Multimodal misinformation detection methods often rely on different preprocessing pipelines and auxiliary resources, making experimental comparisons difficult to interpret. We present MMFact, a collection of shared resources and tools that simplifies this research workflow. It brings together 47 detector implementations and baselines, three multimodal datasets, and prepared materials ranging from semantic features to OCR, image descriptions, and external evidence. Compatible methods can use the same materials, reducing variation introduced by preprocessing and supporting fairer comparisons. A modular training and evaluation framework preserves method-specific designs while providing common experimental procedures. By treating prepared materials as a shared research resource, MMFact helps researchers focus on developing and evaluating detection methods.
Caroline Mazini Rodrigues, Nicolas Keriven, Thomas Maugey
With the increasing adoption of deep learning for applications such as image compression, improvements in the rate-distortion trade-off have been achieved at the cost of increasingly larger and more opaque ''black-box'' models. Autoencoders are among the most widely used architectures for this task; however, without a clear understanding of their internal behavior, these models tend to grow in complexity to achieve more performance gains. In this paper, we investigate whether universal behaviors can be detected from the internal operations of bias-free autoencoders through Jacobian analysis. If such behaviors exist, they may be extracted to design low-complexity image compression models inspired by high-complexity deep learning architectures.
An ad break in a live stream is a hard deadline. Once a cue arrives, the ad path has seconds to produce a state the player can use, and starting that work earlier buys time while giving the prepared state longer to expire before the break. Comparing trigger policies means scoring which of these happened; the records easiest to collect do not. We separate trigger, reach, readiness, playback, billability and measurement in an observed-event ledger, fix the reference-deadline rule before candidates are assigned, and account for contribution over every planned break-session pair, reached or not. Readiness means the declared state is valid at that deadline. We prove when scoring completion by the deadline instead reverses a policy comparison: for monotone playback with a shared latency law, a fixed state lifetime and common coefficients, the reversal occurs when the earlier policy's extra trigger cost falls between its validity gain and its completion gain. A synthetic study of 45,000 scripted sessions matches this condition on monotone paths, a check of its generator, and, outside its assumptions, finds the same reversal in mixtures with restarts, seeks, late joins and pauses. Playhead summaries misclassify up to 41% of reach indicators yet reverse no comparison in the study: maximum-position errors cancel exactly, and terminal-position errors change only magnitudes. Completion scoring reverses the comparison in three of nine short-lifetime conditions, all at the one latency setting where the proved condition holds. The evidence is synthetic by design; code, seeds and event histories are public.
Recent research on Multimodal Sentiment Analysis (MSA) has focused on learning from language, visual, and acoustic modalities with incomplete data to infer human sentiment. Most studies typically compensate for missing information by reconstructing modality features or designing complicated fusion mechanisms. However, these methods still suffer from spurious generation and noisy guidance due to the lack of high-level semantic grounding in partially observed multimodal evidence. To address these issues, we propose SemMSA, a latent semantic-aided framework that constructs rich sentiment-relevant semantics with LLMs, fully integrating with all modalities via anchor-free spectral alignment. It mainly consists of Cross-modal Semantic Refinement (CSR) and Cross-modal Spectral Alignment (CSA). Specifically, CSR first adaptively extracts visual and acoustic representations by corresponding adapters to form a unified multimodal prefix with language in the frozen LLM embedding space. It then iteratively produces continuous discriminative semantic states through a token-efficient latent refinement process without decoding explicit text. Next, CSA simultaneously aligns the refined semantics with all modalities by enhancing the dominant spectral component of their kernel Gram matrix. This captures global nonlinear dependencies among all representations without relying on a predefined anchor modality. In addition, an instance-level spectral separation constraint preserves cross-sample discriminability and mitigates representation collapse. Extensive experiments on SIMS, MOSI, and MOSEI benchmarks demonstrate that SemMSA achieves state-of-the-art performance.
