Abstract
Interoperability issue is a significant problem in Building Information Modeling (BIM). Object type, as a kind of critical semantic information needed in multiple BIM applications like scan-to-BIM and code compliance checking, also suffers when exchanging BIM data or creating models using software of other domains. It can be supplemented using deep learning. Current deep learning methods mainly learn from the shape information of BIM objects for classification, leaving relational information inherent in the BIM context unused. To address this issue, we introduce a two-branch geometric-relational deep learning framework. It boosts previous geometric classification methods with relational information. We also present a BIM object dataset IFCNet++, which contains both geometric and relational information about the objects. Experiments show that our framework can be flexibly adapted to different geometric methods. And relational features do act as a bonus to general geometric learning methods, obviously improving their classification performance, thus reducing the manual labor of checking models and improving the practical value of enriched BIM models.
Keyword: image retrieval
StructVPR: Distill Structural Knowledge with Weighting Samples for Visual Place Recognition
Abstract
Visual place recognition (VPR) is usually considered as a specific image retrieval problem. Limited by existing training frameworks, most deep learning-based works cannot extract sufficiently stable global features from RGB images and rely on a time-consuming re-ranking step to exploit spatial structural information for better performance. In this paper, we propose StructVPR, a novel training architecture for VPR, to enhance structural knowledge in RGB global features and thus improve feature stability in a constantly changing environment. Specifically, StructVPR uses segmentation images as a more definitive source of structural knowledge input into a CNN network and applies knowledge distillation to avoid online segmentation and inference of seg-branch in testing. Considering that not all samples contain high-quality and helpful knowledge, and some even hurt the performance of distillation, we partition samples and weigh each sample's distillation loss to enhance the expected knowledge precisely. Finally, StructVPR achieves impressive performance on several benchmarks using only global retrieval and even outperforms many two-stage approaches by a large margin. After adding additional re-ranking, ours achieves state-of-the-art performance while maintaining a low computational cost.
Abstract
Extracting the relevant information out of a large number of documents is a challenging and tedious task. The quality of results generated by the traditionally available full-text search engine and text-based image retrieval systems is not optimal. Information retrieval (IR) tasks become more challenging with the nontraditional language scripts, as in the case of Indic scripts. The authors have developed OCR (Optical Character Recognition) Search Engine to make an Information Retrieval & Extraction (IRE) system that replicates the current state-of-the-art methods using the IRE and Natural Language Processing (NLP) techniques. Here we have presented the study of the methods used for performing search and retrieval tasks. The details of this system, along with the statistics of the dataset (source: National Digital Library of India or NDLI), is also presented. Additionally, the ideas to further explore and add value to research in IRE are also discussed.
Keyword: self-supervised
Progressive Feature Upgrade in Semi-supervised Learning on Tabular Domain
Authors: Morteza Mohammady Gharasuie, Fenjiao Wang
Abstract
Recent semi-supervised and self-supervised methods have shown great success in the image and text domain by utilizing augmentation techniques. Despite such success, it is not easy to transfer this success to tabular domains. It is not easy to adapt domain-specific transformations from image and language to tabular data due to mixing of different data types (continuous data and categorical data) in the tabular domain. There are a few semi-supervised works on the tabular domain that have focused on proposing new augmentation techniques for tabular data. These approaches may have shown some improvement on datasets with low-cardinality in categorical data. However, the fundamental challenges have not been tackled. The proposed methods either do not apply to datasets with high-cardinality or do not use an efficient encoding of categorical data. We propose using conditional probability representation and an efficient progressively feature upgrading framework to effectively learn representations for tabular data in semi-supervised applications. The extensive experiments show superior performance of the proposed framework and the potential application in semi-supervised settings.
ObjectStitch: Generative Object Compositing
Authors: Yizhi Song, Zhifei Zhang, Zhe Lin, Scott Cohen, Brian Price, Jianming Zhang, Soo Ye Kim, Daniel Aliaga
Subjects: Computer Vision and Pattern Recognition (cs.CV)
Abstract
Object compositing based on 2D images is a challenging problem since it typically involves multiple processing stages such as color harmonization, geometry correction and shadow generation to generate realistic results. Furthermore, annotating training data pairs for compositing requires substantial manual effort from professionals, and is hardly scalable. Thus, with the recent advances in generative models, in this work, we propose a self-supervised framework for object compositing by leveraging the power of conditional diffusion models. Our framework can hollistically address the object compositing task in a unified model, transforming the viewpoint, geometry, color and shadow of the generated object while requiring no manual labeling. To preserve the input object's characteristics, we introduce a content adaptor that helps to maintain categorical semantics and object appearance. A data augmentation method is further adopted to improve the fidelity of the generator. Our method outperforms relevant baselines in both realism and faithfulness of the synthesized result images in a user study on various real-world images.
