Qiao Wang

Aditi, Niket Agarwal, Arslan Ali et al.

We introduce Cosmos 3, a family of omnimodal world models designed to jointly process and generate language, image, video, audio, and action sequences within a unified mixture-of-transformers architecture. By supporting highly flexible input-output configurations, Cosmos 3 seamlessly unifies critical modalities for Physical AI -- effectively subsuming vision-language models, video generators, world simulators, and world-action models into a single framework. Our evaluation demonstrates that Cosmos 3 establishes a new state-of-the-art across a diverse suite of understanding and generation tasks, demonstrating omnimodal world models as scalable, general-purpose backbones for embodied agents. Our post-trained Cosmos 3 models were ranked as the best open-source Text-to-Image and Image-to-Video models by Artificial Analysis, and the best policy model by RoboArena at the time the technical report was written. To accelerate open research and deployment in Physical AI, we make our code, model checkpoints, curated synthetic datasets, and evaluation benchmark available under the Linux Foundation's OpenMDW-1.1 https://openmdw.ai/license/1-1/ License at https://github.com/nvidia/cosmos}{github.com/nvidia/cosmos and https://huggingface.co/collections/nvidia/cosmos3 . The project website is available at https://research.nvidia.com/labs/cosmos-lab/cosmos3 .

Qiao Wang

We uncover a finite-temperature extension of de Bruijn's identity -- the classical relation $\frac{d}{dt}h(X+\sqrt{t}Z)=\frac{1}{2}J(X)$ connecting differential entropy and Fisher information. Our framework is the spectral theory of Blahut--Arimoto (BA) dynamics, recently developed by Wang~\cite{Wang2026} for the analysis of rate-distortion optimization. The central observation is elementary yet profound: for Gaussian sources, the spectral gap $\lam$ of the BA relaxation kernel $\G$ satisfies $\lam = 1/(2βσ^2)$~\cite{Wang2026}, while the Fisher information of the source is $J = 1/σ^2$. Hence \[ {\lam = \frac{J}{2β}} \] for all inverse temperatures $β> 1/(2σ^2)$. This identifies the BA spectral gap as a \emph{finite-temperature regularization of Fisher information}. From this observation we derive an exact finite-temperature de Bruijn identity: \[ \frac{\partial F_β}{\partial σ^2} = \frac{1}{2βσ^2} = \lam, \] where $F_β$ is the BA free energy. This identity holds for all finite $β$ without any limit procedure. The classical de Bruijn identity follows as the exact consequence $β\,\partial F_β/\partialσ^2 = J/2$. The significance is structural: classical de Bruijn is not an isolated fact about Gaussian convolutions, but the $β\to\infty$ shadow of a one-parameter family of exact identities living in the spectral geometry of rate-distortion optimization. We discuss implications for the entropy power inequality, the $χ^2$-dissipation structure of BA dynamics, and the geometric unification of information inequalities.

Seth Karten, Jake Grigsby, Tersoo Upaa et al.

We present the PokeAgent Challenge, a large-scale benchmark for decision-making research built on Pokemon's multi-agent battle system and expansive role-playing game (RPG) environment. Partial observability, game-theoretic reasoning, and long-horizon planning remain open problems for frontier AI, yet few benchmarks stress all three simultaneously under realistic conditions. PokeAgent targets these limitations at scale through two complementary tracks: our Battling Track, which calls for strategic reasoning and generalization under partial observability in competitive Pokemon battles, and our Speedrunning Track, which requires long-horizon planning and sequential decision-making in the Pokemon RPG. Our Battling Track supplies a dataset of 20M+ battle trajectories alongside a suite of heuristic, RL, and LLM-based baselines capable of high-level competitive play. Our Speedrunning Track provides the first standardized evaluation framework for RPG speedrunning, including an open-source multi-agent orchestration system for modular, reproducible comparisons of harness-based LLM approaches. Our NeurIPS 2025 competition validates both the quality of our resources and the research community's interest in Pokemon, with over 100 teams competing across both tracks and winning solutions detailed in our paper. Participant submissions and our baselines reveal considerable gaps between generalist (LLM), specialist (RL), and elite human performance. Analysis against the BenchPress evaluation matrix shows that Pokemon battling is nearly orthogonal to standard LLM benchmarks, measuring capabilities not captured by existing suites and positioning Pokemon as an unsolved benchmark that can drive RL and LLM research forward. We transition to a living benchmark with a live leaderboard for Battling and self-contained evaluation for Speedrunning at https://pokeagentchallenge.com.

Jun Saito, Jiefeng Li, Michael de Ruyter et al.

