Abstract
Self-supervised representation learning follows a paradigm of withholding some part of the data and tasking the network to predict it from the remaining part. Towards this end, masking has emerged as a generic and powerful tool where content is withheld along the sequential dimension, e.g., spatial in images, temporal in audio, and syntactic in language. In this paper, we explore the orthogonal channel dimension for generic data augmentation. The data for each channel is quantized through a non-uniform quantizer, with the quantized value sampled randomly within randomly sampled quantization bins. From another perspective, quantization is analogous to channel-wise masking, as it removes the information within each bin, but preserves the information across bins. We apply the randomized quantization in conjunction with sequential augmentations on self-supervised contrastive models. This generic approach achieves results on par with modality-specific augmentation on vision tasks, and state-of-the-art results on 3D point clouds as well as on audio. We also demonstrate this method to be applicable for augmenting intermediate embeddings in a deep neural network on the comprehensive DABS benchmark which is comprised of various data modalities. Code is availabel at this http URL
Multi-person 3D pose estimation from unlabelled data
Authors: Daniel Rodriguez-Criado, Pilar Bachiller, George Vogiatzis, Luis J. Manso
Subjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
Abstract
Its numerous applications make multi-human 3D pose estimation a remarkably impactful area of research. Nevertheless, assuming a multiple-view system composed of several regular RGB cameras, 3D multi-pose estimation presents several challenges. First of all, each person must be uniquely identified in the different views to separate the 2D information provided by the cameras. Secondly, the 3D pose estimation process from the multi-view 2D information of each person must be robust against noise and potential occlusions in the scenario. In this work, we address these two challenges with the help of deep learning. Specifically, we present a model based on Graph Neural Networks capable of predicting the cross-view correspondence of the people in the scenario along with a Multilayer Perceptron that takes the 2D points to yield the 3D poses of each person. These two models are trained in a self-supervised manner, thus avoiding the need for large datasets with 3D annotations.
ColoristaNet for Photorealistic Video Style Transfer
Abstract
Photorealistic style transfer aims to transfer the artistic style of an image onto an input image or video while keeping photorealism. In this paper, we think it's the summary statistics matching scheme in existing algorithms that leads to unrealistic stylization. To avoid employing the popular Gram loss, we propose a self-supervised style transfer framework, which contains a style removal part and a style restoration part. The style removal network removes the original image styles, and the style restoration network recovers image styles in a supervised manner. Meanwhile, to address the problems in current feature transformation methods, we propose decoupled instance normalization to decompose feature transformation into style whitening and restylization. It works quite well in ColoristaNet and can transfer image styles efficiently while keeping photorealism. To ensure temporal coherency, we also incorporate optical flow methods and ConvLSTM to embed contextual information. Experiments demonstrates that ColoristaNet can achieve better stylization effects when compared with state-of-the-art algorithms.
APOLLO: A Simple Approach for Adaptive Pretraining of Language Models for Logical Reasoning
Abstract
Logical reasoning of text is an important ability that requires understanding the information present in the text, their interconnections, and then reasoning through them to infer new conclusions. Prior works on improving the logical reasoning ability of language models require complex processing of training data (e.g., aligning symbolic knowledge to text), yielding task-specific data augmentation solutions that restrict the learning of general logical reasoning skills. In this work, we propose APOLLO, an adaptively pretrained language model that has improved logical reasoning abilities. We select a subset of Wikipedia, based on a set of logical inference keywords, for continued pretraining of a language model. We use two self-supervised loss functions: a modified masked language modeling loss where only specific parts-of-speech words, that would likely require more reasoning than basic language understanding, are masked, and a sentence-level classification loss that teaches the model to distinguish between entailment and contradiction types of sentences. The proposed training paradigm is both simple and independent of task formats. We demonstrate the effectiveness of APOLLO by comparing it with prior baselines on two logical reasoning datasets. APOLLO performs comparably on ReClor and outperforms baselines on LogiQA.
