This book is the result of a seminar in which we reviewed multimodal approaches and attempted to create a solid overview of the field, starting with the current state-of-the-art approaches in the two subfields of Deep Learning individually. Further, modeling frameworks are discussed where one modality is transformed into the other, as well as models in which one modality is utilized to enhance representation learning for the other. To conclude the second part, architectures with a focus on handling both modalities simultaneously are introduced. Finally, we also cover other modalities as well as general-purpose multi-modal models, which are able to handle different tasks on different modalities within one unified architecture. One interesting application (Generative Art) eventually caps off this booklet.
The advances in remote sensing technologies have boosted applications for Earth observation. These technologies provide multiple observations or views with different levels of information. They might contain static or temporary views with different levels of resolution, in addition to having different types and amounts of noise due to sensor calibration or deterioration. A great variety of deep learning models have been applied to fuse the information from these multiple views, known as deep multi-view or multi-modal fusion learning. However, the approaches in the literature vary greatly since different terminology is used to refer to similar concepts or different illustrations are given to similar techniques. This article gathers works on multi-view fusion for Earth observation by focusing on the common practices and approaches used in the literature. We summarize and structure insights from several different publications concentrating on unifying points and ideas. In this manuscript, we provide a harmonized terminology while at the same time mentioning the various alternative terms that are used in literature. The topics covered by the works reviewed focus on supervised learning with the use of neural network models. We hope this review, with a long list of recent references, can support future research and lead to a unified advance in the area.
The ability to jointly learn from multiple modalities, such as text, audio, and visual data, is a defining feature of intelligent systems. While there have been promising advances in designing neural networks to harness multimodal data, the enormous success of data augmentation currently remains limited to single-modality tasks like image classification. Indeed, it is particularly difficult to augment each modality while preserving the overall semantic structure of the data; for example, a caption may no longer be a good description of an image after standard augmentations have been applied, such as translation. Moreover, it is challenging to specify reasonable transformations that are not tailored to a particular modality. In this paper, we introduce LeMDA, Learning Multimodal Data Augmentation, an easy-to-use method that automatically learns to jointly augment multimodal data in feature space, with no constraints on the identities of the modalities or the relationship between modalities. We show that LeMDA can (1) profoundly improve the performance of multimodal deep learning architectures, (2) apply to combinations of modalities that have not been previously considered, and (3) achieve state-of-the-art results on a wide range of applications comprised of image, text, and tabular data.
Multimodal deep learning has been used to predict clinical endpoints and diagnoses from clinical routine data. However, these models suffer from scaling issues: they have to learn pairwise interactions between each piece of information in each data type, thereby escalating model complexity beyond manageable scales. This has so far precluded a widespread use of multimodal deep learning. Here, we present a new technical approach of "learnable synergies", in which the model only selects relevant interactions between data modalities and keeps an "internal memory" of relevant data. Our approach is easily scalable and naturally adapts to multimodal data inputs from clinical routine. We demonstrate this approach on three large multimodal datasets from radiology and ophthalmology and show that it outperforms state-of-the-art models in clinically relevant diagnosis tasks. Our new approach is transferable and will allow the application of multimodal deep learning to a broad set of clinically relevant problems.
Autonomous driving is a popular research area within the computer vision research community. Since autonomous vehicles are highly safety-critical, ensuring robustness is essential for real-world deployment. While several public multimodal datasets are accessible, they mainly comprise two sensor modalities (camera, LiDAR) which are not well suited for adverse weather. In addition, they lack far-range annotations, making it harder to train neural networks that are the base of a highway assistant function of an autonomous vehicle. Therefore, we introduce a multimodal dataset for robust autonomous driving with long-range perception. The dataset consists of 176 scenes with synchronized and calibrated LiDAR, camera, and radar sensors covering a 360-degree field of view. The collected data was captured in highway, urban, and suburban areas during daytime, night, and rain and is annotated with 3D bounding boxes with consistent identifiers across frames. Furthermore, we trained unimodal and multimodal baseline models for 3D object detection. Data are available at \url{https://github.com/aimotive/aimotive_dataset}.