FedCoCo: A Memory Efficient Federated Self-supervised Framework for On-Device Visual Representation Learning
Abstract
The ubiquity of edge devices has led to a growing amount of unlabeled data produced at the edge. Deep learning models deployed on edge devices are required to learn from these unlabeled data to continuously improve accuracy. Self-supervised representation learning has achieved promising performances using centralized unlabeled data. However, the increasing awareness of privacy protection limits centralizing the distributed unlabeled image data on edge devices. While federated learning has been widely adopted to enable distributed machine learning with privacy preservation, without a data selection method to efficiently select streaming data, the traditional federated learning framework fails to handle these huge amounts of decentralized unlabeled data with limited storage resources on edge. To address these challenges, we propose a Federated on-device Contrastive learning framework with Coreset selection, which we call FedCoCo, to automatically select a coreset that consists of the most representative samples into the replay buffer on each device. It preserves data privacy as each client does not share raw data while learning good visual representations. Experiments demonstrate the effectiveness and significance of the proposed method in visual representation learning.
CLeBPI: Contrastive Learning for Bug Priority Inference
Abstract
Automated bug priority inference can reduce the time overhead of bug triagers for priority assignments, improving the efficiency of software maintenance. Currently, there are two orthogonal lines for this task, i.e., traditional machine learning based (TML-based) and neural network based (NN-based) approaches. Although these approaches achieve competitive performance, our observation finds that existing approaches face the following two issues: 1) TML-based approaches require much manual feature engineering and cannot learn the semantic information of bug reports; 2) Both TML-based and NN-based approaches cannot effectively address the label imbalance problem because they are difficult to distinguish the semantic difference between bug reports with different priorities. In this paper, we propose CLeBPI (Contrastive Learning for Bug Priority Inference), which leverages pre-trained language model and contrastive learning to tackle the above-mentioned two issues. Specifically, CLeBPI is first pre-trained on a large-scale bug report corpus in a self-supervised way, thus it can automatically learn contextual representations of bug reports without manual feature engineering. Afterward, it is further pre-trained by a contrastive learning objective, which enables it to distinguish semantic differences between bug reports, learning more precise contextual representations for each bug report. When finishing pre-training, we can connect a classification layer to CLeBPI and fine-tune it for bug priority inference in a supervised way. To verify the effectiveness of CLeBPI, we choose four baseline approaches and conduct comparison experiments on a public dataset. The experimental results show that CLeBPI outperforms all baseline approaches by 23.86%-77.80% in terms of weighted average F1-score, showing its effectiveness.
General Framework for Self-Supervised Model Priming for Parameter-Efficient Fine-tuning
Abstract
Parameter-efficient methods (like Prompt or Adapters) for adapting pre-trained language models to downstream tasks have been popular recently. However, hindrances still prevent these methods from reaching their full potential. For example, two significant challenges are few-shot adaptation and cross-task generalization ability. To tackle these issues, we propose a general framework to enhance the few-shot adaptation and cross-domain generalization ability of parameter-efficient methods. In our framework, we prime the self-supervised model for parameter-efficient methods to rapidly adapt to various downstream few-shot tasks. To evaluate the authentic generalization ability of these parameter-efficient methods, we conduct experiments on a few-shot cross-domain benchmark containing 160 diverse NLP tasks. The experiment result reveals that priming by tuning PLM only with extra training tasks leads to the best performance. Also, we perform a comprehensive analysis of various parameter-efficient methods under few-shot cross-domain scenarios.
A General Purpose Supervisory Signal for Embodied Agents
Abstract
Training effective embodied AI agents often involves manual reward engineering, expert imitation, specialized components such as maps, or leveraging additional sensors for depth and localization. Another approach is to use neural architectures alongside self-supervised objectives which encourage better representation learning. In practice, there are few guarantees that these self-supervised objectives encode task-relevant information. We propose the Scene Graph Contrastive (SGC) loss, which uses scene graphs as general-purpose, training-only, supervisory signals. The SGC loss does away with explicit graph decoding and instead uses contrastive learning to align an agent's representation with a rich graphical encoding of its environment. The SGC loss is generally applicable, simple to implement, and encourages representations that encode objects' semantics, relationships, and history. Using the SGC loss, we attain significant gains on three embodied tasks: Object Navigation, Multi-Object Navigation, and Arm Point Navigation. Finally, we present studies and analyses which demonstrate the ability of our trained representation to encode semantic cues about the environment.