Parametric human body models are foundational to human reconstruction, animation, and simulation, yet they remain mutually incompatible: SMPL, SMPL-X, MHR, Anny, and related models each diverge in mesh topology, skeletal structure, shape parameterization, and unit convention, making it impractical to exploit their complementary strengths within a single pipeline. We present SOMA, a unified body layer that bridges these heterogeneous representations through three abstraction layers. Mesh topology abstraction maps any source model's identity to a shared canonical mesh in constant time per vertex. Skeletal abstraction recovers a full set of identity-adapted joint transforms from any body shape, whether in rest pose or an arbitrary posed configuration, in a single closed-form pass, with no iterative optimization or per-model training. Pose abstraction inverts the skinning pipeline to recover unified skeleton rotations directly from posed vertices of any supported model, enabling heterogeneous motion datasets to be consumed without custom retargeting. Together, these layers reduce the $O(M^2)$ per-pair adapter problem to $O(M)$ single-backend connectors, letting practitioners freely mix identity sources and pose data at inference time. The entire pipeline is fully differentiable end-to-end and GPU-accelerated via NVIDIA-Warp.

Haoran Shen, Chen Zeng, Jiahui Wang et al.

It has been observed that even a small amount of noise introduced into the dataset can significantly degrade the performance of KAN. In this brief note, we aim to quantitatively evaluate the performance when noise is added to the dataset. We propose an oversampling technique combined with denoising to alleviate the impact of noise. Specifically, we employ kernel filtering based on diffusion maps for pre-filtering the noisy data for training KAN network. Our experiments show that while adding i.i.d. noise with any fixed SNR, when we increase the amount of training data by a factor of $r$, the test-loss (RMSE) of KANs will exhibit a performance trend like $\text{test-loss} \sim \mathcal{O}(r^{-\frac{1}{2}})$ as $r\to +\infty$. We conclude that applying both oversampling and filtering strategies can reduce the detrimental effects of noise. Nevertheless, determining the optimal variance for the kernel filtering process is challenging, and enhancing the volume of training data substantially increases the associated costs, because the training dataset needs to be expanded multiple times in comparison to the initial clean data. As a result, the noise present in the data ultimately diminishes the effectiveness of Kolmogorov-Arnold networks.

Rui Cao, Qiao Wang

This research examines the use of Large Language Models (LLMs) in predicting time series, with a specific focus on the LLMTIME model. Despite the established effectiveness of LLMs in tasks such as text generation, language translation, and sentiment analysis, this study highlights the key challenges that large language models encounter in the context of time series prediction. We assess the performance of LLMTIME across multiple datasets and introduce classical almost periodic functions as time series to gauge its effectiveness. The empirical results indicate that while large language models can perform well in zero-shot forecasting for certain datasets, their predictive accuracy diminishes notably when confronted with diverse time series data and traditional signals. The primary finding of this study is that the predictive capacity of LLMTIME, similar to other LLMs, significantly deteriorates when dealing with time series data that contain both periodic and trend components, as well as when the signal comprises complex frequency components.

Xiang Zhang, Qiao Wang

We consider the problem of spectral clustering under group fairness constraints, where samples from each sensitive group are approximately proportionally represented in each cluster. Traditional fair spectral clustering (FSC) methods consist of two consecutive stages, i.e., performing fair spectral embedding on a given graph and conducting $k$means to obtain discrete cluster labels. However, in practice, the graph is usually unknown, and we need to construct the underlying graph from potentially noisy data, the quality of which inevitably affects subsequent fair clustering performance. Furthermore, performing FSC through separate steps breaks the connections among these steps, leading to suboptimal results. To this end, we first theoretically analyze the effect of the constructed graph on FSC. Motivated by the analysis, we propose a novel graph construction method with a node-adaptive graph filter to learn graphs from noisy data. Then, all independent stages of conventional FSC are integrated into a single objective function, forming an end-to-end framework that inputs raw data and outputs discrete cluster labels. An algorithm is developed to jointly and alternately update the variables in each stage. Finally, we conduct extensive experiments on synthetic, benchmark, and real data, which show that our model is superior to state-of-the-art fair clustering methods.

Qiao Wang, Liang Liu, Mitsuo Yoshida

Social video platforms shape how people access information, while recommendation systems can narrow exposure and increase the risk of toxic interaction. Previous research has often examined text or users in isolation, overlooking the structural context in which such toxic interactions occur. Without considering who interacts with whom and around what content, it is difficult to explain why negative expressions cluster within particular communities. To address this issue, this study focuses on the Chinese social video platform Bilibili, incorporating video-level information as the environment for user expression, modeling users and videos in an interaction matrix. After normalization and dimensionality reduction, we perform separate clustering on both sides of the video-user interaction matrix with K-means. Cluster assignments facilitate comparisons of user behavior, including message length, posting frequency, and source (barrage and comment), as well as textual features such as sentiment and toxicity, and video attributes defined by uploaders. Such a clustering approach integrates structural ties with content signals to identify stable groups of videos and users. We find clear stratification in interaction style (message length, comment ratio) across user clusters, while sentiment and toxicity differences are weak or inconsistent across video clusters. Across video clusters, viewing volume exhibits a clear hierarchy, with higher exposure groups concentrating more toxic expressions. For such a group, platforms should require timely intervention during periods of rapid growth. Across user clusters, comment ratio and message length form distinct hierarchies, and several clusters with longer and comment-oriented messages exhibit lower toxicity. For such groups, platforms should strengthen mechanisms that sustain rational dialogue and encourage engagement across topics.