Speaking Style Conversion With Discrete Self-Supervised Units
Authors: Gallil Maimon, Yossi Adi
Subjects: Sound (cs.SD); Computation and Language (cs.CL); Machine Learning (cs.LG); Audio and Speech Processing (eess.AS)
Abstract
Voice Conversion (VC) is the task of making a spoken utterance by one speaker sound as if uttered by a different speaker, while keeping other aspects like content unchanged. Current VC methods, focus primarily on spectral features like timbre, while ignoring the unique speaking style of people which often impacts prosody. In this study, we introduce a method for converting not only the timbre, but also prosodic information (i.e., rhythm and pitch changes) to those of the target speaker. The proposed approach is based on a pretrained, self-supervised, model for encoding speech to discrete units, which make it simple, effective, and easy to optimise. We consider the many-to-many setting with no paired data. We introduce a suite of quantitative and qualitative evaluation metrics for this setup, and empirically demonstrate the proposed approach is significantly superior to the evaluated baselines. Code and samples can be found under https://pages.cs.huji.ac.il/adiyoss-lab/dissc/ .
Keyword: vision transformer
There is no result
Keyword: multimodal
EffMulti: Efficiently Modeling Complex Multimodal Interactions for Emotion Analysis
Authors: Feng Qiu, Chengyang Xie, Yu Ding, Wanzeng Kong
Subjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
Abstract
Humans are skilled in reading the interlocutor's emotion from multimodal signals, including spoken words, simultaneous speech, and facial expressions. It is still a challenge to effectively decode emotions from the complex interactions of multimodal signals. In this paper, we design three kinds of multimodal latent representations to refine the emotion analysis process and capture complex multimodal interactions from different views, including a intact three-modal integrating representation, a modality-shared representation, and three modality-individual representations. Then, a modality-semantic hierarchical fusion is proposed to reasonably incorporate these representations into a comprehensive interaction representation. The experimental results demonstrate that our EffMulti outperforms the state-of-the-art methods. The compelling performance benefits from its well-designed framework with ease of implementation, lower computing complexity, and less trainable parameters.
iCub! Do you recognize what I am doing?: multimodal human action recognition on multisensory-enabled iCub robot
Abstract
This study uses multisensory data (i.e., color and depth) to recognize human actions in the context of multimodal human-robot interaction. Here we employed the iCub robot to observe the predefined actions of the human partners by using four different tools on 20 objects. We show that the proposed multimodal ensemble learning leverages complementary characteristics of three color cameras and one depth sensor that improves, in most cases, recognition accuracy compared to the models trained with a single modality. The results indicate that the proposed models can be deployed on the iCub robot that requires multimodal action recognition, including social tasks such as partner-specific adaptation, and contextual behavior understanding, to mention a few.
Exploring Workplace Behaviors through Speaking Patterns using Large-scale Multimodal Wearable Recordings: A Study of Healthcare Providers
Authors: Tiantian Feng, Shrikanth Narayanan
Subjects: Sound (cs.SD); Multimedia (cs.MM); Audio and Speech Processing (eess.AS)
Abstract
Interpersonal spoken communication is central to human interaction and the exchange of information. Such interactive processes involve not only speech and spoken language but also non-verbal cues such as hand gestures, facial expressions, and nonverbal vocalization, that are used to express feelings and provide feedback. These multimodal communication signals carry a variety of information about the people: traits like gender and age as well as about physical and psychological states and behavior. This work uses wearable multimodal sensors to investigate interpersonal communication behaviors focusing on speaking patterns among healthcare providers with a focus on nurses. We analyze longitudinal data collected from $99$ nurses in a large hospital setting over ten weeks. The results indicate that speaking pattern differences across shift schedules and working units. Moreover, results show that speaking patterns combined with physiological measures can be used to predict affect measures and life satisfaction scores. The implementation of this work can be accessed at https://github.com/usc-sail/tiles-audio-arousal.
Medical Diagnosis with Large Scale Multimodal Transformers -- Leveraging Diverse Data for More Accurate Diagnosis
Authors: Firas Khader, Gustav Mueller-Franzes, Tianci Wang, Tianyu Han, Soroosh Tayebi Arasteh, Christoph Haarburger, Johannes Stegmaier, Keno Bressem, Christiane Kuhl, Sven Nebelung, Jakob Nikolas Kather, Daniel Truhn
Abstract
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.
