多模态生成与编辑

Text-Image-to-Video (TI2V) models are an emerging unified architecture, where a single model simultaneously supports text-to-video (T2V) and image-to-video (I2V) generation. Given a high-quality first frame or a detailed textual prompt, TI2V models unlock substantially better visual quality than their T2V mode, raising a natural question: can the capability elicited by such privileged conditions be internalized into the model's own base generation ability? A common approach toward this goal is model self-distillation. However, the most straightforward solution, supervised fine-tuning, follows an off-policy strategy: its supervision is confined to teacher-generated endpoints from a fixed offline distribution rather than student-visited states, lacking precise correction tailored to the evolving policy. Recent on-policy distillation methods instead suffer from condition-state mismatch, where supervision is steered toward the given first frame instead of the student's actual content, misleading the correction. To achieve self-distillation that absorbs the teacher's privileged prior while retaining precise policy correction, in this work, we propose Hybrid-Policy Self-Distillation (HPSD), a novel self-distillation framework where a single TI2V model acts as both teacher and student under different conditions: the teacher operates in TI2V mode with a high-quality first frame and an enhanced prompt, while the student runs in the base T2V mode with only the vanilla prompt. Specifically, the student inherits off-policy teacher trajectory points as anchors, locally refines them toward its own policy, and finally receives velocity-level supervision on these self-generated roll-outs. Extensive experiments demonstrate that HPSD significantly improves T2V performance while also delivering notable TI2V gains, effectively strengthening the model's base generation ability.

Limit order book (LOB) simulators are most useful to practitioners when they combine realistic market dynamics, computationally efficient sampling, controllable scenario generation, and the ability to generalize beyond the instruments seen during training---properties that existing agent-based and deep generative simulators provide only partially. We present FlowLOB, a conditional flow-matching generator of LOB trajectories, trained on multiple Hong Kong Exchange (HKEX) symbols at three sampling frequencies ($0.1$s, $1$s, $10$s) in tick-relative representation that transfers to unseen instruments. Because flow and diffusion models admit a common formulation, we train both with identical data, architecture, and budget, and sample both through the same fixed-step ODE solvers, yielding a controlled comparison of sampling efficiency and fidelity. Flow matching attains its best quality with only $10$ ODE-solver steps, whereas diffusion needs many more function evaluations to approach the same fidelity. At this efficient operating point, FlowLOB improves realism over baselines, two learned and two agent-based models, in most distributional metrics at the two finer sampling frequencies. We evaluate counterfactual controllability with a distributional test that asks whether changing a scenario condition moves the generated statistic toward the corresponding real tail regime; FlowLOB satisfies this criterion in most tested settings. Both realism and control effects transfer zero-shot on a held-out symbol. We additionally conduct ablation studies on the network architecture and the learning rate.

This report presents an improved version of AlayaWorld. While the backbone architecture, chunk-wise autoregressive generation scheme, and training data remain unchanged from the previous release, we substantially revise how conditioning signals are represented and integrated into the model. The new design is guided by a simple principle: conditioning signals should match the generated content as closely as possible in both latent representation and temporal structure. To this end, we make two major changes. First, we replace the previous depth-warping-based spatial memory with a streaming 3D point-cache renderer. Second, we redesign the conditioning pipeline so that visual conditions are encoded in the same causal-VAE latent space, with temporal statistics consistent with those of the generated video. Concretely, the new version introduces six modifications: (1) replacing static-frame image conditioning with motion-aware latent conditioning; (2) causally encoding re-rendered spatial memory as a continuous sequence; (3) aligning the temporal-memory window in pixel space; (4) adopting hard memory dropout that removes memory tokens rather than zeroing them; (5) unifying the VAE encoding and decoding protocol across training and inference; and (6) removing the camera AdaLN branch, such that viewpoint control is provided entirely through the re-rendered spatial condition.

Diffusion models have achieved dominant performance in visual generation but suffer from substantial inference overhead. While cache-based acceleration has emerged as a promising solution, existing policies rely on local similarity heuristics, which we identify as being significantly misaligned with final generation quality. This discrepancy stems from the non-uniform propagation and accumulation of errors along the denoising trajectory. To address this, we propose Global-Impact Cache (GCache). We first establish a rigorous theoretical characterization of the error propagation upper bound. Recognizing that this bound can be overly conservative for complex, highly non-convex diffusion models, we further reparameterize the propagation exponent with a Bernstein form and reformulate cache policy search as a bilevel optimization problem. In detail, GCache identifies an optimal reuse policy in the inner objective while aligning the error-weighting function with generation quality loss in the outer objective. This framework effectively reconciles theoretical rigor with empirical performance, learning to prioritize computation where it most impacts visual fidelity. Extensive experiments demonstrate that GCache consistently outperforms prior caching strategies on both video and image generation. Notably, on the state-of-the-art Wan2.1 video diffusion model, GCache maintains a 2.17x speedup while significantly enhancing generation quality, reducing LPIPS from 0.1095 to 0.0316.