Ruizhe Wang, Chunliang Hua, Tomakayev Shingys et al.

The photorealistic reconstruction and rendering of architectural scenes have extensive applications in industries such as film, games, and transportation. It also plays an important role in urban planning, architectural design, and the city's promotion, especially in protecting historical and cultural relics. The 3D Gaussian Splatting, due to better performance over NeRF, has become a mainstream technology in 3D reconstruction. Its only input is a set of images but it relies heavily on geometric parameters computed by the SfM process. At the same time, there is an existing abundance of raw 3D models, that could inform the structural perception of certain buildings but cannot be applied. In this paper, we propose a straightforward method to harness these raw 3D models to guide 3D Gaussians in capturing the basic shape of the building and improve the visual quality of textures and details when photos are captured non-systematically. This exploration opens up new possibilities for improving the effectiveness of 3D reconstruction techniques in the field of architectural design.

Xiang Zhang, Qiao Wang

We consider the problem of inferring graph topology from smooth graph signals in a novel but practical scenario where data are located in distributed clients and prohibited from leaving local clients due to factors such as privacy concerns. The main difficulty in this task is how to exploit the potentially heterogeneous data of all clients under data silos. To this end, we first propose an auto-weighted multiple graph learning model to jointly learn a personalized graph for each local client and a single consensus graph for all clients. The personalized graphs match local data distributions, thereby mitigating data heterogeneity, while the consensus graph captures the global information. Moreover, the model can automatically assign appropriate contribution weights to local graphs based on their similarity to the consensus graph. We next devise a tailored algorithm to solve the induced problem, where all raw data are processed locally without leaving clients. Theoretically, we establish a provable estimation error bound and convergence analysis for the proposed model and algorithm. Finally, extensive experiments on synthetic and real data are carried out, and the results illustrate that our approach can learn graphs effectively in the target scenario.

Chen Zeng, Jiahui Wang, Haoran Shen et al.

In this paper, we compare the performance of Kolmogorov-Arnold Networks (KAN) and Multi-Layer Perceptron (MLP) networks on irregular or noisy functions. We control the number of parameters and the size of the training samples to ensure a fair comparison. For clarity, we categorize the functions into six types: regular functions, continuous functions with local non-differentiable points, functions with jump discontinuities, functions with singularities, functions with coherent oscillations, and noisy functions. Our experimental results indicate that KAN does not always perform best. For some types of functions, MLP outperforms or performs comparably to KAN. Furthermore, increasing the size of training samples can improve performance to some extent. When noise is added to functions, the irregular features are often obscured by the noise, making it challenging for both MLP and KAN to extract these features effectively. We hope these experiments provide valuable insights for future neural network research and encourage further investigations to overcome these challenges.

Adnan Labib, Qiao Wang, Yixuan Huang et al.

Automated assistants for Grammatical Error Correction are now embedded in educational platforms serving millions of learners, yet three critical gaps remain in this domain: (1) latest-generation Large Language Models (LLMs) lack comprehensive evaluation on grammar correction tasks; (2) whether combining these LLMs improves correction quality is unexplored; and (3) the extent to which reference-based metrics underestimate GEC system performance has not been adequately quantified. In this study, first, we evaluate latest-generation LLMs on edit precision, fluency preservation, and meaning retention, showing fine-tuned GPT-4o achieves state-of-the-art performance across all three dimensions. Second, through grammatical error type analysis we demonstrate that individual LLMs exhibit highly similar error correction patterns ($ρ=0.947$). Third, we show that reference-based metrics underestimate GEC performance with 73.76% of GPT-4o corrections different from gold standards being equally valid or even superior. These GEC evaluation findings equip educators with guidance for selecting GEC assistants that enhance rather than constrain student linguistic development. We make our data, code, and models publicly available.

Ahmed Karim, Qiao Wang, Zheng Yuan

Automated Essay Scoring (AES) systems now reach near human agreement on some public benchmarks, yet real-world adoption, especially in high-stakes examinations, remains limited. A principal obstacle is that most models output a single score without any accompanying measure of confidence or explanation. We address this gap with conformal prediction, a distribution-free wrapper that equips any classifier with set-valued outputs and formal coverage guarantees. Two open-source large language models (Llama-3 8B and Qwen-2.5 3B) are fine-tuned on three diverse corpora (ASAP, TOEFL11, Cambridge-FCE) and calibrated at a 90 percent risk level. Reliability is assessed with UAcc, an uncertainty-aware accuracy that rewards models for being both correct and concise. To our knowledge, this is the first work to combine conformal prediction and UAcc for essay scoring. The calibrated models consistently meet the coverage target while keeping prediction sets compact, indicating that open-source, mid-sized LLMs can already support teacher-in-the-loop AES; we discuss scaling and broader user studies as future work.

Xinxin Zhuo, Mengyuan Niu, Ruizhe Wang et al.

