English-based Vision-Language Pre-training (VLP) has achieved great success in various downstream tasks. Some efforts have been taken to generalize this success to non-English languages through Multilingual Vision-Language Pre-training (M-VLP). However, due to the large number of languages, M-VLP models often require huge computing resources and cannot be flexibly extended to new languages. In this work, we propose a \textbf{M}ulti\textbf{L}ingual \textbf{A}cquisition (MLA) framework that can easily generalize a monolingual Vision-Language Pre-training model into multilingual. Specifically, we design a lightweight language acquisition encoder based on state-of-the-art monolingual VLP models. We further propose a two-stage training strategy to optimize the language acquisition encoder, namely the Native Language Transfer stage and the Language Exposure stage. With much less multilingual training data and computing resources, our model achieves state-of-the-art performance on multilingual image-text and video-text retrieval benchmarks.
Lexical and semantic matching capture different successful approaches to text retrieval and the fusion of their results has proven to be more effective and robust than either alone. Prior work performs hybrid retrieval by conducting lexical and semantic matching using different systems (e.g., Lucene and Faiss, respectively) and then fusing their model outputs. In contrast, our work integrates lexical representations with dense semantic representations by densifying high-dimensional lexical representations into what we call low-dimensional dense lexical representations (DLRs). Our experiments show that DLRs can effectively approximate the original lexical representations, preserving effectiveness while improving query latency. Furthermore, we can combine dense lexical and semantic representations to generate dense hybrid representations (DHRs) that are more flexible and yield faster retrieval compared to existing hybrid techniques. In addition, we explore it jointly training lexical and semantic representations in a single model and empirically show that the resulting DHRs are able to combine the advantages of the individual components. Our best DHR model is competitive with state-of-the-art single-vector and multi-vector dense retrievers in both in-domain and zero-shot evaluation settings. Furthermore, our model is both faster and requires smaller indexes, making our dense representation framework an attractive approach to text retrieval. Our code is available at https://github.com/castorini/dhr.
A ranker plays an indispensable role in the de facto 'retrieval & rerank' pipeline, but its training still lags behind -- learning from moderate negatives or/and serving as an auxiliary module for a retriever. In this work, we first identify two major barriers to a robust ranker, i.e., inherent label noises caused by a well-trained retriever and non-ideal negatives sampled for a high-capable ranker. Thereby, we propose multiple retrievers as negative generators improve the ranker's robustness, where i) involving extensive out-of-distribution label noises renders the ranker against each noise distribution, and ii) diverse hard negatives from a joint distribution are relatively close to the ranker's negative distribution, leading to more challenging thus effective training. To evaluate our robust ranker (dubbed R$^2$anker), we conduct experiments in various settings on the popular passage retrieval benchmark, including BM25-reranking, full-ranking, retriever distillation, etc. The empirical results verify the new state-of-the-art effectiveness of our model.
Data augmentation is a necessity to enhance data efficiency in deep learning. For vision-language pre-training, data is only augmented either for images or for text in previous works. In this paper, we present MixGen: a joint data augmentation for vision-language representation learning to further improve data efficiency. It generates new image-text pairs with semantic relationships preserved by interpolating images and concatenating text. It's simple, and can be plug-and-played into existing pipelines. We evaluate MixGen on four architectures, including CLIP, ViLT, ALBEF and TCL, across five downstream vision-language tasks to show its versatility and effectiveness. For example, adding MixGen in ALBEF pre-training leads to absolute performance improvements on downstream tasks: image-text retrieval (+6.2% on COCO fine-tuned and +5.3% on Flicker30K zero-shot), visual grounding (+0.9% on RefCOCO+), visual reasoning (+$0.9% on NLVR2), visual question answering (+0.3% on VQA2.0), and visual entailment (+0.4% on SNLI-VE).