DeepFT: Fault-Tolerant Edge Computing using a Self-Supervised Deep Surrogate Model
Authors: Shreshth Tuli, Giuliano Casale, Ludmila Cherkasova, Nicholas R. Jennings
Subjects: Distributed, Parallel, and Cluster Computing (cs.DC); Artificial Intelligence (cs.AI)
Abstract
The emergence of latency-critical AI applications has been supported by the evolution of the edge computing paradigm. However, edge solutions are typically resource-constrained, posing reliability challenges due to heightened contention for compute and communication capacities and faulty application behavior in the presence of overload conditions. Although a large amount of generated log data can be mined for fault prediction, labeling this data for training is a manual process and thus a limiting factor for automation. Due to this, many companies resort to unsupervised fault-tolerance models. Yet, failure models of this kind can incur a loss of accuracy when they need to adapt to non-stationary workloads and diverse host characteristics. To cope with this, we propose a novel modeling approach, called DeepFT, to proactively avoid system overloads and their adverse effects by optimizing the task scheduling and migration decisions. DeepFT uses a deep surrogate model to accurately predict and diagnose faults in the system and co-simulation based self-supervised learning to dynamically adapt the model in volatile settings. It offers a highly scalable solution as the model size scales by only 3 and 1 percent per unit increase in the number of active tasks and hosts. Extensive experimentation on a Raspberry-Pi based edge cluster with DeFog benchmarks shows that DeepFT can outperform state-of-the-art baseline methods in fault-detection and QoS metrics. Specifically, DeepFT gives the highest F1 scores for fault-detection, reducing service deadline violations by up to 37\% while also improving response time by up to 9%.
ColD Fusion: Collaborative Descent for Distributed Multitask Finetuning
Abstract
Pretraining has been shown to scale well with compute, data size and data diversity. Multitask learning trains on a mixture of supervised datasets and produces improved performance compared to self-supervised pretraining. Until now, massively multitask learning required simultaneous access to all datasets in the mixture and heavy compute resources that are only available to well-resourced teams. In this paper, we propose ColD Fusion, a method that provides the benefits of multitask learning but leverages distributed computation and requires limited communication and no sharing of data. Consequentially, ColD Fusion can create a synergistic loop, where finetuned models can be recycled to continually improve the pretrained model they are based on. We show that ColD Fusion yields comparable benefits to multitask pretraining by producing a model that (a) attains strong performance on all of the datasets it was multitask trained on and (b) is a better starting point for finetuning on unseen datasets. We find ColD Fusion outperforms RoBERTa and even previous multitask models. Specifically, when training and testing on 35 diverse datasets, ColD Fusion-based model outperforms RoBERTa by 2.45 points in average without any changes to the architecture.
Keyword: vision transformer
There is no result
Keyword: multimodal
UniT3D: A Unified Transformer for 3D Dense Captioning and Visual Grounding
Abstract
Performing 3D dense captioning and visual grounding requires a common and shared understanding of the underlying multimodal relationships. However, despite some previous attempts on connecting these two related tasks with highly task-specific neural modules, it remains understudied how to explicitly depict their shared nature to learn them simultaneously. In this work, we propose UniT3D, a simple yet effective fully unified transformer-based architecture for jointly solving 3D visual grounding and dense captioning. UniT3D enables learning a strong multimodal representation across the two tasks through a supervised joint pre-training scheme with bidirectional and seq-to-seq objectives. With a generic architecture design, UniT3D allows expanding the pre-training scope to more various training sources such as the synthesized data from 2D prior knowledge to benefit 3D vision-language tasks. Extensive experiments and analysis demonstrate that UniT3D obtains significant gains for 3D dense captioning and visual grounding.
Weakly Supervised Annotations for Multi-modal Greeting Cards Dataset
Authors: Sidra Hanif, Longin Jan Latecki
Subjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
Abstract
In recent years, there is a growing number of pre-trained models trained on a large corpus of data and yielding good performance on various tasks such as classifying multimodal datasets. These models have shown good performance on natural images but are not fully explored for scarce abstract concepts in images. In this work, we introduce an image/text-based dataset called Greeting Cards. Dataset (GCD) that has abstract visual concepts. In our work, we propose to aggregate features from pretrained images and text embeddings to learn abstract visual concepts from GCD. This allows us to learn the text-modified image features, which combine complementary and redundant information from the multi-modal data streams into a single, meaningful feature. Secondly, the captions for the GCD dataset are computed with the pretrained CLIP-based image captioning model. Finally, we also demonstrate that the proposed the dataset is also useful for generating greeting card images using pre-trained text-to-image generation model.