Micro-scale street-level economic assessment is fundamental for precision spatial resource allocation. While Street View Imagery (SVI) advances urban sensing, existing approaches remain semantically superficial and overlook brand hierarchy heterogeneity and structural recession. To address this, we propose a visual-semantic and field-based spatiotemporal framework, operationalized via the Street Economic Vitality Index (SEVI). Our approach integrates physical and semantic streetscape parsing through instance segmentation of signboards, glass interfaces, and storefront closures. A dual-stage VLM-LLM pipeline standardizes signage into global hierarchies to quantify a spatially smoothed brand premium index. To overcome static SVI limitations, we introduce a temporal lag design using Location-Based Services (LBS) data to capture realized demand. Combined with a category-weighted Gaussian spillover model, we construct a three-dimensional diagnostic system covering Commercial Activity, Spatial Utilization, and Physical Environment. Experiments based on time-lagged geographically weighted regression across eight tidal periods in Nanjing reveal quasi-causal spatiotemporal heterogeneity. Street vibrancy arises from interactions between hierarchical brand clustering and mall-induced externalities. High-quality interfaces show peak attraction during midday and evening, while structural recession produces a lagged nighttime repulsion effect. The framework offers evidence-based support for precision spatial governance.

Jianbo Qi, Mengyao Li, Baogui Jiang et al.

Satellite Earth observation has accumulated massive spatiotemporal archives essential for monitoring environmental change, yet these remain organized as discrete raster files, making them costly to store, transmit, and query. We present GeoNDC, a queryable neural data cube that encodes planetary-scale Earth observation data as a continuous spatiotemporal implicit neural field, enabling on-demand queries and continuous-time reconstruction without full decompression. Experiments on a 20-year global MODIS MCD43A4 reflectance record (7 bands, 5\,km, 8-day sampling) show that the learned representation supports direct spatiotemporal queries on consumer hardware. On Sentinel-2 imagery (10\,m), continuous temporal parameterization recovers cloud-free dynamics with high fidelity ($R^2 > 0.85$) under simulated 2-km cloud occlusion. On HiGLASS biophysical products (LAI and FPAR), GeoNDC attains near-perfect accuracy ($R^2 > 0.98$). The representation compresses the 20-year MODIS archive to 0.44\,GB -- approximately 95:1 relative to an optimized Int16 baseline -- with high spectral fidelity (mean $R^2 > 0.98$, mean RMSE $= 0.021$). These results suggest GeoNDC offers a unified AI-native representation for planetary-scale Earth observation, complementing raw archives with a compact, analysis-ready data layer integrating query, reconstruction, and compression in a single framework.

Qiao Wang

This paper uncovers an exact $χ^2$ dissipation identity for the Blahut--Arimoto (BA) flow and establishes its fundamental information-geometric structure. While prior works have analyzed BA convergence via the monotone decrease of the Kullback--Leibler divergence, we establish that the continuous-time flow $\partial_τq = \tilde{q}_β[q] - q$ is fundamentally governed by the Pearson $χ^2$ divergence. Specifically, we prove the governing identity $\frac{d}{dτ} Φ_β= -χ^2(\tilde{q}_β\| q)$, elevating $χ^2$ from a mere local approximation to the global non-asymptotic driver of the dynamics. We establish a \emph{unified dissipation identity} revealing that the $χ^2$ and the symmetrised Jeffreys divergence $D_J$ differ only by a cubic residual $\mathcal{N}$, which vanishes \emph{exactly} for the BA flow. This algebraic cancellation identifies the Jacobian $K_q = D\tilde{q}_β[q]$ and its complement $\mathcal{A}_q = I - K_q$ as the Fisher--Rao Hessian of the free energy $Φ_β$. The spectral gap of $\mathcal{A}_{q^*}$ is proved to determine both the local exponential convergence rate (a universal value of $2$ in artificial time) and the stability of the attraction basin. For Gaussian sources, the theory yields three concrete consequences: (i)~the reproduction variance obeys an exact closed ODE; (ii)~the Gaussian distribution emerges as the unique dynamical attractor; and (iii)~in high dimensions, convergence is bottlenecked by the maximum source variance -- \emph{spectral stiffness}. Significantly, this framework provides new dissipative interpretations for fundamental problems including MIMO channel capacity, Wyner--Ziv coding, and the information bottleneck. These results establish the BA flow as a canonical dissipative relaxation and provide a geometric foundation for the dynamical theory of information-theoretic optimisation.

Chen Zeng, Jiahui Wang, Qiao Wang

Existing theory suggests that Kolmogorov-Arnold Networks (KANs) can overcome the spectral bias commonly observed in neural networks under the assumption that inputs are statistically independent. However, this assumption does not hold in time series forecasting (TSF), where inputs are lagged observations with strong temporal autocorrelation. Through theoretical analysis and empirical validation, we obtain an unexpected finding: temporal autocorrelation reintroduces spectral bias in KANs, and the bias becomes increasingly pronounced as the degree of autocorrelation increases. This suggests that standard KANs may face substantial difficulties in TSF with strongly autocorrelated inputs. To address this problem, we introduce the Discrete Cosine Transform (DCT) to reduce the correlations among the network inputs. As expected, experimental results reveal that DCT preprocessing substantially reduces the observed low-frequency preference in TSF. This result also corroborates that the spectral bias of KANs in TSF tasks is indeed induced by the autocorrelation among input variables.