Context. Video ads reach streaming and connected-TV apps through VAST documents, and companies count ads from the HTTP requests the player's client library sends. A tracking error can leave playback intact and surface later as disagreeing counts. Objective. To measure how far open-source VAST clients meet the player duties of VAST 4.3, whether verdicts depend on the runtime, what generated inputs add, and when deviations were introduced. Method. A catalog of 108 requirements was extracted from the specification. Six open-source clients in eight configurations ran against 69 crafted test vectors and the IAB's sample tags, under Node.js and in five browsers on three engines, with every request recorded. Grids, fuzzing and a sweep over 36 releases extend the catalog; Google's closed IMA SDK is a comparison. Results. The open-source clients have 78 findings and 3 to 21 must-level deviations per configuration. None of the tested open-source clients that follow redirect (Wrapper) chains fills the required macros in a Wrapper's request; the archived IMA 3.794.0 configuration tested does. One library's timeout throws an exception no caller can catch, ending a Node.js process. In a pod (ads played back to back), one player credits a finishing ad's completion to the next ad, fixed in release 5.3.0. A library can conform under jsdom and deviate in every tested Chrome- and Safari-family browser. Conclusions. The tested clients diverge on the duties other companies use to count and diagnose ads. A shared conformance suite run in real browsers could expose most graded deviations; the harness and vectors are released.
Adaptive streaming of 360-degree video relies on viewport prediction to allocate bandwidth efficiently. Current approaches predominantly use visual saliency or historical gaze patterns, neglecting the role of spatial audio in guiding user attention. This paper presents a self-learning framework for detecting "surprising" auditory events, moments that deviate from learned temporal expectations, and demonstrates their utility for viewport prediction. The proposed architecture combines SE(3)-equivariant graph neural networks with recurrent temporal modeling, trained via a dual self-supervised objective. A key feature is the natural modeling of temporal attention decay: surprise is high at event onset but diminishes as the listener adapts. Experiments on the AVTrack360 dataset show that integrating audio surprise with visual cues reduces bitrate waste by up to 18% compared to visual-only methods.
Cryptographically verifiable C2PA-based Content Credentials (CCs) can be attached to freshly acquired or generated digital images but also to already existing ones with no previous provenance information, for instance after processing through trusted C2PA-enabled software tools. The latter case leads to images with so-called partial CCs, which capture only the last operations applied, while prior provenance remains unknown and can be investigated through signal forensics. This paper addresses the forensic analysis of such images, highlighting the opportunity to leverage the partial CCs as additional trusted knowledge to adapt forensic detectors towards the statistical characteristics of the image under analysis and mitigate potential distribution shifts. To this end, a two-step strategy is envisioned: (i) reconstructing the operations applied by the last signer through a structural analysis of the partial CCs; (ii) exploiting the obtained reconstruction to enhance the forensic detector of the analyst through an adaptation process. An empirical analysis on partial CCs produced by commercial C2PA-enabled software tools through different editing operations is carried out, providing evidence on the feasibility of the reconstruction in practical cases. An experimental case study on the forensic task of synthetic image detection is then presented, showing that the proposed strategy can significantly enhance the performance of state-of-the-art detectors on test images that are postprocessed via C2PA-enabled software tools after being generated/acquired.
Video-queried sound separation (VQSS) aims to isolate a target sound from an audio mixture using its corresponding video as a query. However, scene-level video queries can become ambiguous when multiple sounding objects are visible, particularly when the target and interferer are semantically similar. We propose SelectSep, a target-selective VQSS framework that jointly uses text and a single click coordinate to specify the target source and instance without requiring segmentation at inference time. We adapt the visual encoder through target-feature distillation with text and click conditioning, and then fine-tune the separator using the adapted visual features. We also introduce an on-screen interference setting, where both the target and interferer are visible, to evaluate target selectivity. Experiments on standard benchmarks and the on-screen interference setting demonstrate consistent improvements in separation quality and target selectivity, particularly for mixtures containing semantically similar sources.