Foundational image-language models have generated considerable interest due to their efficient adaptation to downstream tasks by prompt learning. Prompt learning treats part of the language model input as trainable while freezing the rest, and optimizes an Empirical Risk Minimization objective. However, Empirical Risk Minimization is known to suffer from distributional shifts which hurt generalizability to prompts unseen during training. By leveraging the regularization ability of Bayesian methods, we frame prompt learning from the Bayesian perspective and formulate it as a variational inference problem. Our approach regularizes the prompt space, reduces overfitting to the seen prompts and improves the prompt generalization on unseen prompts. Our framework is implemented by modeling the input prompt space in a probabilistic manner, as an a priori distribution which makes our proposal compatible with prompt learning approaches that are unconditional or conditional on the image. We demonstrate empirically on 15 benchmarks that Bayesian prompt learning provides an appropriate coverage of the prompt space, prevents learning spurious features, and exploits transferable invariant features. This results in better generalization of unseen prompts, even across different datasets and domains. Code available at: https://github.com/saic-fi/Bayesian-Prompt-Learning
Soft prompt learning has recently emerged as one of the methods of choice for adapting V&L models to a downstream task using a few training examples. However, current methods significantly overfit the training data, suffering from large accuracy degradation when tested on unseen classes from the same domain. To this end, in this paper, we make the following 4 contributions: (1) To alleviate base class overfitting, we propose a novel Language-Aware Soft Prompting (LASP) learning method by means of a text-to-text cross-entropy loss that maximizes the probability of the learned prompts to be correctly classified with respect to pre-defined hand-crafted textual prompts. (2) To increase the representation capacity of the prompts, we propose grouped LASP where each group of prompts is optimized with respect to a separate subset of textual prompts. (3) We identify a visual-language misalignment introduced by prompt learning and LASP, and more importantly, propose a re-calibration mechanism to address it. (4) We show that LASP is inherently amenable to including, during training, virtual classes, i.e. class names for which no visual samples are available, further increasing the robustness of the learned prompts. Through evaluations on 11 datasets, we show that our approach (a) significantly outperforms all prior works on soft prompting, and (b) matches and surpasses, for the first time, the accuracy on novel classes obtained by hand-crafted prompts and CLIP for 8 out of 11 test datasets. Code will be made available at https://www.adrianbulat.com/lasp
When answering a question, humans utilize the information available across different modalities to synthesize a consistent and complete chain of thought (CoT). This process is normally a black box in the case of deep learning models like large-scale language models. Recently, science question benchmarks have been used to diagnose the multi-hop reasoning ability and interpretability of an AI system. However, existing datasets fail to provide annotations for the answers, or are restricted to the textual-only modality, small scales, and limited domain diversity. To this end, we present Science Question Answering (ScienceQA), a new benchmark that consists of ~21k multimodal multiple choice questions with a diverse set of science topics and annotations of their answers with corresponding lectures and explanations. We further design language models to learn to generate lectures and explanations as the chain of thought (CoT) to mimic the multi-hop reasoning process when answering ScienceQA questions. ScienceQA demonstrates the utility of CoT in language models, as CoT improves the question answering performance by 1.20% in few-shot GPT-3 and 3.99% in fine-tuned UnifiedQA. We also explore the upper bound for models to leverage explanations by feeding those in the input; we observe that it improves the few-shot performance of GPT-3 by 18.96%. Our analysis further shows that language models, similar to humans, benefit from explanations to learn from fewer data and achieve the same performance with just 40% of the data. The data and code are available at https://scienceqa.github.io.
We introduce LAVIS, an open-source deep learning library for LAnguage-VISion research and applications. LAVIS aims to serve as a one-stop comprehensive library that brings recent advancements in the language-vision field accessible for researchers and practitioners, as well as fertilizing future research and development. It features a unified interface to easily access state-of-the-art image-language, video-language models and common datasets. LAVIS supports training, evaluation and benchmarking on a rich variety of tasks, including multimodal classification, retrieval, captioning, visual question answering, dialogue and pre-training. In the meantime, the library is also highly extensible and configurable, facilitating future development and customization. In this technical report, we describe design principles, key components and functionalities of the library, and also present benchmarking results across common language-vision tasks. The library is available at: https://github.com/salesforce/LAVIS.
AI creation, such as poem or lyrics generation, has attracted increasing attention from both industry and academic communities, with many promising models proposed in the past few years. Existing methods usually estimate the outputs based on single and independent visual or textual information. However, in reality, humans usually make creations according to their experiences, which may involve different modalities and be sequentially correlated. To model such human capabilities, in this paper, we define and solve a novel AI creation problem based on human experiences. More specifically, we study how to generate texts based on sequential multi-modal information. Compared with the previous works, this task is much more difficult because the designed model has to well understand and adapt the semantics among different modalities and effectively convert them into the output in a sequential manner. To alleviate these difficulties, we firstly design a multi-channel sequence-to-sequence architecture equipped with a multi-modal attention network. For more effective optimization, we then propose a curriculum negative sampling strategy tailored for the sequential inputs. To benchmark this problem and demonstrate the effectiveness of our model, we manually labeled a new multi-modal experience dataset. With this dataset, we conduct extensive experiments by comparing our model with a series of representative baselines, where we can demonstrate significant improvements in our model based on both automatic and human-centered metrics. The code and data are available at: \url{https://github.com/Aman-4-Real/MMTG}.