Qiao Wang, Ralph Rose, Naho Orita et al.

A common way of assessing language learners' mastery of vocabulary is via multiple-choice cloze (i.e., fill-in-the-blank) questions. But the creation of test items can be laborious for individual teachers or in large-scale language programs. In this paper, we evaluate a new method for automatically generating these types of questions using large language models (LLM). The VocaTT (vocabulary teaching and training) engine is written in Python and comprises three basic steps: pre-processing target word lists, generating sentences and candidate word options using GPT, and finally selecting suitable word options. To test the efficiency of this system, 60 questions were generated targeting academic words. The generated items were reviewed by expert reviewers who judged the well-formedness of the sentences and word options, adding comments to items judged not well-formed. Results showed a 75% rate of well-formedness for sentences and 66.85% rate for suitable word options. This is a marked improvement over the generator used earlier in our research which did not take advantage of GPT's capabilities. Post-hoc qualitative analysis reveals several points for improvement in future work including cross-referencing part-of-speech tagging, better sentence validation, and improving GPT prompts.

Qiao Wang

This study examines the convexification version of the backward differential flow algorithm for the global minimization of polynomials, introduced by O. Arikan \textit{et al} in \cite{ABK}. It investigates why this approach might fail with high-degree polynomials yet succeeds with quartic polynomials. We employ the heat evolution method for convexification combined with Gaussian filtering, which acts as a cumulative form of Steklov's regularization. In this context, we apply the fingerprint theory from computer vision. Originally developed by A.L. Yuille and T. Poggio in the 1980s for computer vision, the fingerprint theory, particularly the fingerprint trajectory equation, is used to illustrate the scaling (temporal) evolution of minimizers. In the case of general polynomials, our research has led to the creation of the Yuille-Poggio flow and a broader interpretation of the fingerprint concepts, in particular we establish the condition both sufficient and necessary for the convexified backward differential flow algorithms to successfully achieve global minimization. For quartic polynomials, our analysis not only reflects the results of O. Arikan et al. \cite{ABK} but also presents a significantly simpler version of Newton's method that can always globally minimize quartic polynomials without convexification.

Niluthpol Chowdhury Mithun, Tuan Pham, Qiao Wang et al.

Recent advancements in 3D Gaussian Splatting (3DGS) and Neural Radiance Fields (NeRF) have achieved impressive results in real-time 3D reconstruction and novel view synthesis. However, these methods struggle in large-scale, unconstrained environments where sparse and uneven input coverage, transient occlusions, appearance variability, and inconsistent camera settings lead to degraded quality. We propose GS-Diff, a novel 3DGS framework guided by a multi-view diffusion model to address these limitations. By generating pseudo-observations conditioned on multi-view inputs, our method transforms under-constrained 3D reconstruction problems into well-posed ones, enabling robust optimization even with sparse data. GS-Diff further integrates several enhancements, including appearance embedding, monocular depth priors, dynamic object modeling, anisotropy regularization, and advanced rasterization techniques, to tackle geometric and photometric challenges in real-world settings. Experiments on four benchmarks demonstrate that GS-Diff consistently outperforms state-of-the-art baselines by significant margins.

Haolong Yan, Jia Wang, Xin Huang et al.

Recent advances in multimodal large language models unlock unprecedented opportunities for GUI automation. However, a fundamental challenge remains: how to efficiently acquire high-quality training data while maintaining annotation reliability? We introduce a self-evolving training pipeline powered by the Calibrated Step Reward System, which converts model-generated trajectories into reliable training signals through trajectory-level calibration, achieving >90% annotation accuracy with 10-100x lower cost. Leveraging this pipeline, we introduce Step-GUI, a family of models (4B/8B) that achieves state-of-the-art GUI performance (8B: 80.2% AndroidWorld, 48.5% OSWorld, 62.6% ScreenShot-Pro) while maintaining robust general capabilities. As GUI agent capabilities improve, practical deployment demands standardized interfaces across heterogeneous devices while protecting user privacy. To this end, we propose GUI-MCP, the first Model Context Protocol for GUI automation with hierarchical architecture that combines low-level atomic operations and high-level task delegation to local specialist models, enabling high-privacy execution where sensitive data stays on-device. Finally, to assess whether agents can handle authentic everyday usage, we introduce AndroidDaily, a benchmark grounded in real-world mobile usage patterns with 3146 static actions and 235 end-to-end tasks across high-frequency daily scenarios (8B: static 89.91%, end-to-end 52.50%). Our work advances the development of practical GUI agents and demonstrates strong potential for real-world deployment in everyday digital interactions.