Vision-language (VL) pre-training has recently received considerable attention. However, most existing end-to-end pre-training approaches either only aim to tackle VL tasks such as image-text retrieval, visual question answering (VQA) and image captioning that test high-level understanding of images, or only target region-level understanding for tasks such as phrase grounding and object detection. We present FIBER (Fusion-In-the-Backbone-based transformER), a new VL model architecture that can seamlessly handle both these types of tasks. Instead of having dedicated transformer layers for fusion after the uni-modal backbones, FIBER pushes multimodal fusion deep into the model by inserting cross-attention into the image and text backbones, bringing gains in terms of memory and performance. In addition, unlike previous work that is either only pre-trained on image-text data or on fine-grained data with box-level annotations, we present a two-stage pre-training strategy that uses both these kinds of data efficiently: (i) coarse-grained pre-training based on image-text data; followed by (ii) fine-grained pre-training based on image-text-box data. We conduct comprehensive experiments on a wide range of VL tasks, ranging from VQA, image captioning, and retrieval, to phrase grounding, referring expression comprehension, and object detection. Using deep multimodal fusion coupled with the two-stage pre-training, FIBER provides consistent performance improvements over strong baselines across all tasks, often outperforming methods using magnitudes more data. Code is available at https://github.com/microsoft/FIBER.
To build an artificial neural network like the biological intelligence system, recent works have unified numerous tasks into a generalist model, which can process various tasks with shared parameters and do not have any task-specific modules. While generalist models achieve promising results on various benchmarks, they have performance degradation on some tasks compared with task-specialized models. In this work, we find that interference among different tasks and modalities is the main factor to this phenomenon. To mitigate such interference, we introduce the Conditional Mixture-of-Experts (Conditional MoEs) to generalist models. Routing strategies under different levels of conditions are proposed to take both the training/inference cost and generalization ability into account. By incorporating the proposed Conditional MoEs, the recently proposed generalist model Uni-Perceiver can effectively mitigate the interference across tasks and modalities, and achieves state-of-the-art results on a series of downstream tasks via prompt tuning on 1% of downstream data. Moreover, the introduction of Conditional MoEs still holds the generalization ability of generalist models to conduct zero-shot inference on new tasks, e.g., video-text retrieval and video caption. Code and pre-trained generalist models shall be released.
Image-text retrieval is a fundamental and crucial branch in information retrieval. Although much progress has been made in bridging vision and language, it remains challenging because of the difficult intra-modal reasoning and cross-modal alignment. Existing modality interaction methods have achieved impressive results on public datasets. However, they heavily rely on expert experience and empirical feedback towards the design of interaction patterns, therefore, lacking flexibility. To address these issues, we develop a novel modality interaction modeling network based upon the routing mechanism, which is the first unified and dynamic multimodal interaction framework towards image-text retrieval. In particular, we first design four types of cells as basic units to explore different levels of modality interactions, and then connect them in a dense strategy to construct a routing space. To endow the model with the capability of path decision, we integrate a dynamic router in each cell for pattern exploration. As the routers are conditioned on inputs, our model can dynamically learn different activated paths for different data. Extensive experiments on two benchmark datasets, i.e., Flickr30K and MS-COCO, verify the superiority of our model compared with several state-of-the-art baselines.
Video-Language Pretraining (VLP), which aims to learn transferable representation to advance a wide range of video-text downstream tasks, has recently received increasing attention. Best performing works rely on large-scale, 3rd-person video-text datasets, such as HowTo100M. In this work, we exploit the recently released Ego4D dataset to pioneer Egocentric VLP along three directions. (i) We create EgoClip, a 1st-person video-text pretraining dataset comprising 3.8M clip-text pairs well-chosen from Ego4D, covering a large variety of human daily activities. (ii) We propose a novel pretraining objective, dubbed EgoNCE, which adapts video-text contrastive learning to the egocentric domain by mining egocentric-aware positive and negative samples. (iii) We introduce EgoMCQ, a development benchmark that is close to EgoClip and hence can support effective validation and fast exploration of our design decisions in EgoClip and EgoNCE. Furthermore, we demonstrate strong performance on five egocentric downstream tasks across three datasets: video-text retrieval on EPIC-KITCHENS-100; action recognition on Charades-Ego; natural language query, moment query, and object state change classification on Ego4D challenge benchmarks. The dataset and code are available at https://github.com/showlab/EgoVLP.