Fake/synthetic images are increasingly prevalent, but it remains unclear whether Convolutional Neural Networks (CNNs) process real and fake images in the same internal manner. This work examines the hypothesis that CNNs represent real and fake images differently, such that fake images induce different hidden-layer activation patterns even when semantic content is preserved. The hypothesis is evaluated in scene recognition settings using trained CNN models. Dense-layer activations are extracted, and neurosymbolic methods assign semantic labels to selected neurons. For each real test image, corresponding fake images are generated with similar semantic content using object-label-guided text-to-image and image-to-image generation based on Stable Diffusion variants. Paired real-fake activation patterns are then compared statistically. Additional experiments with another dataset, CNN architecture, generative model, and JPEG/blur degradation analysis assess robustness. Results suggest that fake images evoke different hidden-neuron activations, and these differences are not explained only by simple image degradation. Overall, the findings indicate that real and fake images differ in CNN hidden-layer activation behavior at least in some settings, which opens the door for follow-up work on making use of this different behavior to improve fake image detection.

Recent breakthroughs in generative models have made text-to-voice generation (TTV) possible, enabling the synthesis of speech directly from textual voice descriptions. However, existing systems face two key challenges. First, they struggle to generate a diverse range of voices, spanning real-world human speakers and fictional characters. Second, they lack robust and flexible voice editing capabilities, such as voice cloning and the ability to modify attributes like emotion and tone. In this paper, we propose VoiceDesigner, a unified framework for text-to-voice generation and editing that supports diverse and controllable voice design. To tackle the above challenges, we propose solutions from two perspectives. First, we develop a hybrid data pipeline that leverages digital signal processing techniques and speech generation models to construct a diverse voice dataset covering both real-world and fictional voices. Second, we introduce a diffusion transformer with architectural improvements to better handle complex conditioning and enhance multi-task performance, enabling unified voice generation and editing. Through subjective and objective evaluations, VoiceDesigner achieves superior prompt alignment with both voice descriptions and editing instructions, while maintaining competitive perceptual quality and voice usability compared to state-of-the-art TTV models.

Visual content is the dominant medium of communication, yet without audio descriptions (ADs), it remains inaccessible to blind and low-vision people. ADs narrate context-relevant visual events during natural audio pauses. Manually creating ADs is expensive, limiting coverage to a small fraction of available content. Most existing automatic AD generation methods frame the task as video clip captioning, requiring ground-truth timestamps and additional context cues such as character databases. Current benchmarks reinforce this framing, consisting of short video segments paired with automatic or task-mismatched annotations. We introduce StrAD, a benchmark for long-form AD generation on full-length videos spanning diverse genres such as movies, documentaries, short films, performances, and video games. We reformulate AD generation as streaming dense video captioning. Our approach processes full-length videos with a sliding window, inserting ADs into existing transcripts without ground-truth timestamps, and supports both fine-tuned models and zero-shot prompting of vision-language models. On the segment-level task with given timestamps, our fine-tuned StrAD-FT sets the state of the art on CMD-AD with 36.3 CIDEr (+10.0 over Shot-by-shot), establishes a reference point on StrAD (51.0 CIDEr), and remains competitive on MAD-Eval at 24.9 CIDEr. On the full-video streaming task, StrAD-FT reaches a SODA score of 2.4 against 1.1 for our zero-shot baseline StrAD-Zero, though both exhibit limitations in temporal localization and narrative coherence. While prior work has tackled full-video AD generation in an offline, multi-stage fashion, ours is the first streaming approach, generating ADs on the fly without ground-truth timestamps. StrAD makes progress on full-video AD generation measurable, a prerequisite for scaling accessibility.