Masked Contrastive Pre-Training for Efficient Video-Text Retrieval
Authors: Fangxun Shu, Biaolong Chen, Yue Liao, Ke Gao, Shuwen Xiao, Wenyu Sun, Xiaobo Li, Yousong Zhu, Jinqiao Wang, Si Liu
Subjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
Abstract
We present a simple yet effective end-to-end Video-language Pre-training (VidLP) framework, Masked Contrastive Video-language Pretraining (MAC), for video-text retrieval tasks. Our MAC aims to reduce video representation's spatial and temporal redundancy in the VidLP model by a mask sampling mechanism to improve pre-training efficiency. Comparing conventional temporal sparse sampling, we propose to randomly mask a high ratio of spatial regions and only feed visible regions into the encoder as sparse spatial sampling. Similarly, we adopt the mask sampling technique for text inputs for consistency. Instead of blindly applying the mask-then-prediction paradigm from MAE, we propose a masked-then-alignment paradigm for efficient video-text alignment. The motivation is that video-text retrieval tasks rely on high-level alignment rather than low-level reconstruction, and multimodal alignment with masked modeling encourages the model to learn a robust and general multimodal representation from incomplete and unstable inputs. Coupling these designs enables efficient end-to-end pre-training: reduce FLOPs (60% off), accelerate pre-training (by 3x), and improve performance. Our MAC achieves state-of-the-art results on various video-text retrieval datasets, including MSR-VTT, DiDeMo, and ActivityNet. Our approach is omnivorous to input modalities. With minimal modifications, we achieve competitive results on image-text retrieval tasks.
Coevolutionary Framework for Generalized Multimodal Multi-objective Optimization
Abstract
Most multimodal multi-objective evolutionary algorithms (MMEAs) aim to find all global Pareto optimal sets (PSs) for a multimodal multi-objective optimization problem (MMOP). However, in real-world problems, decision makers (DMs) may be also interested in local PSs. Also, searching for both global and local PSs is more general in view of dealing with MMOPs, which can be seen as a generalized MMOP. In addition, the state-of-the-art MMEAs exhibit poor convergence on high-dimension MMOPs. To address the above two issues, in this study, a novel coevolutionary framework termed CoMMEA for multimodal multi-objective optimization is proposed to better obtain both global and local PSs, and simultaneously, to improve the convergence performance in dealing with high-dimension MMOPs. Specifically, the CoMMEA introduces two archives to the search process, and coevolves them simultaneously through effective knowledge transfer. The convergence archive assists the CoMMEA to quickly approaching the Pareto optimal front (PF). The knowledge of the converged solutions is then transferred to the diversity archive which utilizes the local convergence indicator and the $\epsilon$-dominance-based method to obtain global and local PSs effectively. Experimental results show that CoMMEA is competitive compared to seven state-of-the-art MMEAs on fifty-four complex MMOPs.
Keyword: CLIP
3D-LDM: Neural Implicit 3D Shape Generation with Latent Diffusion Models
Authors: Gimin Nam, Mariem Khlifi, Andrew Rodriguez, Alberto Tono, Linqi Zhou, Paul Guerrero
Subjects: Computer Vision and Pattern Recognition (cs.CV)
Abstract
Diffusion models have shown great promise for image generation, beating GANs in terms of generation diversity, with comparable image quality. However, their application to 3D shapes has been limited to point or voxel representations that can in practice not accurately represent a 3D surface. We propose a diffusion model for neural implicit representations of 3D shapes that operates in the latent space of an auto-decoder. This allows us to generate diverse and high quality 3D surfaces. We additionally show that we can condition our model on images or text to enable image-to-3D generation and text-to-3D generation using CLIP embeddings. Furthermore, adding noise to the latent codes of existing shapes allows us to explore shape variations.
