Premise-based Multimodal Reasoning: Conditional Inference on Joint Textual and Visual Clues
원문 초록
It is a common practice for recent works in vision language cross-modal reasoning to adopt a binary or multi-choice classification formulation taking as input a set of source image(s) and textual query. In this work, we take a sober look at such an unconditional formulation in the sense that no prior knowledge is specified with respect to the source image(s). Inspired by the designs of both visual commonsense reasoning and natural language inference tasks, we propose a new task termed Premise-based Multi-modal Reasoning(PMR) where a textual premise is the background presumption on each source image. The PMR dataset contains 15,360 manually annotated samples which are created by a multi-phase crowd-sourcing process. With selected high-quality movie screenshots and human-curated premise templates from 6 pre-defined categories, we ask crowd-source workers to write one true hypothesis and three distractors (4 choices) given the premise and image through a cross-check procedure. Besides, we generate adversarial samples to alleviate the annotation artifacts and double the size of PMR. We benchmark various state-of-the-art (pretrained) multi-modal inference models on PMR and conduct comprehensive experimental analyses to showcase the utility of our dataset.
함께 살펴볼 논문
공통 태그를 근거로 연결했습니다. 실제 인용 관계나 논문의 우수성을 뜻하지 않습니다.
Preemptive Hallucination Reduction: An Input-Level Approach for Multimodal Language Model
Skywork-R1V3 Technical Report
DS@GT at CheckThat! 2025: Evaluating Context and Tokenization Strategies for Numerical Fact Verification
DreamVLA: A Vision-Language-Action Model Dreamed with Comprehensive World Knowledge
GLM-4.1V-Thinking: Towards Versatile Multimodal Reasoning with Scalable Reinforcement Learning
Thinking with Images for Multimodal Reasoning: Foundations, Methods, and Future Frontiers
토론
이 논문과 연결된 포럼 글이 아직 없습니다.