The increasing use of transformer-based large language models brings forward the challenge of processing long sequences. In document visual question answering (DocVQA), leading methods focus on the single-page setting, while documents can span hundreds of pages. We present GRAM, a method that seamlessly extends pre-trained single-page models to the multi-page setting, without requiring computationally-heavy pretraining. To do so, we leverage a single-page encoder for local page-level understanding, and enhance it with document-level designated layers and learnable tokens, facilitating the flow of information across pages for global reasoning. To enforce our model to utilize the newly introduced document tokens, we propose a tailored bias adaptation method. For additional computational savings during decoding, we introduce an optional compression stage using our compression-transformer (C-Former),reducing the encoded sequence length, thereby allowing a tradeoff between quality and latency. Extensive experiments showcase GRAM's state-of-the-art performance on the benchmarks for multi-page DocVQA, demonstrating the effectiveness of our approach.
AWM: Answerable Working Memory for Long-Document VQA Agents
University of Oslo(奥斯陆大学);Bosch Center for AI(博世人工智能中心);University of Stuttgart(斯图加特大学);Hong Kong University of Science and Technology (Guangzhou)(香港科技大学(广州));Tsinghua University(清华大学)
BinaryVQA: A Versatile Test Set to Evaluate the Out-of-Distribution Generalization of VQA Models
Quintic AI(昆蒂克人工智能)
我们引入了一个新的测试集的视觉问答(VQA)称为BinaryVQA推动VQA模型的限制。我们的数据集包括1,024张图像中的7,800个问题,涵盖了各种各样的对象、主题和概念。为了简化模型评估,我们只考虑二元问题。问题和答案都是人工精心制定和验证的。大约63%的问题得到了肯定的回答。每个图像的问题数和问题长度的中位数分别为7和5。最先进的OFA模型在BinaryVQA数据集上达到了75%的准确率,这明显低于其在VQA v2测试开发数据集上的性能(94.7%)。我们还沿着几个维度分析了模型行为,包括:a)不同类别的性能,如文本,计数和注视方向,b)模型可解释性,c)问题长度对准确性的影响,d)模型对积极答案的偏见和引入称为ShuffleAcc的新分数,以及e)对拼写和语法错误的敏感性。我们的调查证明了我们的数据集的难度,并表明它可以在未来几年内挑战VQA模型。数据和代码可在以下网址公开获取:DATA and CODE。
Doc-V*:Coarse-to-Fine Interactive Visual Reasoning for Multi-Page Document VQA
School of Software Engineering, Huazhong University of Science and Technology(华中科技大学软件工程学院);MiLM Plus, Xiaomi Inc.(小米公司MiLM Plus团队);School of Computer Science and Technology, Huazhong University of Science and Technology(华中科技大学计算机科学与技术学院);School of Data Science, Fudan University(复旦大学数据科学学院)
Quantifying and Mitigating Unimodal Biases in Multimodal Large Language Models: A Causal Perspective
Shanghai AI Laboratory(上海人工智能实验室);State Key Laboratory of General Artificial Intelligence, Peking University(北京大学通用人工智能国家重点实验室);School of Intelligence Science and Technology, Peking University(北京大学智能科学与技术学院);School of Computer Science, Fudan University(复旦大学计算机科学技术学院);Shanghai Artificial Intelligence Laboratory(上海人工智能实验室)
MMSI-Bench: A Benchmark for Multi-Image Spatial Intelligence
Shanghai AI Laboratory(上海人工智能实验室);The Chinese University of Hong Kong(香港中文大学);Zhejiang University(浙江大学);Tsinghua University(清华大学);Shanghai Jiaotong University(上海交通大学);University of Hong Kong(香港大学);Beijing Normal University(北京师范大学)
MicroVQA: A Multimodal Reasoning Benchmark for Microscopy-Based Scientific Research
Stanford University(斯坦福大学);Tsinghua University(清华大学);University of North Carolina at Chapel Hill(北卡罗来纳大学教堂山分校);Princeton University(普林斯顿大学);KTH Royal Institute of Technology(瑞典皇家理工学院);Chan Zuckerberg Biohub Network(陈·扎克翰生物中心网络)
Capturing Global Informativeness in Open Domain Keyphrase Extraction
Department of Electronic Engineering, Tsinghua University(清华大学电子工程系);Microsoft Research(微软研究院);Department of Computer Science and Technology, Tsinghua University(清华大学计算机科学与技术系);Institute for Artificial Intelligence, Tsinghua University(清华大学人工智能研究院);Department of Computer Science and Technology, Northeastern University(东北大学计算机科学与技术系)