Continuous physiological time series underpin modern clinical monitoring, yet many of the most informative signals are invasive, expensive, or simply unavailable for a given patient. Conditional generation offers a remedy: an absent signal can be synthesized from co-recorded signals and routine clinical variables. Existing generators, however, are built around a single conditioning modality and degrade when forced to handle the heterogeneous, irregularly missing mix of time-variant signals and static covariates seen in practice. We propose ReCoGen (Represent Conditions, then Generate), a two-stage framework that decouples multimodal condition representation from target generation. Stage I trains one masked autoencoder per modality, distilling each time-variant condition into a compact and missingness-tolerant token sequence. Stage II trains a flow-matching generator that fuses these tokens with static conditions to synthesize the target signal. Across three physiological benchmarks, including continuous glucose monitoring on AI-READI and arterial blood pressure generation on MIMIC-III and MIMIC-IV, ReCoGen attains the best downstream utility on all sixteen (dataset, task, metric) settings, surpassing six representative conditional generators; on thirteen of them its utility also reaches or exceeds the utility measured on the real signal, a reference we read as an approximate anchor rather than a ceiling. Ablations trace the gains to the conditioning path: learnable cross-attention over the frozen per-modality encoders, and a dual token-plus-AdaLN route for the static conditions. ReCoGen thus turns routinely collected signals into informative surrogates for invasive or unavailable ones, a step toward less invasive, lower-cost continuous clinical monitoring.

Automating empirical research is a long-standing direction of AI. Recent automatic research (AutoResearch) agents bring this goal within reach, as modern LLMs show the capability to independently implement solutions and learn from the execution outcomes. Behind these gains, post-training (especially RL) plays a central role. In this paper, we identify a fundamental tension when scaling RL for these agents: the two components of every AutoResearch trajectory (agent generation and environment execution) scale in very different manners, since all generation shares compute through batching, while each execution occupies its exclusive sandbox and real machine time. As a result, the environment execution dominates the training cost and becomes the bottleneck as trajectories grow. To resolve this tension, we propose World Model RL (WMRL), which replaces environment execution with a world model to remove this bottleneck. Additionally, the world model can be imperfect, as its rewards are corrupted by bias and noise. Therefore, we further equip WMRL with two mitigations, Online Debiasing and Inverse-Variance Denoising, which offset the bias and suppress the noise respectively. Theoretically, we prove that both mitigations of WMRL strictly improve the convergence guarantee. Empirically, WMRL accelerates training by 3-4x on various tasks at different agent scales, while exceeding the performance of standard RL baselines. Moreover, our post-trained 4B and 9B agents outperform much larger open-weight agents of 48B and 120B on held-out benchmarks. Beyond AutoResearch, WMRL also transfers to post-training embodied VLA policies, which demonstrates the generalizability of our method.

In-vehicle music can serve as an adaptive interface to enhance driver experience, attention, and well-being. We present Drive-to-Music, a context-aware system that generates music in real time from multimodal driving signals. Using dashcam imagery and vehicle telemetry, the system extracts scene semantics and driving context, maps them to high-level musical descriptors, and conditions generative audio models to produce contextually aligned soundtracks. The architecture combines perception and generative components to translate visual and kinematic inputs into structured musical attributes and synthesize audio with low latency. It supports smooth transitions as driving conditions evolve, and to ensure robustness and deployment readiness, we incorporate constraint-based controls and safety checks across the generation pipeline. Our results demonstrate the feasibility of real-time, context-aware music generation in automotive settings, providing a foundation for personalized and adaptive in-vehicle audio experiences.

Existing screenshot-to-code systems face a trade-off between flexibility and controllability. Direct multimodal generation can hallucinate visible details, whereas structured pipelines reduce such errors through component-wise decomposition, predefined templates, and customized intermediate representations. These structures, however, introduce additional generative orchestration and restrict outputs to designs covered by the representation. We investigate whether selective tool grounding can improve the fidelity--efficiency trade-off of direct widget-to-code generation. We introduce WidgetGen, a lightweight tool-grounded framework that extracts observable text and color evidence, performs high-level layout and optional chart reasoning, and directly generates executable JavaScript XML (JSX). This design reduces reliance on component-wise generation while avoiding a fixed UI schema. Across six multimodal models and \(1{,}000\) held-out widgets, WidgetGen outperforms direct prompting and the structured Widget2Code pipeline on most visual reconstruction metrics, with consistent gains in area, legibility, and style. Finally, reconstruction-derived image-code pairs improve six Qwen-family open-weight models across every reported metric through supervised fine-tuning. These results establish WidgetGen as a strong lightweight baseline and show that selective evidence grounding offers an effective alternative to extensive representation constraints.