Transferring General Multimodal Pretrained Models to Text Recognition
Abstract
This paper proposes a new method, OFA-OCR, to transfer multimodal pretrained models to text recognition. Specifically, we recast text recognition as image captioning and directly transfer a unified vision-language pretrained model to the end task. Without pretraining on large-scale annotated or synthetic text recognition data, OFA-OCR outperforms the baselines and achieves state-of-the-art performance in the Chinese text recognition benchmark. Additionally, we construct an OCR pipeline with OFA-OCR, and we demonstrate that it can achieve competitive performance with the product-level API. The code (https://github.com/OFA-Sys/OFA) and demo (https://modelscope.cn/studios/damo/ofa_ocr_pipeline/summary) are publicly available.
Keyword: CLIP
3D Point Cloud Pre-training with Knowledge Distillation from 2D Images
Abstract
The recent success of pre-trained 2D vision models is mostly attributable to learning from large-scale datasets. However, compared with 2D image datasets, the current pre-training data of 3D point cloud is limited. To overcome this limitation, we propose a knowledge distillation method for 3D point cloud pre-trained models to acquire knowledge directly from the 2D representation learning model, particularly the image encoder of CLIP, through concept alignment. Specifically, we introduce a cross-attention mechanism to extract concept features from 3D point cloud and compare them with the semantic information from 2D images. In this scheme, the point cloud pre-trained models learn directly from rich information contained in 2D teacher models. Extensive experiments demonstrate that the proposed knowledge distillation scheme achieves higher accuracy than the state-of-the-art 3D pre-training methods for synthetic and real-world datasets on downstream tasks, including object classification, object detection, semantic segmentation, and part segmentation.
Efficient Image Captioning for Edge Devices
Authors: Ning Wang, Jiangrong Xie, Hang Luo, Qinglin Cheng, Jihao Wu, Mingbo Jia, Linlin Li
Subjects: Computer Vision and Pattern Recognition (cs.CV)
Abstract
Recent years have witnessed the rapid progress of image captioning. However, the demands for large memory storage and heavy computational burden prevent these captioning models from being deployed on mobile devices. The main obstacles lie in the heavyweight visual feature extractors (i.e., object detectors) and complicated cross-modal fusion networks. To this end, we propose LightCap, a lightweight image captioner for resource-limited devices. The core design is built on the recent CLIP model for efficient image captioning. To be specific, on the one hand, we leverage the CLIP model to extract the compact grid features without relying on the time-consuming object detectors. On the other hand, we transfer the image-text retrieval design of CLIP to image captioning scenarios by devising a novel visual concept extractor and a cross-modal modulator. We further optimize the cross-modal fusion model and parallel prediction heads via sequential and ensemble distillations. With the carefully designed architecture, our model merely contains 40M parameters, saving the model size by more than 75% and the FLOPs by more than 98% in comparison with the current state-of-the-art methods. In spite of the low capacity, our model still exhibits state-of-the-art performance on prevalent datasets, e.g., 136.6 CIDEr on COCO Karpathy test split. Testing on the smartphone with only a single CPU, the proposed LightCap exhibits a fast inference speed of 188ms per image, which is ready for practical applications.
