Vision-and-language reasoning requires an understanding of visual concepts, language semantics, and, most importantly, the alignment and relationships between these two modalities. We thus propose the LXMERT (Learning Cross-Modality Encoder Representations from Transformers) framework to learn these vision-and-language connections. In LXMERT, we build a large-scale Transformer model that consists of three encoders: an object relationship encoder, a language encoder, and a cross-modality encoder. Next, to endow our model with the capability of connecting vision and language semantics, we pre-train the model with large amounts of image-and-sentence pairs, via five diverse representative pre-training tasks: masked language modeling, masked object prediction (feature regression and label classification), cross-modality matching, and image question answering. These tasks help in learning both intra-modality and cross-modality relationships. After fine-tuning from our pre-trained parameters, our model achieves the state-of-the-art results on two visual question answering datasets (i.e., VQA and GQA). We also show the generalizability of our pre-trained cross-modality model by adapting it to a challenging visual-reasoning task, NLVR2, and improve the previous best result by 22% absolute (54% to 76%). Lastly, we demonstrate detailed ablation studies to prove that both our novel model components and pre-training strategies significantly contribute to our strong results; and also present several attention visualizations for the different encoders. Code and pre-trained models publicly available at: https://github.com/airsplay/lxmert
Beijing University of Posts and Telecommunication（北京邮电大学）；Rutgers University（罗格斯大学）；The Chinese University of Hong Kong（香港中文大学）；Chinese Academy of Sciences（中国科学院）；Hasso Plattner Institute（哈索·普拉特纳研究院）
Several multi-modality representation learning approaches such as LXMERT and ViLBERT have been proposed recently. Such approaches can achieve superior performance due to the high-level semantic information captured during large-scale multimodal pretraining. However, as ViLBERT and LXMERT adopt visual region regression and classification loss, they often suffer from domain gap and noisy label problems, based on the visual features having been pretrained on the Visual Genome dataset. To overcome these issues, we propose unbiased Contrastive Visual-Linguistic Pretraining (CVLP), which constructs …
We present a unified Vision-Language pretrained Model (VLMo) that jointly learns a dual encoder and a fusion encoder with a modular Transformer network. Specifically, we introduce Mixture-of-Modality-Experts (MoME) Transformer, where each block contains a pool of modality-specific experts and a shared self-attention layer. Because of the modeling flexibility of MoME, pretrained VLMo can be fine-tuned as a fusion encoder for vision-language classification tasks, or used as a dual encoder for efficient image-text retrieval. Moreover, we propose a stagewise pre-training strategy, which effectivel…
University of Malta, Institute of Linguistics and Language Technology（马耳他大学语言学与语言技术研究所）；Universiteit Utrecht, Information and Computing Sciences（乌得勒支大学信息与计算科学系）
Images can be described in terms of the objects they contain, or in terms of the types of scene or place that they instantiate. In this paper we address to what extent pretrained Vision and Language models can learn to align descriptions of both types with images. We compare 3 state-of-the-art models, VisualBERT, LXMERT and CLIP. We find that (i) V&L models are susceptible to stylistic biases acquired during pretraining; (ii) only CLIP performs consistently well on both object- and scene-level descriptions. A follow-up ablation study shows that CLIP uses object-level information in the visual …
3D Modality-Aware Pre-training for Vision-Language Model in MRI Multi-organ Abnormality Detection
School of Electronic, Electrical Engineering and Physics, Fujian University of Technology（福建工程学院电子电气工程学院）；School of Computer Science and Engineering, Southeast University, China（东南大学计算机科学与工程学院）；Department of Medical Imaging, Suzhou Traditional Chinese Medicine Hospital, China（苏州中医医院影像科）
Image captioning model is a cross-modality knowledge discovery task, which targets at automatically describing an image with an informative and coherent sentence. To generate the captions, the previous encoder-decoder frameworks directly forward the visual vectors to the recurrent language model, forcing the recurrent units to generate a sentence based on the visual features. Although these sentences are generally readable, they still suffer from the lack of details and highlights, due to the fact that the substantial gap between the image and text modalities is not sufficiently addressed. In …
Beijing Institution of Technology（北京理工大学）；Baidu Inc.（百度公司）；Inception Institute of Artificial Intelligence（起源人工智能研究院）
Answering semantically-complicated questions according to an image is challenging in Visual Question Answering (VQA) task. Although the image can be well represented by deep learning, the question is always simply embedded and cannot well indicate its meaning. Besides, the visual and textual features have a gap for different modalities, it is difficult to align and utilize the cross-modality information. In this paper, we focus on these two problems and propose a Graph Matching Attention (GMA) network. Firstly, it not only builds graph for the image, but also constructs graph for the question …
An Empirical Study on the Language Modal in Visual Question Answering
Huazhong University of Science and Technology（华中科技大学）；Beijing Institute of Technology（北京理工大学）；Ping An Property & Casualty Insurance Company of China, Ltd（中国平安财产保险股份有限公司）
This paper investigates the effectiveness and implementation of modality-specific large-scale pre-trained encoders for multimodal sentiment analysis~(MSA). Although the effectiveness of pre-trained encoders in various fields has been reported, conventional MSA methods employ them for only linguistic modality, and their application has not been investigated. This paper compares the features yielded by large-scale pre-trained encoders with conventional heuristic features. One each of the largest pre-trained encoders publicly available for each modality are used; CLIP-ViT, WavLM, and BERT for vis…
Recently, opinion summarization, which is the generation of a summary from multiple reviews, has been conducted in a self-supervised manner by considering a sampled review as a pseudo summary. However, non-text data such as image and metadata related to reviews have been considered less often. To use the abundant information contained in non-text data, we propose a self-supervised multimodal opinion summarization framework called MultimodalSum. Our framework obtains a representation of each modality using a separate encoder for each modality, and the text decoder generates a summary. To resolv…