Music score files are more difficult to compare than their text-based equivalents, as their content is more complex and organised into hierarchical structures. This explains the lack of tools for processing music scores similar to the widely used old Unix diff command. We propose a new algorithm and tool for comparing digital music scores in various encodings, such as MusicXML or MEI, based on a graph-structured abstract representation called Tree Score Model (TSM). Given two scores, after conversion into TSM, our procedure calculates jointly: A list of differences, localised by time positions in the input scores; A time distance, cumulated duration of time-spans where scores differ; An edit distance, called surface distance, based on notational attributes not related to time. The two later measures are normalised and combined into a similarity distance between scores. Our procedure has been implemented in C++ and is available as a Python module. Our first experiments have shown a very good efficiency even on challenging data. As a proof of concept, we propose two demos based on different versions of some Mozart's piano sonata, and transcriptions of monophonic soli on the same jazz standards by different performers. https://tsm-diff-lbd-7a7e2a.gitlab.io/demo.html
This study aims to build effective drill software by combining a virtual simulation of a disaster with multiple agents capable of expressing emotions. We evaluated how the number of agents and their emotional and behavioral actions during evacuation influence users in a virtual space. The results showed that users tended to act more confidently when many agents evacuated with them, and that the evacuation route chosen by the agents influenced the users' decision-making. These findings indicate that users' evacuation behavior is shaped not only by the environment itself but also by the collective behavior of surrounding agents, underscoring the need to consider how agent behavior affects users' psychological states during training.
In this work, we explore 2D Gaussian splat as a new primitive for representing and compressing videos, showing its promise as an alternative to traditional signal-based primitives. We introduce GSVC, a new Gaussian splatting-based video codec that incorporates common techniques used in classical video codecs to exploit temporal redundancy between nearby frames. In GSVC, each frame is represented by a set of overlapping Gaussian splats, which, when rendered together, approximate the original frame. Gaussian splats for the first frame and subsequent frames following scene changes (called anchor frames) are learned from scratch, while the remaining frames (called propagation frames) are initialized from the Gaussian splats of their previous frames and then fine-tuned. Scene changes are detected through loss differences during the training process. To capture large motion or newly appeared objects, we randomly add Gaussian splats to propagation frames before fine-tuning. GSVC supports a knob for rate-distortion trade-offs through controlling the number of Gaussian splats for each frame. Gaussian splats with low contributions to the frame quality can be removed during training. GSVC supports two-pass encoding by propagating the Gaussian splats from a future propagation frame back to an anchor frame, further improving the video quality for challenging video scenes. To achieve efficient compression, GSVC uses a residual-based Gaussian splats quantization framework that encodes only changes to the Gaussian parameters between frames. Experimental results show that GSVC achieves rate-distortion performance comparable to state-of-the-art video codec implementations for AV1 and VVC. GSVC inherits the advantages of Gaussian splatting in rendering, allowing it to achieve the rendering speed of 1500 fps for a $1920\times 1080$ video. Project page: https://longanwang-cs.github.io/gsvc/.
Music audio generation spans instrumental clips, full songs with vocals, edits of existing recordings, and systems that play alongside a musician in real time. This monograph explains the methods behind these systems for readers familiar with neural-network training who are new to music generation. It introduces neural audio codecs and autoencoders, autoregressive and masked token models, diffusion and flow matching, sampling acceleration, and preference optimization in a shared notation. Worked examples connect training objectives and sampling procedures to published systems. The application sections relate the musical requirements of instrumental generation, songs, editing, and live performance to model design. Comparison tables record representations, methods, inputs, tasks, and release information documented in papers and official sources. Symbolic composition, speech, and sound effects are included where they connect to music audio. We explain how to interpret automatic metrics and learned judges, design and report listening tests, and assess evidence about data provenance, memorization, attribution, and output detection. Selected legal decisions and disputes illustrate the implications for training and release. The reviewed systems combine established generative paradigms with new representations, planning stages, and post-training procedures. Differences in data, scale, and evaluation limit what cross-system comparisons can establish about the sources of reported gains.
