concludes with a set of recommendations and challenges, including how to orchestrate the proposed theoretical work, align quantum AI developments with quantum hardware roadmaps, estimate both classical and quantum resources - especially with the goal of mitigating and optimizing energy consumption - advance this emerging hybrid software engineering discipline, and enhance European industrial competitiveness while considering societal implications.

Multiple importance sampling (MIS) is a vital component of most rendering algorithms. MIS computes a weighted sum of samples from many different techniques to achieve generalization, that is, to handle a wide range of scene types and lighting effects. A key factor to the performance of MIS is the choice of weighting function. The go-to default-the balance heuristic-performs well in many cases, but prior work has shown that it can yield unsatisfactory results. A number of challenges cause this suboptimal performance, including low-variance techniques, sample correlation, and unknown sampling densities. Prior work has suggested improvements for some of these problems, but a general optimal solution has yet to be found. We propose a general and practical weight correction scheme: We optimize, on-the-fly, a set of correction factors that are multiplied into any baseline MIS heuristic (e.g., balance or power). We demonstrate that this approach yields consistently better equal-time performance on two rendering applications: bidirectional algorithms and resampled importance sampling for direct illumination.

Large language models have shown great success on natural language tasks in recent years, but they have also shown great promise when adapted to new modalities, e.g., for scientific machine learning tasks. Even though decoder-only models are more popular within NLP and scale exceedingly well at generating natural language, most proposed approaches for cross-modal adaptation focus on encoder-only models, raising the question of how model architecture affects these approaches. In this paper, we therefore perform a series of ablation studies to answer this question, systematically comparing encoder-only and decoder-only models on cross-modal adaptation for time-dependent simulation tasks based on partial differential equations (PDEs). We find that decoder-only models are far worse than encoder-only models, when existing approaches are applied unmodified. In contrast to several other domains, scaling decoder-only models also does not help. To harness the potential of decoder-only models in this context, we introduce two novel approaches, Parallel Flipping and Sequence Doubling, attempting to mimic bidirectionality in autoregressive models. Both our methods improve overall performance using decoder-only models for all tasks and all cross-model adaptation methods, closing the gap to encoder-only model performance. We hope that our findings broaden the spectrum of models used on cross-modal adaptation tasks to further scientific ML.

Network latency presents a critical challenge for interactive multi-user virtual reality (VR) applications, particularly when transmitting the users' full-body motion data to remote collaborators. Such latency can significantly degrade the quality of interaction, leading to a less immersive and less effective user experience. Pose forecasting, which involves predicting future motion sequences based on observed past motions, offers a potential solution. However, it has not yet been widely adopted for compensating latency for full-body motion transmission. In this work, we extend and evaluate two state-of-the-art neural network architectures for predicting future animation sequences, enabling client-side compensation of network latency. To support our methods, we construct a custom motion capture dataset with VR-specific full-body movements, along with an animation-specific data encoding that integrates both joint position and rotation data. Our experimental results show that predictive latency compensation can reduce the positional error of the reconstructed motion by a factor of up to.2.5 and is robust to noisy network connections containing jitter and packet loss. Despite these promising results, we identify several open challenges and outline directions for future work necessary for successful deployment of predictive compensation in real-world VR systems.