National Laboratory of Pattern Recognition, Institute of Automation, CAS（中国科学院自动化研究所模式识别国家重点实验室）；School of Artificial Intelligence, University of Chinese Academy of Sciences（中国科学院大学人工智能学院）；CAS Center for Excellence in Brain Science and Intelligence Technology（中国科学院脑科学与智能技术卓越创新中心）
End-to-end speech translation aims to translate speech in one language into text in another language via an end-to-end way. Most existing methods employ an encoder-decoder structure with a single encoder to learn acoustic representation and semantic information simultaneously, which ignores the speech-and-text modality differences and makes the encoder overloaded, leading to great difficulty in learning such a model. To address these issues, we propose a Speech-to-Text Adaptation for Speech Translation (STAST) model which aims to improve the end-to-end model performance by bridging the modalit…
Shenzhen International Graduate School, Tsinghua University（清华大学深圳国际研究生院）；Tencent Cloud Xiaowei（腾讯云小微）
Existing text- and image-based multimodal dialogue systems use the traditional Hierarchical Recurrent Encoder-Decoder (HRED) framework, which has an utterance-level encoder to model utterance representation and a context-level encoder to model context representation. Although pioneer efforts have shown promising performances, they still suffer from the following challenges: (1) the interaction between textual features and visual features is not fine-grained enough. (2) the context representation can not provide a complete representation for the context. To address the issues mentioned above, w…
End-to-end speech-to-text translation models are often initialized with pre-trained speech encoder and pre-trained text decoder. This leads to a significant training gap between pre-training and fine-tuning, largely due to the modality differences between speech outputs from the encoder and text inputs to the decoder. In this work, we aim to bridge the modality gap between speech and text to improve translation quality. We propose M-Adapter, a novel Transformer-based module, to adapt speech representations to text. While shrinking the speech sequence, M-Adapter produces features desired for sp…
Columbia University（哥伦比亚大学）；Microsoft Cloud and AI（微软云与人工智能部门）；Microsoft Research（微软研究院）
Large-scale multi-modal contrastive pre-training has demonstrated great utility to learn transferable features for a range of downstream tasks by mapping multiple modalities into a shared embedding space. Typically, this has employed separate encoders for each modality. However, recent work suggests that transformers can support learning across multiple modalities and allow knowledge sharing. Inspired by this, we investigate a variety of Modality-Shared Contrastive Language-Image Pre-training (MS-CLIP) frameworks. More specifically, we question how many parameters of a transformer model can be…
Large scale pretrained language models have demonstrated state-of-the-art performance in language understanding tasks. Their application has recently expanded into multimodality learning, leading to improved representations combining vision and language. However, progress in adapting language models towards conditional Natural Language Generation (NLG) has been limited to a single modality, generally text. We propose MAnTiS, Multimodal Adaptation for Text Synthesis, a general approach for multimodal conditionality in transformer-based NLG models. In this method, we pass inputs from each modali…
We present modality gap, an intriguing geometric phenomenon of the representation space of multi-modal models. Specifically, we show that different data modalities (e.g. images and text) are embedded at arm's length in their shared representation in multi-modal models such as CLIP. Our systematic analysis demonstrates that this gap is caused by a combination of model initialization and contrastive learning optimization. In model initialization, we show empirically and theoretically that the representation of a common deep neural network is restricted to a narrow cone. As a consequence, in a mu…
We propose an approach to build a neural machine translation system with no supervised resources (i.e., no parallel corpora) using multimodal embedded representation over texts and images. Based on the assumption that text documents are often likely to be described with other multimedia information (e.g., images) somewhat related to the content, we try to indirectly estimate the relevance between two languages. Using multimedia as the "pivot", we project all modalities into one common hidden space where samples belonging to similar semantic concepts should come close to each other, whatever th…
In this paper, we propose a novel architecture for multi-modal speech and text input. We combine pretrained speech and text encoders using multi-headed cross-modal attention and jointly fine-tune on the target problem. The resultant architecture can be used for continuous token-level classification or utterance-level prediction acting on simultaneous text and speech. The resultant encoder efficiently captures both acoustic-prosodic and lexical information. We compare the benefits of multi-headed attention-based fusion for multi-modal utterance-level classification against a simple concatenatio…
In this paper, we propose a novel multi-modal multi-task encoder-decoder pre-training framework (MMSpeech) for Mandarin automatic speech recognition (ASR), which employs both unlabeled speech and text data. The main difficulty in speech-text joint pre-training comes from the significant difference between speech and text modalities, especially for Mandarin speech and text. Unlike English and other languages with an alphabetic writing system, Mandarin uses an ideographic writing system where character and sound are not tightly mapped to one another. Therefore, we propose to introduce the phonem…
The Hong Kong Polytechnic University（香港理工大学）；Nanjing University of Aeronautics and Astronautics（南京航空航天大学）
The multimedia communications with texts and images are popular on social media. However, limited studies concern how images are structured with texts to form coherent meanings in human cognition. To fill in the gap, we present a novel concept of cross-modality discourse, reflecting how human readers couple image and text understandings. Text descriptions are first derived from images (named as subtitles) in the multimedia contexts. Five labels -- entity-level insertion, projection and concretization and scene-level restatement and extension -- are further employed to shape the structure of su…
