Jaeyeon Kim

Ka Chun Shum, Jaeyeon Kim, Binh-Son Hua et al.

Neural radiance field is an emerging rendering method that generates high-quality multi-view consistent images from a neural scene representation and volume rendering. Although neural radiance field-based techniques are robust for scene reconstruction, their ability to add or remove objects remains limited. This paper proposes a new language-driven approach for object manipulation with neural radiance fields through dataset updates. Specifically, to insert a new foreground object represented by a set of multi-view images into a background radiance field, we use a text-to-image diffusion model to learn and generate combined images that fuse the object of interest into the given background across views. These combined images are then used for refining the background radiance field so that we can render view-consistent images containing both the object and the background. To ensure view consistency, we propose a dataset updates strategy that prioritizes radiance field training with camera views close to the already-trained views prior to propagating the training to remaining views. We show that under the same dataset updates strategy, we can easily adapt our method for object insertion using data from text-to-3D models as well as object removal. Experimental results show that our method generates photorealistic images of the edited scenes, and outperforms state-of-the-art methods in 3D reconstruction and neural radiance field blending.

Jewon Lee, Wooksu Shin, Seungmin Yang et al.

Efficient processing of high-resolution images is crucial for real-world vision-language applications. However, existing Large Vision-Language Models (LVLMs) incur substantial computational overhead due to the large number of vision tokens. With the advent of "thinking with images" models, reasoning now extends beyond text to the visual domain. This capability motivates our two-stage "coarse-to-fine" reasoning pipeline: first, a downsampled image is analyzed to identify task-relevant regions; then, only these regions are cropped at full resolution and processed in a subsequent reasoning stage. This approach reduces computational cost while preserving fine-grained visual details where necessary. A major challenge lies in inferring which regions are truly relevant to a given query. Recent related methods often fail in the first stage after input-image downsampling, due to perception-driven reasoning, where clear visual information is required for effective reasoning. To address this issue, we propose ERGO (Efficient Reasoning & Guided Observation) that performs reasoning-driven perception-leveraging multimodal context to determine where to focus. Our model can account for perceptual uncertainty, expanding the cropped region to cover visually ambiguous areas for answering questions. To this end, we develop simple yet effective reward components in a reinforcement learning framework for coarse-to-fine perception. Across multiple datasets, our approach delivers higher accuracy than the original model and competitive methods, with greater efficiency. For instance, ERGO surpasses Qwen2.5-VL-7B on the V* benchmark by 4.7 points while using only 23% of the vision tokens, achieving a 3x inference speedup. The code and models can be found at: https://github.com/nota-github/ERGO.

Gaspard Beaudouin, Minghan Li, Jaeyeon Kim et al.

We propose Delta Velocity Rectified Flow (DVRF), a novel inversion-free, path-aware editing framework within rectified flow models for text-to-image editing. DVRF is a distillation-based method that explicitly models the discrepancy between the source and target velocity fields in order to mitigate over-smoothing artifacts rampant in prior distillation sampling approaches. We further introduce a time-dependent shift term to push noisy latents closer to the target trajectory, enhancing the alignment with the target distribution. We theoretically demonstrate that when this shift is disabled, DVRF reduces to Delta Denoising Score, thereby bridging score-based diffusion optimization and velocity-based rectified-flow optimization. Moreover, when the shift term follows a linear schedule under rectified-flow dynamics, DVRF generalizes the Inversion-free method FlowEdit and provides a principled theoretical interpretation for it. Experimental results indicate that DVRF achieves superior editing quality, fidelity, and controllability while requiring no architectural modifications, making it efficient and broadly applicable to text-to-image editing tasks. Code is available at https://github.com/Harvard-AI-and-Robotics-Lab/DeltaVelocityRectifiedFlow.

Jaeyeon Kim, Heeseung Yun, Gunhee Kim

Spatial audio is essential for enhancing the immersiveness of audio-visual experiences, yet its production typically demands complex recording systems and specialized expertise. In this work, we address a novel problem of generating first-order ambisonics, a widely used spatial audio format, directly from silent videos. To support this task, we introduce YT-Ambigen, a dataset comprising 102K 5-second YouTube video clips paired with corresponding first-order ambisonics. We also propose new evaluation metrics to assess the spatial aspect of generated audio based on audio energy maps and saliency metrics. Furthermore, we present Video-to-Spatial Audio Generation (ViSAGe), an end-to-end framework that generates first-order ambisonics from silent video frames by leveraging CLIP visual features, autoregressive neural audio codec modeling with both directional and visual guidance. Experimental results demonstrate that ViSAGe produces plausible and coherent first-order ambisonics, outperforming two-stage approaches consisting of video-to-audio generation and audio spatialization. Qualitative examples further illustrate that ViSAGe generates temporally aligned high-quality spatial audio that adapts to viewpoint changes.

Chao-Han Huck Yang, Sreyan Ghosh, Qing Wang et al.

We present Task 5 of the DCASE 2025 Challenge: an Audio Question Answering (AQA) benchmark spanning multiple domains of sound understanding. This task defines three QA subsets (Bioacoustics, Temporal Soundscapes, and Complex QA) to test audio-language models on interactive question-answering over diverse acoustic scenes. We describe the dataset composition (from marine mammal calls to soundscapes and complex real-world clips), the evaluation protocol (top-1 accuracy with answer-shuffling robustness), and baseline systems (Qwen2-Audio-7B, AudioFlamingo 2, Gemini-2-Flash). Preliminary results on the development set are compared, showing strong variation across models and subsets. This challenge aims to advance the audio understanding and reasoning capabilities of audio-language models toward human-level acuity, which are crucial for enabling AI agents to perceive and interact about the world effectively.

Jaeyeon Kim, Seunggeun Kim, Taekyun Lee et al.

A natural desideratum for generative models is self-correction--detecting and revising low-quality tokens at inference. While Masked Diffusion Models (MDMs) have emerged as a promising approach for generative modeling in discrete spaces, their capacity for self-correction remains poorly understood. Prior attempts to incorporate self-correction into MDMs either require overhauling MDM architectures/training or rely on imprecise proxies for token quality, limiting their applicability. Motivated by this, we introduce PRISM--Plug-in Remasking for Inference-time Self-correction of Masked Diffusions--a lightweight, model-agnostic approach that applies to any pretrained MDM. Theoretically, PRISM defines a self-correction loss that provably learns per-token quality scores, without RL or a verifier. These quality scores are computed in the same forward pass with MDM and used to detect low-quality tokens. Empirically, PRISM advances MDM inference across domains and scales: Sudoku; unconditional text (170M); and code with LLaDA (8B).

Jaeyeon Kim, Binh-Son Hua, Duc Thanh Nguyen et al.

With recent developments of convolutional neural networks, deep learning for 3D point clouds has shown significant progress in various 3D scene understanding tasks, e.g., object recognition, semantic segmentation. In a safety-critical environment, it is however not well understood how such deep learning models are vulnerable to adversarial examples. In this work, we explore adversarial attacks for point cloud-based neural networks. We propose a unified formulation for adversarial point cloud generation that can generalise two different attack strategies. Our method generates adversarial examples by attacking the classification ability of point cloud-based networks while considering the perceptibility of the examples and ensuring the minimal level of point manipulations. Experimental results show that our method achieves the state-of-the-art performance with higher than 89% and 90% of attack success rate on synthetic and real-world data respectively, while manipulating only about 4% of the total points.