paper-with-me

홈 › Papers

Zero-shot 3D Question Answering via Voxel-based Dynamic Token Compression

2025-01-01 · CVPR 2025 1 · Hsiang-Wei Huang, Fu-Chen Chen, Wenhao Chai, Che-Chun Su, Lu Xia, Sanghun Jung, Cheng-Yen Yang, Jenq-Neng Hwang, Min Sun, Cheng-Hao Kuo

Recent advancements in 3D Large Multi-modal Models (3D-LMMs) have driven significant progress in 3D question answering. However, recent multi-frame Vision-Language Models (VLMs) demonstrate superior performance compared to 3D-LMMs on 3D question answering tasks, largely due to the greater scale and diversity of available 2D image data in contrast to the more limited 3D data. Multi-frame VLMs, although achieving superior performance, suffer from the difficulty of retaining all the detailed visual information in the 3D scene while limiting the number of visual tokens. Common methods such as token pooling, reduce visual token usage but often lead to information loss, impairing the model's ability to preserve visual details essential for 3D question answering tasks. To address this, we propose voxel-based Dynamic Token Compression (DTC), which combines 3D spatial priors and visual semantics to achieve over 90% reduction in visual tokens usage for current multi-frame VLMs. Our method maintains performance comparable to state-of-the-art models on 3D question answering benchmarks including OpenEQA and ScanQA, demonstrating its effectiveness.

📄 PDF Abstract BibTeX

Code (0)

등록된 구현이 없습니다.

Tasks

Question Answering

Similar Papers 제목 키워드 기반

DocPrompt: Large-scale continue pretrain for zero-shot and few-shot document question answering

2023-08-21 · Sijin Wu, Dan Zhang, Teng Hu, Shikun Feng

In this paper, we propose Docprompt for document question answering tasks with powerful zero-shot and few-shot performance. We proposed a novel weakly supervised data generation method, a novel multl-stage training metho…

Question Answering

Zero-Shot Visual Question Answering

2016-11-17 · Damien Teney, Anton Van Den Hengel

Part of the appeal of Visual Question Answering (VQA) is its promise to answer new questions about previously unseen images. Most current methods demand training questions that illustrate every possible concept, and will…

Question AnsweringRetrievalVisual Question AnsweringVisual Question Answering (VQA)+1

DynRank: Improving Passage Retrieval with Dynamic Zero-Shot Prompting Based on Question Classification

2024-11-30 · Abdelrahman Abdallah, Jamshid Mozafari, Bhawna Piryani, Mohammed M. Abdelgwad 외

This paper presents DynRank, a novel framework for enhancing passage retrieval in open-domain question-answering systems through dynamic zero-shot question classification. Traditional approaches rely on static prompts an…

Open-Domain Question AnsweringPassage RetrievalQuestion AnsweringRetrieval

Harnessing Adaptive Topology Representations for Zero-Shot Graph Question Answering

2025-08-08 · Yanbin Wei, Jiangyue Yan, Chun Kang, Yang Chen 외 arxiv

Large Multimodal Models (LMMs) have shown generalized zero-shot capabilities in diverse domain question-answering (QA) tasks, including graph QA that involves complex graph topologies. However, most current approaches us…

Graph Question Answering

QADYNAMICS: Training Dynamics-Driven Synthetic QA Diagnostic for Zero-Shot Commonsense Question Answering

2023-10-17 · Haochen Shi, Weiqi Wang, Tianqing Fang, Baixuan Xu 외

Zero-shot commonsense Question-Answering (QA) requires models to reason about general situations beyond specific benchmarks. State-of-the-art approaches fine-tune language models on QA pairs constructed from CommonSense …

DiagnosticQuestion Answering