Understanding how lower-limb muscle groups coordinate is important for studying movement impairment, rehabilitation, and physical performance. Reproducible analysis of this coordination requires multimodal recordings that relate local muscle-related signals with body-level kinematics. Complementing neural-level electrical activation captured by EMG, AMG provides a valuable mechanical approach to monitoring muscle activity. Here, we introduce a synchronized, multimodal dataset for healthy-adult lower-limb activities. For data collection on the left leg, 16 triaxial accelerometers were evenly divided into four muscle-site clusters for AMG recording, complemented by four surface EMG channels. A 15-marker optical motion-capture (MoCap) system captured lower-body kinematics, with the resulting marker trajectories used to compute bilateral knee and ankle joint angles. Our dataset contains 1,918 trials from 30 subjects across 16 task conditions. We benchmark the dataset by estimating four joint angles from 300 ms windows of the 5-100 Hz band-pass-filtered AMG data and assess matched EMG features in a separate modality ablation. In the primary cross subject benchmark, the four reference models achieved mean absolute errors of 8.840$^\circ$-9.591$^\circ$. The benchmark and ablation results characterize performance across subjects, tasks, and joint angles and examine the effects of sensor configuration, modality, the number of training subjects, and frequency representation. The release includes documented timing definitions, processed data, and reproducible benchmark resources. https://dongxutang918-afk.github.io/SAME-Limb/

As Large Vision-Language Models increasingly aim to integrate visual generation and understanding within a single parameter space, evaluating such structural unification in a cohesive manner remains a critical challenge. Current evaluation protocols predominantly treat generative and discriminative capabilities as separate tasks, leaving a gap in system-level evaluation for unified multimodal models (UMMs). In this work, we propose Self-Generative-Understanding (SGU), a novel, annotation-free evaluation framework that probes the integrated capabilities of unified models through a semantic closed-loop challenge. Without requiring new annotations, SGU leverages the dual understanding-and-generation abilities of UMMs by asking them to first perceive an image and produce a textual description, subsequently reconstruct a visual context based on that description, and finally perform reasoning over the self-generated output. This pipeline provides a zero-cost testbed that yields an integrated performance score specifically tailored for evaluating UMMs as unified systems. Extensive experiments show that even high-performing UMMs often struggle to reason over their own generated contexts, revealing limitations that are not captured by separate evaluations of understanding or generation alone. Our work provides a complementary holistic evaluation framework and offers a foundation for benchmarking the development of next-generation unified multimodal models.

Learning world models from offline trajectories enables agents to accomplish different tasks through planning. Object-centric (OC) representations, which decompose a scene into a set of slots that bind to its objects, have been proposed as an inductive bias for world models that are more sample-efficient and generalize better. Yet prior object-centric world models (OCWMs) take the slot encoder as given and evaluate only in-distribution, leaving open whether the object-centric bias actually delivers for planning and what within the OCWM drives it. We conduct a controlled study of OCWMs for visual model-predictive control along two axes: object-centric representation quality and generalization under distribution shift relative to scene-centric models. We find that (i) planning success correlates positively with unsupervised slot-quality metrics (FG-ARI, mBO), though the gains saturate at high slot quality; (ii) with well-bound slots, the auxiliary proprioception inputs and masking inductive bias that prior methods relied on become unnecessary; and (iii) under unseen distribution shifts, the OCWM with well-bound slots is more robust overall than the end-to-end trained scene-centric LeWM, while DINO-WM, built on similar frozen pretrained features, remains competitive — pointing to pretrained features as a key contributor to robustness.

Video-to-audio (V2A) generation faces significant challenges in achieving precise temporal synchronization and high perceptual quality due to the complex, ambiguous relationship between visual and auditory cues. Existing methods typically compress video inputs into single feature representations, leading to significant loss of temporal dynamics and fine-grained visual information. These approaches also rely on reconstruction-based training objectives that poorly correlate with human perceptual judgments of audio quality and appropriateness. We propose HarmoniDPO, a novel framework that integrates preference-based optimization into diffusion-based V2A generation to address these limitations. (1) Our approach leverages a dual video representation: combining global context with frame-wise features to preserve temporal dynamics and semantic detail. (2) Inspired by reinforcement learning from human feedback (RLHF), HarmoniDPO employs online Direct Preference Optimization (online-DPO) to fine-tune a diffusion-based V2A model from preference judgments, enhancing perceptual quality and alignment. (3) Additionally, we introduce Dual-scale Diffusion Search (DDS), a test time scaling algorithm that adaptively optimizes output fidelity during inference. Experiments demonstrate that HarmoniDPO outperforms state-of-the-art methods in audio-video synchronization and subjective audio quality, offering a robust solution for generating realistic, human-preferred audio from video.

