48 papers
This paper introduces the first formal framework for Audio-Visual World Models (AVWM), presenting the AVW-4k dataset and the AV-CDiT model to enable high-fidelity, synchronized simulation of binaural audio and visual dynamics that significantly enhances agent performance in continuous navigation tasks.
This paper introduces DynHiL-EQA, a new dataset for dynamic human-in-the-loop Embodied Question Answering, and proposes DIVRR, a training-free framework that enhances robustness and inference efficiency by refining ambiguous views and selectively managing memory in dynamic environments.
This paper proposes a large language model-driven method for generating dynamic, semantically aligned speech and gestures for pedagogical agents in virtual reality, demonstrating through user experience experiments that such multimodal expressions significantly enhance learning effectiveness, engagement, and social presence while reducing fatigue and boredom.
The paper introduces MORE-R1, a novel Large Vision-Language Model that leverages a two-stage training process combining Supervised Fine-Tuning on automatically constructed stepwise reasoning data and Reinforcement Learning with Group Relative Policy Optimization to achieve state-of-the-art performance in Multimodal Object-Entity Relation Extraction.
This study systematically investigates how network latency degrades the pragmatic and hedonic dimensions of Quality of Experience in networked VR whiteboards by analyzing their varying impacts across sequential and free collaboration modes, as well as between avatar-enhanced and traditional VR or PC-based systems.
This paper proposes the Task-Aware Perceptual Interaction Fluency Model (TPIFM), which evaluates the perceived responsiveness of remote collaborative augmented reality by integrating network impairments with task-specific just-noticeable differences derived from the Free Energy Principle to guide adaptive system design.
This paper investigates how end-to-end latency impacts interaction fluency and social presence in VR conferencing compared to traditional video conferencing through subjective experiments, aiming to clarify the underlying perceptual and cognitive mechanisms to guide system optimization.
This paper proposes a noise-conditioned mixture-of-experts framework for robust speaker verification that decomposes the feature space into specialized noise-aware subspaces using a routing mechanism, expert specialization strategy, and SNR-decaying curriculum learning to outperform conventional unified representation methods under diverse noisy conditions.
The MEGC 2026 challenge introduces two new tasks, Micro-Expression Video Question Answering (ME-VQA) and Micro-Expression Long-Video Question Answering (ME-LVQA), to advance the analysis of facial micro-expressions by leveraging the multimodal reasoning capabilities of large vision-language models on both short and long-duration video sequences.
This paper introduces CDGLT, a training-efficient framework for multimodal metaphor identification that leverages Concept Drift via Spherical Linear Interpolation and adapted LayerNorm tuning to achieve state-of-the-art performance on the MET-Meme benchmark while significantly reducing computational costs compared to existing generative methods.
This paper proposes and evaluates a novel educational approach that uses AI-generated singing and virtual avatars to transform traditional text-based syllabi into engaging audiovisual performances, demonstrating that this method significantly improves student awareness and recall of critical course information.
The paper introduces VoxEmo, a comprehensive benchmark and toolkit for evaluating Speech Large Language Models on speech emotion recognition across 35 corpora and 15 languages, featuring a distribution-aware soft-label protocol that reveals how these models uniquely align with human subjective emotion distributions despite trailing supervised baselines in hard-label accuracy.
The paper proposes PiVOT, a visual prompting mechanism that leverages a pretrained CLIP foundation model to automatically generate and refine online visual prompts, thereby enhancing generic object tracking by effectively suppressing distractors through contrastive guidance.
This paper presents and evaluates a scalable dynamic point cloud streaming system that leverages on-the-fly transcoding, demonstrating how caching and speculative transcoding significantly reduce server loads and improve user Quality of Experience to support a higher number of simultaneous clients.
This paper proposes a novel spectrogram-based Convolutional Neural Network (CNN) approach for multilabel environmental sound classification that significantly outperforms traditional MFCC-based methods on the South Asian SAS-KIIT and UrbanSound8K datasets, offering a more robust solution for complex, overlapping acoustic environments.
This paper proposes a two-stage cascaded framework that generates controllable complex human motion videos by first using an autoregressive model to synthesize 2D skeleton sequences from text descriptions and then employing a pose-conditioned diffusion model with adaptive layer fusion to render high-fidelity videos, supported by a new synthetic dataset designed to overcome limitations in existing benchmarks.
The paper introduces CONSTANT, a novel one-shot handwriting generation framework that leverages Style-Aware Quantization and a latent patch-based contrastive objective within a diffusion model to overcome existing limitations in capturing diverse writer styles and generating high-quality, realistic handwritten images across multiple languages.
This paper introduces Task 5 of the DCASE 2025 Challenge, a multi-domain Audio Question Answering benchmark designed to evaluate and advance the acoustic reasoning capabilities of audio-language models across diverse scenarios including bioacoustics, temporal soundscapes, and complex real-world clips.
This paper introduces TimeSpot, a comprehensive benchmark comprising 1,455 real-world images from 80 countries designed to evaluate the limited geo-temporal reasoning capabilities of current vision-language models in predicting location, time, and environmental context from visual evidence alone.
The paper introduces Emotion Collider (EC-Net), a hyperbolic hypergraph framework that leverages Poincaré-ball embeddings, bidirectional message passing, and contrastive learning to achieve robust and noise-resilient multimodal sentiment analysis by preserving high-order semantic relations and enhancing class separation.