We introduce a vision-language foundation model called VL-BEiT, which is a bidirectional multimodal Transformer learned by generative pretraining. Our minimalist solution conducts masked prediction on both monomodal and multimodal data with a shared Transformer. Specifically, we perform masked vision-language modeling on image-text pairs, masked language modeling on texts, and masked image modeling on images. VL-BEiT is learned from scratch with one unified pretraining task, one shared backbone, and one-stage training. Our method is conceptually simple and empirically effective. Experimental results show that VL-BEiT obtains strong results on various vision-language benchmarks, such as visual question answering, visual reasoning, and image-text retrieval. Moreover, our method learns transferable visual features, achieving competitive performance on image classification, and semantic segmentation.
In this paper, we introduce Cross-View Language Modeling, a simple and effective pre-training framework that unifies cross-lingual and cross-modal pre-training with shared architectures and objectives. Our approach is motivated by a key observation that cross-lingual and cross-modal pre-training share the same goal of aligning two different views of the same object into a common semantic space. To this end, the cross-view language modeling framework considers both multi-modal data (i.e., image-caption pairs) and multi-lingual data (i.e., parallel sentence pairs) as two different views of the same object, and trains the model to align the two views by maximizing the mutual information between them with conditional masked language modeling and contrastive learning. We pre-train CCLM, a Cross-lingual Cross-modal Language Model, with the cross-view language modeling framework. Empirical results on IGLUE, a multi-lingual multi-modal benchmark, and two multi-lingual image-text retrieval datasets show that while conceptually simpler, CCLM significantly outperforms the prior state-of-the-art with an average absolute improvement of over 10%. Moreover, CCLM is the first multi-lingual multi-modal pre-trained model that surpasses the translate-test performance of representative English vision-language models by zero-shot cross-lingual transfer.
Dialogue systems can benefit from being able to search through a corpus of text to find information relevant to user requests, especially when encountering a request for which no manually curated response is available. The state-of-the-art technology for neural dense retrieval or re-ranking involves deep learning models with hundreds of millions of parameters. However, it is difficult and expensive to get such models to operate at an industrial scale, especially for cloud services that often need to support a big number of individually customized dialogue systems, each with its own text corpus. We report our work on enabling advanced neural dense retrieval systems to operate effectively at scale on relatively inexpensive hardware. We compare with leading alternative industrial solutions and show that we can provide a solution that is effective, fast, and cost-efficient.
Unpaired Image Captioning (UIC) has been developed to learn image descriptions from unaligned vision-language sample pairs. Existing works usually tackle this task using adversarial learning and visual concept reward based on reinforcement learning. However, these existing works were only able to learn limited cross-domain information in vision and language domains, which restrains the captioning performance of UIC. Inspired by the success of Vision-Language Pre-Trained Models (VL-PTMs) in this research, we attempt to infer the cross-domain cue information about a given image from the large VL-PTMs for the UIC task. This research is also motivated by recent successes of prompt learning in many downstream multi-modal tasks, including image-text retrieval and vision question answering. In this work, a semantic prompt is introduced and aggregated with visual features for more accurate caption prediction under the adversarial learning framework. In addition, a metric prompt is designed to select high-quality pseudo image-caption samples obtained from the basic captioning model and refine the model in an iterative manner. Extensive experiments on the COCO and Flickr30K datasets validate the promising captioning ability of the proposed model. We expect that the proposed prompt-based UIC model will stimulate a new line of research for the VL-PTMs based captioning.