Robotic manipulation with dexterous hands is a cornerstone of Embodied AI, yet its progress is stifled by the high cost of collecting embodiment-aware teleoperation data. While abundant egocentric videos of human hands offer a scalable alternative, the profound discrepancies in appearance, articulation, and camera viewpoints between human and robotic data raise significant challenges for co-training. Though existing general image-editing models demonstrate strong capabilities, they lack necessary embodiment-specific priors to fully bridge this gap. In this work, we present HandEdit, a unified large-scale embodiment-aware image-editing dataset and benchmark specifically designed to transform human hands and arms into various dexterous robotic embodiments within egocentric frames. HandEdit comprises over 200M editing instances derived from five diverse source datasets, covering 26 distinct URDFs, including 13 hand-only and 13 hand-arm configurations. Alongside the dataset, we establish a unified benchmark protocol with two tracks: Hand-only and Hand-Arm, supporting URDF-conditioned evaluation. We conduct extensive evaluations of 11 representative image-editing baselines using a multi-dimensional metric suite, including generic similarity metrics, VLM-based judgment, and embodiment-aware metrics. HandEdit serves as a critical resource at the intersection of image editing and robotics: it advances embodiment-aware editing models while enabling scalable dexterous robotic learning from abundant human video data, paving the way for more generalizable Embodied AI.

Quantum generative adversarial networks (QGANs) have attracted increasing attention for image generation using parameterized quantum circuits. Existing amplitude-based approaches face two key limitations: pixel locations are typically encoded by computational-basis indices or address qubits, causing quantum resources to grow with image resolution; meanwhile, jointly decoding many pixels from normalized quantum states introduces probability competition among pixels and limits precise pixel-wise control. To address these issues, we reformulate quantum image generation as coordinate-conditioned implicit function learning. Our method takes spatial coordinates and latent variables as inputs, uses a classical embedding network to generate input-dependent circuit parameters, and evaluates a variational quantum circuit at each coordinate. Pixel intensities are directly obtained from the expectation value of a dedicated color qubit, and a complete image is generated by querying all spatial coordinates. This design decouples image resolution from address-qubit requirements and avoids shared probability-normalization constraints across pixels. We further design a specialized variational quantum circuit to provide structural inductive bias for coordinate-conditioned generation. Simulated experiments on two benchmark datasets show that our method outperforms FRQI-based generation and PQWGAN in visual and quantitative quality while using fewer qubits, and also achieves better generation quality than the corresponding classical baseline.

Talking-video character replacement requires coordinated transfer of appearance and voice while preserving the source motion, scene, linguistic content, and audio-video timing. Existing methods use separately optimized models for the two modalities, making audio-visual consistency difficult to enforce. We present UniSwap, the first framework for streaming joint audio-visual identity replacement in talking videos. Given a source video, a reference image, and a reference voice clip, UniSwap transfers the reference appearance and vocal timbre within a single audio-visual diffusion transformer while preserving the source content and dynamics. To address the scarcity of aligned cross-identity training pairs, we introduce a swap-and-reconstruct pipeline that removes visual and vocal identity from real clips and uses the original clips as reconstruction targets. Starting from a bidirectional backbone, we progressively adapt the model through In-context Pretraining for joint replacement, Conditional Streaming Adaptation for block-causal KV-cached generation, and Efficient Self-forcing DMD for mitigating exposure bias and reducing sampling from 30 to 3 denoising steps per block. Efficient Multi-LoRA Switching enables the three DMD roles to share a single frozen backbone. Feature-RoPE Decomposition keeps cached positions within the training range, supporting stable long-form inference. Experiments demonstrate strong audio-visual synchronization, competitive identity preservation, efficient streaming, and stable long-form generation.

Modern black-box Image-to-Video (I2V) models offer powerful capabilities in automated content creation, yet their lack of fine-grained control and reliability presents significant challenges in professional workflows. Their inherent stochasticity causes minor variations in textual prompts or hyperparameters to yield drastically different outputs often necessitating inefficient, brute-force trial-and-error processes. To address these limitations, we introduce the ``Agentic Self-Improvement" framework, which reframes video synthesis into a closed-loop, goal-directed optimization. Our framework systematically navigates the generation parameter space using a novel two-stage approach. In the first stage, an iterative prompt optimization loop uses a multimodal Large Language Model (mLLM) to refine the input prompt. This refinement implements two automated evaluations: Davidsonian Scene Graph (DSG) queries ensure semantic adherence, and Common Mistake Questions (CMQ) for artifact detection. At the second stage, we use Bayesian optimization to efficiently co-optimize stochastic seeds and CFG scales. This search is guided by a suite of quality metrics, including the novel Video-Text Adherence (VTA) score derived from the DSG and CMQ evaluations. Our framework significantly outperforms unguided search methods: in human preference studies, videos generated via our agentic approach were strongly preferred over baseline outputs, achieving win rates up to 69%. This work provides a practical and extensible methodology for enhancing the predictability and control of state-of-the-art video generation models, moving the field beyond speculative curiosities toward reliable, production-ready tools.