Alzheimer's Disease (AD) is the most common neurodegenerative disorder with one of the most complex pathogeneses, making effective and clinically actionable decision support difficult. The objective of this study was to develop a novel multimodal deep learning framework to aid medical professionals in AD diagnosis. We present a Multimodal Alzheimer's Disease Diagnosis framework (MADDi) to accurately detect the presence of AD and mild cognitive impairment (MCI) from imaging, genetic, and clinical data. MADDi is novel in that we use cross-modal attention, which captures interactions between modalities - a method not previously explored in this domain. We perform multi-class classification, a challenging task considering the strong similarities between MCI and AD. We compare with previous state-of-the-art models, evaluate the importance of attention, and examine the contribution of each modality to the model's performance. MADDi classifies MCI, AD, and controls with 96.88% accuracy on a held-out test set. When examining the contribution of different attention schemes, we found that the combination of cross-modal attention with self-attention performed the best, and no attention layers in the model performed the worst, with a 7.9% difference in F1-Scores. Our experiments underlined the importance of structured clinical data to help machine learning models contextualize and interpret the remaining modalities. Extensive ablation studies showed that any multimodal mixture of input features without access to structured clinical information suffered marked performance losses. This study demonstrates the merit of combining multiple input modalities via cross-modal attention to deliver highly accurate AD diagnostic decision support.
Gravitational lensing is the relativistic effect generated by massive bodies, which bend the space-time surrounding them. It is a deeply investigated topic in astrophysics and allows validating theoretical relativistic results and studying faint astrophysical objects that would not be visible otherwise. In recent years Machine Learning methods have been applied to support the analysis of the gravitational lensing phenomena by detecting lensing effects in data sets consisting of images associated with brightness variation time series. However, the state-of-art approaches either consider only images and neglect time-series data or achieve relatively low accuracy on the most difficult data sets. This paper introduces DeepGraviLens, a novel multi-modal network that classifies spatio-temporal data belonging to one non-lensed system type and three lensed system types. It surpasses the current state of the art accuracy results by $\approx 3\%$ to $\approx 11\%$, depending on the considered data set. Such an improvement will enable the acceleration of the analysis of lensed objects in upcoming astrophysical surveys, which will exploit the petabytes of data collected, e.g., from the Vera C. Rubin Observatory.
Recent works on 3D semantic segmentation propose to exploit the synergy between images and point clouds by processing each modality with a dedicated network and projecting learned 2D features onto 3D points. Merging large-scale point clouds and images raises several challenges, such as constructing a mapping between points and pixels, and aggregating features between multiple views. Current methods require mesh reconstruction or specialized sensors to recover occlusions, and use heuristics to select and aggregate available images. In contrast, we propose an end-to-end trainable multi-view aggregation model leveraging the viewing conditions of 3D points to merge features from images taken at arbitrary positions. Our method can combine standard 2D and 3D networks and outperforms both 3D models operating on colorized point clouds and hybrid 2D/3D networks without requiring colorization, meshing, or true depth maps. We set a new state-of-the-art for large-scale indoor/outdoor semantic segmentation on S3DIS (74.7 mIoU 6-Fold) and on KITTI-360 (58.3 mIoU). Our full pipeline is accessible at https://github.com/drprojects/DeepViewAgg, and only requires raw 3D scans and a set of images and poses.
CLIP (Contrastive Language-Image Pre-training) is a very recent multi-modal model that jointly learns representations of images and texts. The model is trained on a massive amount of English data and shows impressive performance on zero-shot classification tasks. Training the same model on a different language is not trivial, since data in other languages might be not enough and the model needs high-quality translations of the texts to guarantee a good performance. In this paper, we present the first CLIP model for the Italian Language (CLIP-Italian), trained on more than 1.4 million image-text pairs. Results show that CLIP-Italian outperforms the multilingual CLIP model on the tasks of image retrieval and zero-shot classification.
The rapidly emerging field of deep learning-based computational pathology has demonstrated promise in developing objective prognostic models from histology whole slide images. However, most prognostic models are either based on histology or genomics alone and do not address how histology and genomics can be integrated to develop joint image-omic prognostic models. Additionally identifying explainable morphological and molecular descriptors from these models that govern such prognosis is of interest. We used multimodal deep learning to integrate gigapixel whole slide pathology images, RNA-seq abundance, copy number variation, and mutation data from 5,720 patients across 14 major cancer types. Our interpretable, weakly-supervised, multimodal deep learning algorithm is able to fuse these heterogeneous modalities for predicting outcomes and discover prognostic features from these modalities that corroborate with poor and favorable outcomes via multimodal interpretability. We compared our model with unimodal deep learning models trained on histology slides and molecular profiles alone, and demonstrate performance increase in risk stratification on 9 out of 14 cancers. In addition, we analyze morphologic and molecular markers responsible for prognostic predictions across all cancer types. All analyzed data, including morphological and molecular correlates of patient prognosis across the 14 cancer types at a disease and patient level are presented in an interactive open-access database (http://pancancer.mahmoodlab.org) to allow for further exploration and prognostic biomarker discovery. To validate that these model explanations are prognostic, we further analyzed high attention morphological regions in WSIs, which indicates that tumor-infiltrating lymphocyte presence corroborates with favorable cancer prognosis on 9 out of 14 cancer types studied.