In the past few years, the emergence of vision-language pre-training (VLP) has brought cross-modal retrieval to a new era. However, due to the latency and computation demand, it is commonly challenging to apply VLP in a real-time online retrieval system. To alleviate the defect, this paper proposes a \textbf{Hi}erarchical \textbf{V}ision-\textbf{}Language \textbf{P}re-Training (\textbf{HiVLP}) for fast Image-Text Retrieval (ITR). Specifically, we design a novel hierarchical retrieval objective, which uses the representation of different dimensions for coarse-to-fine ITR, i.e., using low-dimensional representation for large-scale coarse retrieval and high-dimensional representation for small-scale fine retrieval. We evaluate our proposed HiVLP on two popular image-text retrieval benchmarks, i.e., Flickr30k and COCO. Extensive experiments demonstrate that our HiVLP not only has fast inference speed but also can be easily scaled to large-scale ITR scenarios. The detailed results show that HiVLP is $1,427$$\sim$$120,649\times$ faster than the fusion-based model UNITER and 2$\sim$5 faster than the fastest embedding-based model LightingDot in different candidate scenarios. It also achieves about +4.9 AR on COCO and +3.8 AR on Flickr30K than LightingDot and achieves comparable performance with the state-of-the-art (SOTA) fusion-based model METER.
Large-scale pretrained foundation models have been an emerging paradigm for building artificial intelligence (AI) systems, which can be quickly adapted to a wide range of downstream tasks. This paper presents mPLUG, a new vision-language foundation model for both cross-modal understanding and generation. Most existing pre-trained models suffer from the problems of low computational efficiency and information asymmetry brought by the long visual sequence in cross-modal alignment. To address these problems, mPLUG introduces an effective and efficient vision-language architecture with novel cross-modal skip-connections, which creates inter-layer shortcuts that skip a certain number of layers for time-consuming full self-attention on the vision side. mPLUG is pre-trained end-to-end on large-scale image-text pairs with both discriminative and generative objectives. It achieves state-of-the-art results on a wide range of vision-language downstream tasks, such as image captioning, image-text retrieval, visual grounding and visual question answering. mPLUG also demonstrates strong zero-shot transferability when directly transferred to multiple video-language tasks.
Deep pre-trained language models (e,g. BERT) are effective at large-scale text retrieval task. Existing text retrieval systems with state-of-the-art performance usually adopt a retrieve-then-reranking architecture due to the high computational cost of pre-trained language models and the large corpus size. Under such a multi-stage architecture, previous studies mainly focused on optimizing single stage of the framework thus improving the overall retrieval performance. However, how to directly couple multi-stage features for optimization has not been well studied. In this paper, we design Hybrid List Aware Transformer Reranking (HLATR) as a subsequent reranking module to incorporate both retrieval and reranking stage features. HLATR is lightweight and can be easily parallelized with existing text retrieval systems so that the reranking process can be performed in a single yet efficient processing. Empirical experiments on two large-scale text retrieval datasets show that HLATR can efficiently improve the ranking performance of existing multi-stage text retrieval methods.
Vision-language pre-training (VLP) on large-scale datasets has shown premier performance on various downstream tasks. In contrast to plenty of available benchmarks with English corpus, large-scale pre-training datasets and downstream datasets with Chinese corpus remain largely unexplored. In this work, we build a large-scale high-quality Chinese Cross-Modal Benchmark named CCMB for the research community, which contains the currently largest public pre-training dataset Zero and five human-annotated fine-tuning datasets for downstream tasks. Zero contains 250 million images paired with 750 million text descriptions, plus two of the five fine-tuning datasets are also currently the largest ones for Chinese cross-modal downstream tasks. Along with the CCMB, we also develop a VLP framework named R2D2, applying a pre-Ranking + Ranking strategy to learn powerful vision-language representations and a two-way distillation method (i.e., target-guided Distillation and feature-guided Distillation) to further enhance the learning capability. With the Zero and the R2D2 VLP framework, we achieve state-of-the-art performance on twelve downstream datasets from five broad categories of tasks including image-text retrieval, image-text matching, image caption, text-to-image generation, and zero-shot image classification. The datasets, models, and codes are available at https://github.com/yuxie11/R2D2
How can we learn unified representations for spoken utterances and their written text? Learning similar representations for semantically similar speech and text is important for speech translation. To this end, we propose ConST, a cross-modal contrastive learning method for end-to-end speech-to-text translation. We evaluate ConST and a variety of previous baselines on a popular benchmark MuST-C. Experiments show that the proposed ConST consistently outperforms the previous methods on, and achieves an average BLEU of 29.4. The analysis further verifies that ConST indeed closes the representation gap of different modalities -- its learned representation improves the accuracy of cross-modal speech-text retrieval from 4% to 88%. Code and models are available at https://github.com/ReneeYe/ConST.