Proprietary text-to-image diffusion models are increasingly distributed as hosted services and downloadable checkpoints, making their intellectual property (IP) protection an increasingly critical concern when model leakage, copying, or unauthorized fine-tuning is disputed. In this work, we present a non-invasive model fingerprinting framework based on collapsed generation, a phenomenon where certain input conditions produce highly consistent images across multiple stochastic seeds. We show that collapsed generation is an intrinsic, model-dependent property of the learned generation process. These collapse-prone conditions therefore expose model-specific behavioral signatures, enabling reliable ownership verification without embedding invasive watermarks. After preparing conditions on the source model, the framework verifies a suspect model under two access settings: (1) white-box pipeline access, where optimized continuous embeddings can be injected into the generation process, and (2) black-box API-only access, where natural language prompts are queried through the service interface. In both cases, ownership evidence is measured by whether the suspect model reproduces the source model's collapse behavior across stochastic samplings. Extensive experiments across UNet- and transformer-based diffusion models show that collapsed generation fingerprints can distinguish different source models with low confusion. These fingerprints remain verifiable in fine-tuned derivatives and under common and adaptive model- or query-level obfuscations, while requiring only a modest verification query budget. Together, these results establish collapsed generation as a reliable intrinsic evidence source for non-invasive diffusion model ownership verification.

Generative models like Diffusion Models and Flow Matching have demonstrated remarkable capabilities in synthesizing high-fidelity driving videos, but are severely constrained by high inference latency due to the requirement of extensive sampling steps. We argue that this inefficiency stems from the prevailing reliance on a standard Gaussian source distribution, where consecutive frames are initialized as independent Gaussian noise. This paradigm disregards the rich spatiotemporal correlations inherent in driving videos, compelling the model to regenerate deterministic scene structures existing in previous frames from noise, which is both computationally redundant and prone to geometric inconsistency. To address this problem, we propose GeoFlow, a novel framework designed to achieve efficient driving video generation by harnessing explicit geometric priors. Instead of sampling from standard Gaussian noise, we leverage multi-view geometry and spatially-adaptive noise injection to construct a Geometry-Aligned Prior (GAP) distribution as starting point. This initialization bridges the gap between source distribution and data distribution, yielding a significantly straighter and shorter sampling trajectory. Extensive experiments demonstrate that GeoFlow can achieve remarkable efficiency of both training and inference: merely several hours of fine-tuning on baseline models can significantly boost few-step generation quality, while fully converged training drastically reduces number of inference steps required for state-of-the-art video generation.

Existing streaming video systems often rely on sequential, distillation-centered training pipelines to enable few-step long-video generation. However, this paradigm suffers from two limitations. First, failures or distribution shifts introduced in earlier stages affect later optimization, complicating the training process to converge. Second, the distillation-centric objective favours short-term generation but is prone to quality degradation when autoregressive errors accumulate over long rollouts. We propose Avatar-Forever, a decoupled parallel training framework for high-quality real-time infinite interactive avatars. Instead of coupling generation efficiency and long-horizon robustness under a sequential distillation pipeline, we treat them as two independent capabilities that can be trained in parallel. One branch performs full-parameter distillation to train an efficient generator with high visual quality, while another trains a lightweight long-horizon adapter via Recovery-oriented Rollout Training (RRT), which improves generation robustness under long-horizon inference conditions. Our decoupled parallel training design simplifies the overall training process and avoids unnecessary objective conflicts between few-step generation and long-horizon adaptation. We further introduce ForeverCache, a chunk-wise feature caching mechanism to substantially reduce redundant history computation during streaming inference. Built upon a 22B video foundation model, Avatar-Forever supports unbounded audio-driven avatar generation while maintaining identity consistency, motion coherence, and visual fidelity, enabling an end-to-end throughput of high-resolution 768x512 videos at 27.2 FPS on a single H100 GPU and providing a practical path toward stable digital humans.

Previsualization is an intermediate layer between ideas and production in film, games, architecture, and urban design. It lets creators iteratively refine scenes, actions, cameras, and spatial-temporal dynamics. Yet existing generative methods rely on simple prompts to jointly control all of these factors through one-shot image or video synthesis, offering weak controllability and limited support for iterative editing. Fundamentally, a world comprises multiple elements with geometry, appearance, and other attributes, together with cameras. Different frames are produced through local modifications or recombinations of this shared state, which is otherwise largely reused. Therefore, we argue that the missing component is an explicit and persistent working state. To address this, we present StateFlow, a state-centric framework for generative previsualization. Rather than generating videos in one shot, StateFlow uses an editable 3D world to organize scene structure, evolution, and cameras, while off-the-shelf video models enhance visual quality when higher fidelity is desired. This world is maintained as a persistent structured 3D state of scene elements and camera configurations, serving as the core working representation for previsualization. Built on this insight, StateFlow has three stages to construct, evolve, and access the world state. State construction lifts generated 2D content into a coherent 3D world through prior-guided, conflict-aware dual-view initialization, while State evolution translates user intent into structured state transitions while preserving world memory, avoiding full-scene regeneration for each edit. State access uses render-feedback reflection to refine camera plans into visually feasible trajectories, avoiding reliance on VLM semantics alone. Experiments show that StateFlow produces high-quality 3D worlds for video creation and game-like prototyping.