Focus! Relevant and Sufficient Context Selection for News Image Captioning
Authors: Mingyang Zhou, Grace Luo, Anna Rohrbach, Zhou Yu
Subjects: Computer Vision and Pattern Recognition (cs.CV); Computation and Language (cs.CL)
Abstract
News Image Captioning requires describing an image by leveraging additional context from a news article. Previous works only coarsely leverage the article to extract the necessary context, which makes it challenging for models to identify relevant events and named entities. In our paper, we first demonstrate that by combining more fine-grained context that captures the key named entities (obtained via an oracle) and the global context that summarizes the news, we can dramatically improve the model's ability to generate accurate news captions. This begs the question, how to automatically extract such key entities from an image? We propose to use the pre-trained vision and language retrieval model CLIP to localize the visually grounded entities in the news article and then capture the non-visual entities via an open relation extraction model. Our experiments demonstrate that by simply selecting a better context from the article, we can significantly improve the performance of existing models and achieve new state-of-the-art performance on multiple benchmarks.
Weakly Supervised Annotations for Multi-modal Greeting Cards Dataset
Authors: Sidra Hanif, Longin Jan Latecki
Subjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
Abstract
In recent years, there is a growing number of pre-trained models trained on a large corpus of data and yielding good performance on various tasks such as classifying multimodal datasets. These models have shown good performance on natural images but are not fully explored for scarce abstract concepts in images. In this work, we introduce an image/text-based dataset called Greeting Cards. Dataset (GCD) that has abstract visual concepts. In our work, we propose to aggregate features from pretrained images and text embeddings to learn abstract visual concepts from GCD. This allows us to learn the text-modified image features, which combine complementary and redundant information from the multi-modal data streams into a single, meaningful feature. Secondly, the captions for the GCD dataset are computed with the pretrained CLIP-based image captioning model. Finally, we also demonstrate that the proposed the dataset is also useful for generating greeting card images using pre-trained text-to-image generation model.
3D-TOGO: Towards Text-Guided Cross-Category 3D Object Generation
Abstract
Text-guided 3D object generation aims to generate 3D objects described by user-defined captions, which paves a flexible way to visualize what we imagined. Although some works have been devoted to solving this challenging task, these works either utilize some explicit 3D representations (e.g., mesh), which lack texture and require post-processing for rendering photo-realistic views; or require individual time-consuming optimization for every single case. Here, we make the first attempt to achieve generic text-guided cross-category 3D object generation via a new 3D-TOGO model, which integrates a text-to-views generation module and a views-to-3D generation module. The text-to-views generation module is designed to generate different views of the target 3D object given an input caption. prior-guidance, caption-guidance and view contrastive learning are proposed for achieving better view-consistency and caption similarity. Meanwhile, a pixelNeRF model is adopted for the views-to-3D generation module to obtain the implicit 3D neural representation from the previously-generated views. Our 3D-TOGO model generates 3D objects in the form of the neural radiance field with good texture and requires no time-cost optimization for every single caption. Besides, 3D-TOGO can control the category, color and shape of generated 3D objects with the input caption. Extensive experiments on the largest 3D object dataset (i.e., ABO) are conducted to verify that 3D-TOGO can better generate high-quality 3D objects according to the input captions across 98 different categories, in terms of PSNR, SSIM, LPIPS and CLIP-score, compared with text-NeRF and Dreamfields.
Keyword: metric learning
A Geometric-Relational Deep Learning Framework for BIM Object Classification
Keyword: image retrieval
StructVPR: Distill Structural Knowledge with Weighting Samples for Visual Place Recognition
Information Retrieval from the Digitized Books
Keyword: self-supervised
Progressive Feature Upgrade in Semi-supervised Learning on Tabular Domain
ObjectStitch: Generative Object Compositing
FedCoCo: A Memory Efficient Federated Self-supervised Framework for On-Device Visual Representation Learning
CLeBPI: Contrastive Learning for Bug Priority Inference
General Framework for Self-Supervised Model Priming for Parameter-Efficient Fine-tuning
A General Purpose Supervisory Signal for Embodied Agents
DeepFT: Fault-Tolerant Edge Computing using a Self-Supervised Deep Surrogate Model
ColD Fusion: Collaborative Descent for Distributed Multitask Finetuning
Keyword: vision transformer
There is no result
Keyword: multimodal
UniT3D: A Unified Transformer for 3D Dense Captioning and Visual Grounding
Weakly Supervised Annotations for Multi-modal Greeting Cards Dataset
Masked Contrastive Pre-Training for Efficient Video-Text Retrieval
Coevolutionary Framework for Generalized Multimodal Multi-objective Optimization
Keyword: CLIP
3D-LDM: Neural Implicit 3D Shape Generation with Latent Diffusion Models
Focus! Relevant and Sufficient Context Selection for News Image Captioning
Weakly Supervised Annotations for Multi-modal Greeting Cards Dataset
3D-TOGO: Towards Text-Guided Cross-Category 3D Object Generation
Keyword: DALLE
There is no result