Automatic soccer event spotting supports highlight generation, video indexing, and sports analytics, but existing methods depend on large annotated datasets and soccer-specific training. We present a zero-shot framework that detects, classifies, and temporally localizes events in full-length soccer broadcasts without task-specific parameter training. Each match is segmented into overlapping clips represented by a compact visual storyboard and the aligned commentary transcript, which a general-purpose multimodal model processes jointly. On three manually annotated matches containing 130 events, the best configuration achieves 72.0% F1@5s and outperforms SoccerNet-trained CALF and NetVLAD++ in a cross-dataset setting, while the storyboard representation reduces inference cost by 72% compared with individual frames. We further analyze the effects of model family, input representation, and prompt optimization, and introduce an evidence-grounded excitement-scoring stage that ranks detected events with natural-language explanations for highlight selection.
Image editing models can modify visual content through natural language, but editing dense documents requires precise local changes amid tightly coupled text, tables, and layout. We introduce VDE Bench, an evaluation resource for studying instruction-guided editing of Chinese and English document images. VDE Bench contains 3,506 instruction-guided image editing samples, including 1,578 text-region edits and 1,928 table-region edits. The benchmark pairs editing instructions and reference images with annotated editing regions and structured OCR outputs. Its evaluation protocol combines region-level and full-page measurements to assess textual correctness, structural fidelity, and visual content preservation. By focusing on the interaction between local edits and their surrounding document context, VDE Bench complements existing visual text editing benchmarks and provides a resource for diagnosing document-specific editing failures.
Guowei Zhong, Junjie Li, Huaiyu Zhu, Ruohong Huan, Yun Pan
Multimodal sentiment analysis (MSA) has made considerable progress in recent years. Existing MSA methods mainly learn shared multimodal representations or train unimodal predictors with a common multimodal label. However, assigning the shared label to every modality can obscure modality-specific sentiment, while direct fusion may overemphasize unreliable predictions when modalities disagree. To address these issues, we propose Uncertainty-Guided Multimodal Consensus (UMC), which first recovers continuous unimodal pseudo-labels and then forms a reliability-aware consensus across modalities. Our method consists of two key components: i) Continuous Prototype-Guided Label Recovery (CPLR), which combines heteroscedastic predictions with ordered sentiment prototypes and uses confidence and label conflict to refine continuous unimodal pseudo-labels, and ii) Reliability and Precision Consensus (RPC), which estimates sample-dependent modality reliability from predictive statistics and cross-modal disagreement and combines it with predictive precision through a closed-form consensus solver. Extensive experiments conducted on four publicly available MSA datasets demonstrate the effectiveness and superiority of our proposed model.
This paper explores whether large language models (LLMs) can design video coding tools, a highly challenging task due to the intricate algorithmic coupling of tool modifications. In particular, we present an empirical case study on the Planar mode, a long-standing intra prediction tool in video coding standards. Our experiments operate within a generation-and-evaluation loop, with the LLM generating new Planar predictors, encoder trials evaluating their coding performance, and the LLM re-generating refined implementations based on the evaluation feedback. We first examine directly replacing the default Planar mode in the Fraunhofer Versatile Video Encoder (VVenC) under its faster preset. Experimental results demonstrate that the LLM-generated mode can outperform the conventional Planar mode on this lightweight toolset, achieving 0.18% bitrate savings with 0.4% complexity overhead on the standard benchmark. We further extend our evaluation to the Enhanced Compression Model (ECM). Leveraging newly introduced directional Planar modes, we investigate two integration strategies: directly replacing them, and introducing the LLM-generated predictor as an additional prediction mode with new syntax elements. The empirical results suggest that both strategies can yield coding gains under a constrained low-resolution setting. Overall, this study offers preliminary evidence and practical insights, highlighting both the potential and open challenges of LLM-based coding tool design.