Multimodal sentiment analysis aims to extract and integrate semantic information collected from multiple modalities to recognize the expressed emotions and sentiment in multimodal data. This research area's major concern lies in developing an extraordinary fusion scheme that can extract and integrate key information from various modalities. However, one issue that may restrict previous work to achieve a higher level is the lack of proper modeling for the dynamics of the competition between the independence and relevance among modalities, which could deteriorate fusion outcomes by causing the collapse of modality-specific feature space or introducing extra noise. To mitigate this, we propose the Bi-Bimodal Fusion Network (BBFN), a novel end-to-end network that performs fusion (relevance increment) and separation (difference increment) on pairwise modality representations. The two parts are trained simultaneously such that the combat between them is simulated. The model takes two bimodal pairs as input due to the known information imbalance among modalities. In addition, we leverage a gated control mechanism in the Transformer architecture to further improve the final output. Experimental results on three datasets (CMU-MOSI, CMU-MOSEI, and UR-FUNNY) verifies that our model significantly outperforms the SOTA. The implementation of this work is available at https://github.com/declare-lab/multimodal-deep-learning.
Recently a number of studies demonstrated impressive performance on diverse vision-language multi-modal tasks such as image captioning and visual question answering by extending the BERT architecture with multi-modal pre-training objectives. In this work we explore a broad set of multi-modal representation learning tasks in the medical domain, specifically using radiology images and the unstructured report. We propose Medical Vision Language Learner (MedViLL), which adopts a BERT-based architecture combined with a novel multi-modal attention masking scheme to maximize generalization performance for both vision-language understanding tasks (diagnosis classification, medical image-report retrieval, medical visual question answering) and vision-language generation task (radiology report generation). By statistically and rigorously evaluating the proposed model on four downstream tasks with three radiographic image-report datasets (MIMIC-CXR, Open-I, and VQA-RAD), we empirically demonstrate the superior downstream task performance of MedViLL against various baselines, including task-specific architectures. The source code is publicly available at: https://github.com/SuperSupermoon/MedViLL
We propose a novel, efficient, modular and scalable framework for content based visual media retrieval systems by leveraging the power of Deep Learning which is flexible to work both for images and videos conjointly and we also introduce an efficient comparison and filtering metric for retrieval. We put forward our findings from critical performance tests comparing our method to the predominant conventional approach to demonstrate the feasibility and efficiency of the proposed solution with best practices, possible improvements that may further augment the ability of retrieval architectures.
The assignment of importance scores to particular frames or (short) segments in a video is crucial for summarization, but also a difficult task. Previous work utilizes only one source of visual features. In this paper, we suggest a novel model architecture that combines three feature sets for visual content and motion to predict importance scores. The proposed architecture utilizes an attention mechanism before fusing motion features and features representing the (static) visual content, i.e., derived from an image classification model. Comprehensive experimental evaluations are reported for two well-known datasets, SumMe and TVSum. In this context, we identify methodological issues on how previous work used these benchmark datasets, and present a fair evaluation scheme with appropriate data splits that can be used in future work. When using static and motion features with parallel attention mechanism, we improve state-of-the-art results for SumMe, while being on par with the state of the art for the other dataset.
Having access to multi-modal cues (e.g. vision and audio) empowers some cognitive tasks to be done faster compared to learning from a single modality. In this work, we propose to transfer knowledge across heterogeneous modalities, even though these data modalities may not be semantically correlated. Rather than directly aligning the representations of different modalities, we compose audio, image, and video representations across modalities to uncover richer multi-modal knowledge. Our main idea is to learn a compositional embedding that closes the cross-modal semantic gap and captures the task-relevant semantics, which facilitates pulling together representations across modalities by compositional contrastive learning. We establish a new, comprehensive multi-modal distillation benchmark on three video datasets: UCF101, ActivityNet, and VGGSound. Moreover, we demonstrate that our model significantly outperforms a variety of existing knowledge distillation methods in transferring audio-visual knowledge to improve video representation learning. Code is released here: https://github.com/yanbeic/CCL.