A common practice for text retrieval is to use an encoder to map the documents and the query to a common vector space and perform a nearest neighbor search (NNS); multi-hop retrieval also often adopts the same paradigm, usually with a modification of iteratively reformulating the query vector so that it can retrieve different documents at each hop. However, such a bi-encoder approach has limitations in multi-hop settings; (1) the reformulated query gets longer as the number of hops increases, which further tightens the embedding bottleneck of the query vector, and (2) it is prone to error propagation. In this paper, we focus on alleviating these limitations in multi-hop settings by formulating the problem in a fully generative way. We propose an encoder-decoder model that performs multi-hop retrieval by simply generating the entire text sequences of the retrieval targets, which means the query and the documents interact in the language model's parametric space rather than L2 or inner product space as in the bi-encoder approach. Our approach, Generative Multi-hop Retrieval(GMR), consistently achieves comparable or higher performance than bi-encoder models in five datasets while demonstrating superior GPU memory and storage footprint.
Dominant pre-training work for video-text retrieval mainly adopt the "dual-encoder" architectures to enable efficient retrieval, where two separate encoders are used to contrast global video and text representations, but ignore detailed local semantics. The recent success of image BERT pre-training with masked visual modeling that promotes the learning of local visual context, motivates a possible solution to address the above limitation. In this work, we for the first time investigate masked visual modeling in video-text pre-training with the "dual-encoder" architecture. We perform Masked visual modeling with Injected LanguagE Semantics (MILES) by employing an extra snapshot video encoder as an evolving "tokenizer" to produce reconstruction targets for masked video patch prediction. Given the corrupted video, the video encoder is trained to recover text-aligned features of the masked patches via reasoning with the visible regions along the spatial and temporal dimensions, which enhances the discriminativeness of local visual features and the fine-grained cross-modality alignment. Our method outperforms state-of-the-art methods for text-to-video retrieval on four datasets with both zero-shot and fine-tune evaluation protocols. Our approach also surpasses the baseline models significantly on zero-shot action recognition, which can be cast as video-to-text retrieval.
Image retrieval with hybrid-modality queries, also known as composing text and image for image retrieval (CTI-IR), is a retrieval task where the search intention is expressed in a more complex query format, involving both vision and text modalities. For example, a target product image is searched using a reference product image along with text about changing certain attributes of the reference image as the query. It is a more challenging image retrieval task that requires both semantic space learning and cross-modal fusion. Previous approaches that attempt to deal with both aspects achieve unsatisfactory performance. In this paper, we decompose the CTI-IR task into a three-stage learning problem to progressively learn the complex knowledge for image retrieval with hybrid-modality queries. We first leverage the semantic embedding space for open-domain image-text retrieval, and then transfer the learned knowledge to the fashion-domain with fashion-related pre-training tasks. Finally, we enhance the pre-trained model from single-query to hybrid-modality query for the CTI-IR task. Furthermore, as the contribution of individual modality in the hybrid-modality query varies for different retrieval scenarios, we propose a self-supervised adaptive weighting strategy to dynamically determine the importance of image and text in the hybrid-modality query for better retrieval. Extensive experiments show that our proposed model significantly outperforms state-of-the-art methods in the mean of Recall@K by 24.9% and 9.5% on the Fashion-IQ and Shoes benchmark datasets respectively.