CLIP is Also an Efficient Segmenter: A Text-Driven Approach for Weakly Supervised Semantic Segmentation
Authors: Yuqi Lin, Minghao Chen, Wenxiao Wang, Boxi Wu, Ke Li, Binbin Lin, Haifeng Liu, Xiaofei He
Subjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
Abstract
Weakly supervised semantic segmentation (WSSS) with image-level labels is a challenging task in computer vision. Mainstream approaches follow a multi-stage framework and suffer from high training costs. In this paper, we explore the potential of Contrastive Language-Image Pre-training models (CLIP) to localize different categories with only image-level labels and without any further training. To efficiently generate high-quality segmentation masks from CLIP, we propose a novel framework called CLIP-ES for WSSS. Our framework improves all three stages of WSSS with special designs for CLIP: 1) We introduce the softmax function into GradCAM and exploit the zero-shot ability of CLIP to suppress the confusion caused by non-target classes and backgrounds. Meanwhile, to take full advantage of CLIP, we re-explore text inputs under the WSSS setting and customize two text-driven strategies: sharpness-based prompt selection and synonym fusion. 2) To simplify the stage of CAM refinement, we propose a real-time class-aware attention-based affinity (CAA) module based on the inherent multi-head self-attention (MHSA) in CLIP-ViTs. 3) When training the final segmentation model with the masks generated by CLIP, we introduced a confidence-guided loss (CGL) to mitigate noise and focus on confident regions. Our proposed framework dramatically reduces the cost of training for WSSS and shows the capability of localizing objects in CLIP. Our CLIP-ES achieves SOTA performance on Pascal VOC 2012 and MS COCO 2014 while only taking 10% time of previous methods for the pseudo mask generation. Code is available at https://github.com/linyq2117/CLIP-ES.
MIST: Multi-modal Iterative Spatial-Temporal Transformer for Long-form Video Question Answering
Authors: Difei Gao, Luowei Zhou, Lei Ji, Linchao Zhu, Yi Yang, Mike Zheng Shou
Subjects: Computer Vision and Pattern Recognition (cs.CV)
Abstract
To build Video Question Answering (VideoQA) systems capable of assisting humans in daily activities, seeking answers from long-form videos with diverse and complex events is a must. Existing multi-modal VQA models achieve promising performance on images or short video clips, especially with the recent success of large-scale multi-modal pre-training. However, when extending these methods to long-form videos, new challenges arise. On the one hand, using a dense video sampling strategy is computationally prohibitive. On the other hand, methods relying on sparse sampling struggle in scenarios where multi-event and multi-granularity visual reasoning are required. In this work, we introduce a new model named Multi-modal Iterative Spatial-temporal Transformer (MIST) to better adapt pre-trained models for long-form VideoQA. Specifically, MIST decomposes traditional dense spatial-temporal self-attention into cascaded segment and region selection modules that adaptively select frames and image regions that are closely relevant to the question itself. Visual concepts at different granularities are then processed efficiently through an attention module. In addition, MIST iteratively conducts selection and attention over multiple layers to support reasoning over multiple events. The experimental results on four VideoQA datasets, including AGQA, NExT-QA, STAR, and Env-QA, show that MIST achieves state-of-the-art performance and is superior at computation efficiency and interpretability.
Keyword: metric learning
There is no result
Keyword: image retrieval
There is no result
Keyword: self-supervised
Randomized Quantization for Data Agnostic Representation Learning
Multi-person 3D pose estimation from unlabelled data
ColoristaNet for Photorealistic Video Style Transfer
APOLLO: A Simple Approach for Adaptive Pretraining of Language Models for Logical Reasoning
Speaking Style Conversion With Discrete Self-Supervised Units
Keyword: vision transformer
There is no result
Keyword: multimodal
EffMulti: Efficiently Modeling Complex Multimodal Interactions for Emotion Analysis
iCub! Do you recognize what I am doing?: multimodal human action recognition on multisensory-enabled iCub robot
Exploring Workplace Behaviors through Speaking Patterns using Large-scale Multimodal Wearable Recordings: A Study of Healthcare Providers
Medical Diagnosis with Large Scale Multimodal Transformers -- Leveraging Diverse Data for More Accurate Diagnosis
Transferring General Multimodal Pretrained Models to Text Recognition
Keyword: CLIP
3D Point Cloud Pre-training with Knowledge Distillation from 2D Images
Efficient Image Captioning for Edge Devices
CLIP is Also an Efficient Segmenter: A Text-Driven Approach for Weakly Supervised Semantic Segmentation
MIST: Multi-modal Iterative Spatial-Temporal Transformer for Long-form Video Question Answering
Keyword: DALLE
There is no result