Video models answer from sampled frames and may miss the decisive moment. A common remedy is to commit to an answer, re-examine the video, and confirm or revise it, a loop now mainly learned with reinforcement learning (RL). We find that RL erodes the initial answer, a phenomenon we term confidence erosion, where confidence means the label-verified reliability of the initial answer instead of self-reported certainty: because errors can be repaired later, a correct initial answer earns less reward, and the initial answer stalls even as the final answer's gain over it rises. We propose the principle confident first, reflective second: the initial answer must stay reliable on its own, above a floor set by a reference policy, and reflection may only add to it. Our method, CoFire, implements this principle with one mechanism for each erosion channel: reward incentives, parameter drift, and gradient interference. CoFire reaches 67.0% mean accuracy on Video-MME, MLVU, and LongVideoBench, 1.7 points above the strongest compared 7B method, and improves initial-answer and final accuracy over outcome-reward RL by 4.1 and 3.7 points on average: a stronger initial answer carries through to a stronger final one.
Video stylization typically transforms entire scenes, while local artistic editing requires precise region--style association, spatial control, and temporal consistency. We introduce LociArt, a unified paradigm for free-style local video stylization that explicitly specifies where to stylize and what style to apply while preserving non-target content. At its core, GRNStyle builds on Generative Refinement Networks (GRN) and introduces a polar style-aware mask that embeds style identities as equal-norm directions within selected regions. Complementary representation--geometry conditioning combines encoded mask features with raw-mask spatial support, integrating spatiotemporal context with geometric control. To provide paired supervision, we develop an automated pipeline combining global stylization, video object removal, matting, and compositing, and construct the LociArt-100K Dataset with 100K local stylization pairs. We further establish the LociArt Benchmark, comprising 80 source videos and 84 target masks covering 168 foreground/background stylization cases, with a Style Prototype Similarity (SPS) evaluation scheme for regional style quality and complementary measures of non-target preservation and temporal consistency. Experiments show that GRNStyle outperforms existing methods in regional style quality and spatial control while maintaining strong temporal consistency. Qualitative examples further illustrate joint stylization of multiple regions with the same or different learned styles, despite training only on single-region, single-style pairs. Together, these components provide a unified foundation for learning and evaluating controllable local video stylization. Project page: https://higherhu.github.io/lociart/.
Wang Tang, Fethiye Irmak Dogan, Linbo Qing, Ce Zhu, Hatice Gunes
Dyadic relationship recognition aims to identify the latent social relationship between two individuals from their multimodal interactions. Existing methods typically project heterogeneous cues into a shared representation and optimise a flat classification objective, overlooking that relationships may rely on different interaction patterns. For instance, distant relationships can often be inferred from direct behavioural correspondence, whereas close relationships are better reflected by coordinated affective, behavioural, and multimodal responses. To address this limitation, we propose HiRe-DRR, a Hierarchical Multimodal Interaction Reasoning framework that organises relationship evidence from three complementary perspectives: feature level, interaction type, and relationship hierarchy. HiRe-DRR decomposes facial, bodily, acoustic, and textual cues into low- and high-level representations and constructs homogeneous and heterogeneous cross-participant interactions. We further introduce the Mixture-of-Interaction Transformer (MoI-Former), which treats diverse interactions as specialised experts and employs relation-aware sparse routing to identify relationship-relevant evidence while preserving contextual interaction information. The proposed Hierarchical Conditional Loss (HC-Loss) exploits the hierarchical nature of social relationships through coarse-to-fine supervision, enabling structured relationship reasoning. Extensive experiments on three dyadic relationship benchmarks demonstrate that HiRe-DRR outperforms recent representative image-, video-, and multimodal methods. Our routing analysis further reveals that distant relationships tend to rely on low-level homogeneous interactions, whereas closer relationships focus on high-level heterogeneous interactions. Our code is publicly available at: https://github.com/tw-repository/HiRe-DRR.