Multimodal large language models (MLLMs) have advanced image geolocalization mainly by improving how they reason about geographic cues. How that reasoning isdecoded into coordinates, however, has lagged behind. Predicting a place name for a geocoding API is discrete and lossy: it ignores image evidence and collapses multi-granular semantics into a coarse lookup. We argue that the bottleneck has shifted from what a model reasons to how that reasoning is represented for a continuous, geometry-aware decoder. We present GeoBridge, a role-decoupled conditioning mechanism that connects a frozen semantic MLLM to a frozen Riemannian flow-matching head that generates coordinates on the sphere. The central obstacle is arole conflict: supervising the condition with discrete semantic labels biases its representation toward class-discriminative geometry, at odds with the smooth manifold the generative head requires. GeoBridge keeps the semantic supervision decoupled from the condition interface: a separate projection forms the continuous condition the frozen head expects, injecting geographic priors without disturbing the spherical decoder. On IM2GPS3K, GeoBridge reaches 38.67/52.89/70.37 at the 25/200/750 km thresholds, improving over a place-name-to-API pipeline and reasoning-augmented direct prediction at these precision-relevant scales. GeoBridge is a decode-side algorithmic contribution, orthogonal and complementary to chain-of-thought reasoning. Code will be made publicly available.

Text-to-image (T2I) models are increasingly asked to illustrate literary and cultural content, yet we cannot measure how well an image renders the meaning of a poem. The task is many-sided: a good illustration must be visually sound, faithful to the poem's imagery and scene, culturally and stylistically apt, free of spurious text, and true to its emotion, and its deepest requirements, imagery and especially implicit emotion, are never stated in the words. Existing metrics (CLIPScore, BLIPScore, VQAScore) reward literal text-image correspondence and so cannot tell whether an illustration succeeds, let alone why, or even separate the best model from the worst. We introduce TangPoetryBench, a multi-dimensional benchmark of 1,280 images (320 classical Chinese Tang poems x 4 state-of-the-art T2I models) with quality-controlled human annotations across ten dimensions. Analyzing this data, we reveal the shared and model-specific strengths and weaknesses of current T2I models, including their ability to evoke a poem's implicit emotion. We further introduce PoemAutoEvaluator (PAE), an open, rubric-conditioned evaluator that reaches parity with a strong proprietary judge (Claude), generalizes to an unseen generator and a second poetic tradition (Song Ci), and lets the benchmark scale to new images without fresh human annotation. We release the benchmark, annotations, and evaluator.

Extended Reality (XR) systems increasingly deliver high-fidelity visual and auditory experiences, yet tactile perception remains comparatively underutilized as a modality for enriching embodied interaction. This work presents a visual-to-haptic wearable glove and a feature-based visual-to-haptic mapping algorithm that translates spatial and temporal visual features from images and videos into distributed vibrotactile patterns. The proposed method extracts motion, edge, and brightness cues and fuses them into actuator-level intensity maps aligned with a 29-actuator glove arranged in a five-by-seven layout. The system is implemented through a modular four-layer architecture comprising the XR environment, media content handling, visual-to-haptic processing, and embedded haptic hardware. A within-subject user study (N = 20) compared visual-only interaction with visual-plus-haptic augmentation across texture-based and dynamic video scenarios. Results indicate that tactile augmentation significantly improves perceived realism in dynamic video scenarios and enhances immersion and visual-tactile correspondence across conditions, with stronger and more consistent effects observed for dynamic visual events. While the current implementation operates in a single-user, offline-synchronized configuration, the findings demonstrate that vision-driven tactile augmentation can function as a perceptual enhancement layer within multimodal XR systems. Such a layer may provide a foundation for future socially enriched XR environments where coherent multisensory grounding supports higher-level interaction and communication.