Qiao Wang, Joseph George

Monitoring household water use in rapidly urbanizing regions is hampered by costly, time-intensive enumeration methods and surveys. We investigate whether publicly available imagery-satellite tiles, Google Street View (GSV) segmentation-and simple geospatial covariates (nightlight intensity, population density) can be utilized to predict household water consumption in Hubballi-Dharwad, India. We compare four approaches: survey features (benchmark), CNN embeddings (satellite, GSV, combined), and GSV semantic maps with auxiliary data. Under an ordinal classification framework, GSV segmentation plus remote-sensing covariates achieves 0.55 accuracy for water use, approaching survey-based models (0.59 accuracy). Error analysis shows high precision at extremes of the household water consumption distribution, but confusion among middle classes is due to overlapping visual proxies. We also compare and contrast our estimates for household water consumption to that of household subjective income. Our findings demonstrate that open-access imagery, coupled with minimal geospatial data, offers a promising alternative to obtaining reliable household water consumption estimates using surveys in urban analytics.

Qiao Wang, Adnan Labib, Robert Swier et al.

GenQuest is a generative text adventure game that leverages Large Language Models (LLMs) to facilitate second language learning through immersive, interactive storytelling. The system engages English as a Foreign Language (EFL) learners in a collaborative "choose-your-own-adventure" style narrative, dynamically generated in response to learner choices. Game mechanics such as branching decision points and story milestones are incorporated to maintain narrative coherence while allowing learner-driven plot development. Key pedagogical features include content generation tailored to each learner's proficiency level, and a vocabulary assistant that provides in-context explanations of learner-queried text strings, ranging from words and phrases to sentences. Findings from a pilot study with university EFL students in China indicate promising vocabulary gains and positive user perceptions. Also discussed are suggestions from participants regarding the narrative length and quality, and the request for multi-modal content such as illustrations.

Mengyuan Niu, Xinxin Zhuo, Ruizhe Wang et al.

Urban modeling is essential for city planning, scene synthesis, and gaming. Existing image-based methods generate diverse layouts but often lack geometric continuity and scalability, while graph-based methods capture structural relations yet overlook parcel semantics. We present a controllable framework for large-scale 3D vector urban layout generation, conditioned on both geometry and semantics. By fusing geometric and semantic attributes, introducing edge weights, and embedding building height in the graph, our method extends 2D layouts to realistic 3D structures. It also enables users to directly control the output by modifying semantic attributes. Experiments show that it produces valid, large-scale urban models, offering an effective tool for data-driven planning and design.

Chen Zeng, Tiehang Xu, Qiao Wang

Traditional neural networks struggle to capture the spectral structure of complex signals. Fourier neural networks (FNNs) attempt to address this by embedding Fourier series components, yet many real-world signals are almost-periodic with non-commensurate frequencies, posing additional challenges. Building on prior work showing that ARIMA outperforms large language models (LLMs) for forecasting, we extend the comparison to neural predictors and find ARIMA still superior. We therefore propose the Autoregressive-Weight-Enhanced Kolmogorov-Arnold Network (AR-KAN), which integrates a pre-trained AR module for temporal memory with a KAN for nonlinear representation. The AR module preserves essential temporal features while reducing redundancy. Experiments demonstrate that AR-KAN matches ARIMA on almost-periodic functions and achieves the best results on $72\%$ of Rdatasets series, with a clear advantage on data with periodic structure. These results highlight AR-KAN as a robust and effective framework for time series forecasting.

Ruizhe Wang, Junyan Yang, Qiao Wang

We introduce GeoTexBuild, a modular generative framework for creating 3D building models from footprints derived from site planning or map designs. The system is designed for architects and city planners, offering a seamless solution that directly converts map features into 3D buildings. The proposed framework employs a three-stage process comprising height map generation, geometry reconstruction, and appearance stylization, culminating in building models with detailed geometry and appearance attributes. By integrating customized ControlNet, Neural style field (NSF), and Multi-view diffusion model, we explore effective methods for controlling both geometric and visual attributes during the generation process. Our approach eliminates the problem of structural variations in a single facade image in existing 3D generation techniques for buildings. Experimental results at each stage validate the capability of GeoTexBuild to generate detailed and accurate building models from footprints.

Qiao Wang, Zheng Yuan

In this study, we evaluated the performance of the state-of-the-art sequence tagging grammar error detection and correction model (SeqTagger) using Japanese university students' writing samples. With an automatic annotation toolkit, ERRANT, we first evaluated SeqTagger's performance on error correction with human expert correction as the benchmark. Then a human-annotated approach was adopted to evaluate Seqtagger's performance in error detection using a subset of the writing dataset. Results indicated a precision of 63.66% and a recall of 20.19% for error correction in the full dataset. For the subset, after manual exclusion of irrelevant errors such as semantic and mechanical ones, the model shows an adjusted precision of 97.98% and an adjusted recall of 42.98% for error detection, indicating the model's high accuracy but also its conservativeness. Thematic analysis on errors undetected by the model revealed that determiners and articles, especially the latter, were predominant. Specifically, in terms of context-independent errors, the model occasionally overlooked basic ones and faced challenges with overly erroneous or complex structures. Meanwhile, context-dependent errors, notably those related to tense and noun number, as well as those possibly influenced by the students' first language (L1), remained particularly challenging.