Lennart Schiweck, Adrian Kempf, Michelle Retig, Moritz Feucht, Tim Mangliers, Benjamin Himpel
Editing multi-camera concerts requires coordinated decisions about musical relevance, visual quality and cut timing. We present a modular framework that combines instrument activity from multitrack audio, visible instruments and image sharpness with beat-based timing and probabilistic camera selection. An interpretable scoring model links these components to the choice of views over complete shot intervals. Component tests and an exploratory expert study on a four-camera concert yield mixed assessments across the editing variants. The findings motivate extending local audiovisual relevance with contextual and sequence-level criteria. The architecture provides a basis for testing these extensions and studying how component behaviour affects the quality of complete edits.
Head-mounted mixed reality (MR) often prioritises individual immersion, potentially drawing attention away from others sharing the room. This paper asks how MR can support co-located spatial dialogue through embodied interaction and activity visible to non-headset observers. We introduce Immersive Commons, a design framework, and Atmos Messages, a prototype combining gesture-based, ephemeral spatial messages with voice transcription, preset text, and a public floor-plan overview. Following a Research through Design approach, fifteen participants created 236 messages across two sessions, including 48 voice-authored messages. Questionnaire responses highlighted playfulness and social awareness, while observations suggested opportunities for exchange grounded in the architectural setting. Interaction clustered near the public display, although this was also where participants put on their headsets. These findings suggest the potential of ephemeral spatial messaging to support co-located exchange, with playfulness, shared visibility, and architectural context informing the design of more social and situated MR experiences.
Mansi Sharma, Dinar Orazgaliyev, Antonio Krüger, Philipp Müller
Visual search is ubiquitous in daily life, from locating keys in a messy apartment to finding a tool in a cluttered workshop or an application icon on a desktop screen. While visual attention is to a large extent driven by image content, research in psychology indicates that user cognitive states have a significant influence on attention allocation. A promising way to access information about cognitive states is via Electroencephalography (EEG) recordings. Despite this potential, previous works in computational attention modeling have largely disregarded EEG, and related studies in psychology operate with abstract, simplistic stimuli, or even artificially guided search procedures. We bridge this gap with two contributions. First, we present EEGaze 140, the largest publicly available multimodal dataset of free visual search with realistic stimuli. The dataset contains synchronized EEG and eye tracking recordings from 36 participants across 140 naturalistic scenes, covering two scenarios: searching for desktop icons and finding tools in cluttered workshop scenes. Second, we introduce EEGazeNet, the first method for human attention prediction in visual search that integrates EEG signals. Our approach consists of a EEG guided refinement stage that adapts the output of a state of the art saliency prediction method to the user current cognitive state. In comprehensive evaluations on EEGaze 140, we show that EEGazeNet consistently outperforms all baselines and ablation conditions.
Aman Tyagi, Hemanth Boinpally, Jonathan Chen, Douglas Gebert, Steven Hickson
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.
Pretrained image-generation models have been successfully repurposed for various downstream visual tasks. Whether their learned representations and generative priors can transfer beyond vision to other modalities remains largely unresolved, particularly for speech synthesis, which requires linguistic fidelity and preservation of speaker identity from a short reference. We present ImageTTS, which repurposes FLUX.2 for speech synthesis by encoding pseudo-RGB log-mel spectrograms with its frozen image VAE and adapting only its pretrained Diffusion Transformer (DiT) using conditional latent flow matching. This formulation disentangles two forms of vision-to-speech transfer: representation transfer through the image VAE and generative-prior transfer through DiT initialization. The frozen image VAE largely preserves intelligibility and speaker identity with 0.99 STOI and 0.95 speaker similarity, while image-pretrained DiT initialization accelerates convergence and improves linguistic accuracy, reducing WER from 4.2\% to 2.7\% at 100k matched updates relative to random initialization. The same framework supports both zero-shot and instruction-guided TTS and, when scaled to 100K hours of speech data, achieves 1.12\% WER on Seed-TTS test-en and a Style-ACC of 0.66 on CapSpeech, outperforming speech-native baselines trained on comparable data across both tasks. These results suggest that pretrained image models can transfer beyond architectural reuse, supplying useful latent representations and generative priors for intelligible and controllable speech synthesis.