4D generation synthesizes dynamic 3D scenes from conditions such as text or images. Existing methods either reconstruct generated RGB videos with a separate 4D model or adapt a particular video generator to predict geometry directly. The former suffers from distribution mismatch and error propagation, whereas the latter ties 4D prediction to a specific generator and may require retraining when the generator or conditioning regime changes. We ask whether the final denoised latents of video models that share a variational autoencoder (VAE) can instead provide a reusable interface to explicit 4D prediction. Building on this insight, we introduce direct latent-to-4D generation and instantiate it as Latent-to-4D, which bypasses RGB by aligning a video latent with the token grid of a pretrained 4D decoder and refining it through frame-wise and global spatiotemporal attention. Trained on roughly 1K existing reconstruction clips, a single checkpoint transfers unchanged across multiple video diffusion transformers within the same VAE family. On Text4D-200 and I4D-200, Latent-to-4D surpasses matched same-latent Wan+4RC cascades in projection-based DINO-F1 by 2.88--3.45 and 5.81 points, respectively, while also being preferred by human raters for geometry, temporal stability, and overall quality.

Recognizing human behavior across levels of abstraction, from atomic actions to long-horizon intentions, requires data annotated along a semantic hierarchy. Large corpora provide isolated, atomically labeled clips without temporal composition, whereas recorded composite-activity corpora offer shallow, domain-narrow, fixedhierarchies. A benchmark-generation and evaluation frameworkis proposed that synthesizes a four-level hierarchical-intention benchmark, spanning actions, activities, low-level intentions (LLIs), and high-level intentions (HLIs), from a flat single-label action corpus while retaining real pre-extracted features at the action level. Episodes are assembled by a transition model under a subject-consistency constraint, and a coverage-aware sampler reduces the subject usage Gini from 0.566 to 0.248. Synthesizing such a benchmark raises a circular-supervision risk that recorded datasets avoid: if the rules generating the episodes also govern the evaluation, models can succeed by recovering the generator rather than through genuine reasoning. Validity is addressed by design, holding sequence-generation rules disjoint from the first-order-logic rules used at evaluation. The instantiation yields 15,002 episodes. Four reference baselines from different model families characterize difficulty, not as recognition methods. A compositional held-out gap of 0.13 to 0.17 macro-F1 appears across all baselines, including a graph-aware model that recognizes best yet does not close the gap, indicating a structural property of the benchmark rather than a model artifact. A logic-free baseline still violates the held-out semantic rules above their intrinsic data rate, and the order-destroying control changes macro-F1 within seed variation, serving as a generator-consistency check. Theontology, transition model, and generator are released so the benchmark can beregenerated and extended.

Text-only training is a popular paradigm in zero-shot video captioning, where the video distribution is not available to the model during training, leading to a cross-modal gap between the training (text-only) and the inference (video-only). Previous works attempt to bridge the gap through simple linear transformations. However, the inherent gap between text and video makes cross-modal representation space alignment insufficient, resulting in inaccurate sentences. To address this issue, we propose a novel zero-shot video captioning framework (WSV) consisting of two training stages, which first generates corresponding synthetic video latent representations via a pretrained text-to-video generation model. To strengthen the fidelity of the latent representations, we propose a polisher capable of bridging the gap between real and synthetic video distributions. Subsequently, we design a prompter that conditions GPT-2 on the polished latent representations to generate the captions in the second training stage. During inference, an input video is encoded by a pretrained 3D Causal VAE and then fed directly into the prompter, which in turn guides GPT-2 to produce the final caption. Experimental results conducted on MSVD, MSR-VTT, and VATEX datasets demonstrate that our proposed method achieves scores of 52 and 95.7 on the B@4 and CIDEr metrics, respectively.

Embodied artificial intelligence aims to develop agents that perceive, reason, and act through continuous interaction with the physical world. However, most embodied systems are still evaluated within conservative safety margins or moderate interaction regimes, leaving their capability boundaries under extreme conditions insufficiently understood. Autonomous racing provides a stringent testbed by combining high-frequency localization and perception, adversarial interaction, near-saturated vehicle dynamics, and strict safety constraints. Existing systems push high-speed performance but rarely model and refine cognitive and physical limits jointly. Here we show that a world-model-centric autonomous racing agent provides a concrete step toward exploring these coupled limits. The framework learns predictive world models from near-limit successes and failures to capture interaction evolution, ego dynamics, and feasible-motion boundaries, coupling world-state construction, future-aware reasoning, and near-limit control in a closed-loop refinement process. Training data were collected from real-vehicle autonomous racing, where the onboard system maintained robust localization and perception at speeds up to 256.3 km/h and peak lateral acceleration of 26.8 m/s$^2$. In full-scale simulated racing, the well trained world-model-centric agent achieves an 88.3% interaction success rate across various challenging simulated racing scenarios. Closed-loop refinement of the world model and policy further improved utilization of cognitive-physical limits, recovery from failure modes, and generalization across varying conditions and unseen circuits. These results suggest a boundary-aware methodology in which world models help embodied agents represent, predict, and continually refine their capability boundaries for safer real-world deployment.