Xiang Zhang, Qiao Wang

This paper endeavors to learn time-varying graphs by using structured temporal priors that assume underlying relations between arbitrary two graphs in the graph sequence. Different from many existing chain structure based methods in which the priors like temporal homogeneity can only describe the variations of two consecutive graphs, we propose a structure named \emph{temporal graph} to characterize the underlying real temporal relations. Under this framework, the chain structure is actually a special case of our temporal graph. We further proposed Alternating Direction Method of Multipliers (ADMM), a distributed algorithm, to solve the induced optimization problem. Numerical experiments demonstrate the superiorities of our method.

Ying Wang, Tingzhen Liu, Zepeng Bu et al.

In large-scale image retrieval, many indexing methods have been proposed to narrow down the searching scope of retrieval. The features extracted from images usually are of high dimensions or unfixed sizes due to the existence of key points. Most of existing index structures suffer from the dimension curse, the unfixed feature size and/or the loss of semantic similarity. In this paper a new classification-based indexing structure, called Semantic Indexing Structure (SIS), is proposed, in which we utilize the semantic categories rather than clustering centers to create database partitions, such that the proposed index SIS can be combined with feature extractors without the restriction of dimensions. Besides, it is observed that the size of each semantic partition is positively correlated with the semantic distribution of database. Along this way, we found that when the partition number is normalized to five, the proposed algorithm performed very well in all the tests. Compared with state-of-the-art models, SIS achieves outstanding performance.

Xiang Zhang, Yinfei Xu, Qinghe Liu et al.

Graphs are playing a crucial role in different fields since they are powerful tools to unveil intrinsic relationships among signals. In many scenarios, an accurate graph structure representing signals is not available at all and that motivates people to learn a reliable graph structure directly from observed signals. However, in real life, it is inevitable that there exists uncertainty in the observed signals due to noise measurements or limited observability, which causes a reduction in reliability of the learned graph. To this end, we propose a graph learning framework using Wasserstein distributionally robust optimization (WDRO) which handles uncertainty in data by defining an uncertainty set on distributions of the observed data. Specifically, two models are developed, one of which assumes all distributions in uncertainty set are Gaussian distributions and the other one has no prior distributional assumption. Instead of using interior point method directly, we propose two algorithms to solve the corresponding models and show that our algorithms are more time-saving. In addition, we also reformulate both two models into Semi-Definite Programming (SDP), and illustrate that they are intractable in the scenario of large-scale graph. Experiments on both synthetic and real world data are carried out to validate the proposed framework, which show that our scheme can learn a reliable graph in the context of uncertainty.

Zhicheng Liu, Fabio Miranda, Weiting Xiong et al.

Predicting commuting flows based on infrastructure and land-use information is critical for urban planning and public policy development. However, it is a challenging task given the complex patterns of commuting flows. Conventional models, such as gravity model, are mainly derived from physics principles and limited by their predictive power in real-world scenarios where many factors need to be considered. Meanwhile, most existing machine learning-based methods ignore the spatial correlations and fail to model the influence of nearby regions. To address these issues, we propose Geo-contextual Multitask Embedding Learner (GMEL), a model that captures the spatial correlations from geographic contextual information for commuting flow prediction. Specifically, we first construct a geo-adjacency network containing the geographic contextual information. Then, an attention mechanism is proposed based on the framework of graph attention network (GAT) to capture the spatial correlations and encode geographic contextual information to embedding space. Two separate GATs are used to model supply and demand characteristics. A multitask learning framework is used to introduce stronger restrictions and enhance the effectiveness of the embedding representation. Finally, a gradient boosting machine is trained based on the learned embeddings to predict commuting flows. We evaluate our model using real-world datasets from New York City and the experimental results demonstrate the effectiveness of our proposal against the state of the art.

Zheng Wang, Qiao Wang, Tingzhang Zhao et al.

Feature selection, an effective technique for dimensionality reduction, plays an important role in many machine learning systems. Supervised knowledge can significantly improve the performance. However, faced with the rapid growth of newly emerging concepts, existing supervised methods might easily suffer from the scarcity and validity of labeled data for training. In this paper, the authors study the problem of zero-shot feature selection (i.e., building a feature selection model that generalizes well to "unseen" concepts with limited training data of "seen" concepts). Specifically, they adopt class-semantic descriptions (i.e., attributes) as supervision for feature selection, so as to utilize the supervised knowledge transferred from the seen concepts. For more reliable discriminative features, they further propose the center-characteristic loss which encourages the selected features to capture the central characteristics of seen concepts. Extensive experiments conducted on various real-world datasets demonstrate the effectiveness of the method.