Museum audiovisual archives are important records of visual culture, yet in-gallery video remains difficult to search because it lacks catalogue-compatible metadata. We propose CaGMulA, catalogue-grounded multimodal attribution for museum video: a locally deployable pipeline that combines video-language models, structured catalogue retrieval, and abstention-aware decision policies. The system summarises visible artworks, generates catalogue-style descriptions, and attributes title and artist only when evidence is sufficient. We evaluate VideoLLaMA2 and Qwen2-VL across four model configurations and four attribution modes on 18 in-gallery videos. Open-ended generation fails under conservative conditions, producing 0\% coverage across models. By contrast, closed-set catalogue retrieval reaches 100\% coverage with 62\% accepted title--artist correctness, while a verifier-gated hybrid mode reduces coverage to 50\% but improves accepted correctness to 78\%. The results show that reliable museum-video attribution is driven less by unrestricted generation than by catalogue grounding, visual retrieval, and selective abstention for curatorial review.
Speech interaction systems have gradually evolved from traditional cascaded systems that use text as intermediate information to end-to-end speech language models that directly process speech representations. In some applications, the speech frontend runs on the user's device and sends speech representations to a remote server for inference, while the original audio remains local. However, these representations may still contain speaker identity information. If an attacker obtains them, it may use this information to imitate the user's voice, threatening their voice privacy. To investigate this risk, we propose \stia, a Speech Token Inversion Attack that generates speech with attacker-specified text in a victim's voice using a single short leaked representation, without access to the original audio or its transcript. Specifically, \stia converts discrete, continuous, or mixed speech representations into conditioning sequences and trains a conditional flow-matching speech generation model using pairs of different utterances from the same speaker in public speech datasets. The model combines speaker characteristics contained in the leaked representation with the specified text to generate speech in the victim's voice. Experiments show that, with a Moshi representation leak corresponding to 3 seconds of speech, \stia achieves an average CAM++ speaker cosine similarity of 0.78 and a target-text word error rate of 3.80\%. These results show that speech frontend representations reveal reusable speaker identity information and highlight the need to consider privacy risks when designing speech frontend representation methods.
Co-Director: Agentic Generative Video Storytelling
Yale Song, Yiwen Song, Nick Losier, Nathan Hodson, Ye Jin, Rhyard Zhu, Yan Xu, Daniel Vlasic 等 16 位作者
AI 速览：提出分层多智能体框架 Co-Director，将视频叙事生成形式化为全局优化问题，以缓解链式生成中的语义漂移与级联失败。
While diffusion models generate high-fidelity video clips, transforming them into coherent storytelling engines remains challenging. Current agentic pipelines automate this via chained modules but suffer from semantic drift and cascading failures due to independent, handcrafted prompting. We present Co-Director, a hierarchical multi-agent framework formalizing video storytelling as a global optimization problem. To ensure semantic coherence, we introduce hierarchical parameterization: a multi-armed bandit globally identifies promising creative directions, while a local multimodal self-refinement loop mitigates identity drift and ensures sequence-level consistency. This balances the exploration of novel narrative strategies with the exploitation of effective creative configurations. For evaluation, we introduce GenAD-Bench, a 400-scenario dataset of fictional products for personalized advertising. Experiments demonstrate that Co-Director significantly outperforms state-of-the-art baselines, offering a principled approach that seamlessly generalizes to broader cinematic narratives. Project Page: https://co-director-agent.github.io/