Suhas Lohit, Qiao Wang, Pavan Turaga

Many time-series classification problems involve developing metrics that are invariant to temporal misalignment. In human activity analysis, temporal misalignment arises due to various reasons including differing initial phase, sensor sampling rates, and elastic time-warps due to subject-specific biomechanics. Past work in this area has only looked at reducing intra-class variability by elastic temporal alignment. In this paper, we propose a hybrid model-based and data-driven approach to learn warping functions that not just reduce intra-class variability, but also increase inter-class separation. We call this a temporal transformer network (TTN). TTN is an interpretable differentiable module, which can be easily integrated at the front end of a classification network. The module is capable of reducing intra-class variance by generating input-dependent warping functions which lead to rate-robust representations. At the same time, it increases inter-class variance by learning warping functions that are more discriminative. We show improvements over strong baselines in 3D action recognition on challenging datasets using the proposed framework. The improvements are especially pronounced when training sets are smaller.

Ning Wang, Xianhan Zeng, Renjie Xie et al.

Street architectures play an essential role in city image and streetscape analysing. However, existing approaches are all supervised which require costly labeled data. To solve this, we propose a street architectural unsupervised classification framework based on Information maximizing Generative Adversarial Nets (InfoGAN), in which we utilize the auxiliary distribution $Q$ of InfoGAN as an unsupervised classifier. Experiments on database of true street view images in Nanjing, China validate the practicality and accuracy of our framework. Furthermore, we draw a series of heuristic conclusions from the intrinsic information hidden in true images. These conclusions will assist planners to know the architectural categories better.

Renjie Xie, Yanzhi Chen, Yan Wo et al.

Deep neural networks (DNN) have been a de facto standard for nowadays biometric recognition solutions. A serious, but still overlooked problem in these DNN-based recognition systems is their vulnerability against adversarial attacks. Adversarial attacks can easily cause the output of a DNN system to greatly distort with only tiny changes in its input. Such distortions can potentially lead to an unexpected match between a valid biometric and a synthetic one constructed by a strategic attacker, raising security issue. In this work, we show how this issue can be resolved by learning robust biometric features through a deep, information-theoretic framework, which builds upon the recent deep variational information bottleneck method but is carefully adapted to biometric recognition tasks. Empirical evaluation demonstrates that our method not only offers stronger robustness against adversarial attacks but also provides better recognition performance over state-of-the-art approaches.

Qiao Wang, Zheng Wang, Xiaojun Ye

LINE [1], as an efficient network embedding method, has shown its effectiveness in dealing with large-scale undirected, directed, and/or weighted networks. Particularly, it proposes to preserve both the local structure (represented by First-order Proximity) and global structure (represented by Second-order Proximity) of the network. In this study, we prove that LINE with these two proximities (LINE(1st) and LINE(2nd)) are actually factoring two different matrices separately. Specifically, LINE(1st) is factoring a matrix M (1), whose entries are the doubled Pointwise Mutual Information (PMI) of vertex pairs in undirected networks, shifted by a constant. LINE(2nd) is factoring a matrix M (2), whose entries are the PMI of vertex and context pairs in directed networks, shifted by a constant. We hope this finding would provide a basis for further extensions and generalizations of LINE.

Dorian Cazau, Yuancheng Wang, Olivier Adam et al.

Many methods for automatic music transcription involves a multi-pitch estimation method that estimates an activity score for each pitch. A second processing step, called note segmentation, has to be performed for each pitch in order to identify the time intervals when the notes are played. In this study, a pitch-wise two-state on/off firstorder Hidden Markov Model (HMM) is developed for note segmentation. A complete parametrization of the HMM sigmoid function is proposed, based on its original regression formulation, including a parameter alpha of slope smoothing and beta? of thresholding contrast. A comparative evaluation of different note segmentation strategies was performed, differentiated according to whether they use a fixed threshold, called "Hard Thresholding" (HT), or a HMM-based thresholding method, called "Soft Thresholding" (ST). This evaluation was done following MIREX standards and using the MAPS dataset. Also, different transcription scenarios and recording natures were tested using three units of the Degradation toolbox. Results show that note segmentation through a HMM soft thresholding with a data-based optimization of the {alpha,beta} parameter couple significantly enhances transcription performance.

Adrien Gaidon, Qiao Wang, Yohann Cabon et al.

Modern computer vision algorithms typically require expensive data acquisition and accurate manual labeling. In this work, we instead leverage the recent progress in computer graphics to generate fully labeled, dynamic, and photo-realistic proxy virtual worlds. We propose an efficient real-to-virtual world cloning method, and validate our approach by building and publicly releasing a new video dataset, called Virtual KITTI (see http://www.xrce.xerox.com/Research-Development/Computer-Vision/Proxy-Virtual-Worlds), automatically labeled with accurate ground truth for object detection, tracking, scene and instance segmentation, depth, and optical flow. We provide quantitative experimental evidence suggesting that (i) modern deep learning algorithms pre-trained on real data behave similarly in real and virtual worlds, and (ii) pre-training on virtual data improves performance. As the gap between real and virtual worlds is small, virtual worlds enable measuring the impact of various weather and imaging conditions on recognition performance, all other things being equal. We show these factors may affect drastically otherwise high-performing deep models for tracking.