We present a neural network approach to transfer the motion from a single image of an articulated object to a rest-state (i.e., unarticulated) 3D model. Our network learns to predict the object's pose, part segmentation, and corresponding motion parameters to reproduce the articulation shown in the input image. The network is composed of three distinct branches that take a shared joint image-shape embedding and is trained end-to-end. Unlike previous methods, our approach is independent of the topology of the object and can work with objects from arbitrary categories. Our method, trained with only synthetic data, can be used to automatically animate a mesh, infer motion from real images, and transfer articulation to functionally similar but geometrically distinct 3D models at test time.
Federated learning (FL) is a method to train model with distributed data from numerous participants such as IoT devices. It inherently assumes a uniform capacity among participants. However, participants have diverse computational resources in practice due to different conditions such as different energy budgets or executing parallel unrelated tasks. It is necessary to reduce the computation overhead for participants with inefficient computational resources, otherwise they would be unable to finish the full training process. To address the computation heterogeneity, in this paper we propose a strategy for estimating local models without computationally intensive iterations. Based on it, we propose Computationally Customized Federated Learning (CCFL), which allows each participant to determine whether to perform conventional local training or model estimation in each round based on its current computational resources. Both theoretical analysis and exhaustive experiments indicate that CCFL has the same convergence rate as FedAvg without resource constraints. Furthermore, CCFL can be viewed of a computation-efficient extension of FedAvg that retains model performance while considerably reducing computation overhead.
A challenge in spoken language translation is that plenty of spoken content is long-form, but short units are necessary for obtaining high-quality translations. To address this mismatch, we fine-tune a general-purpose, large language model to split long ASR transcripts into segments that can be independently translated so as to maximize the overall translation quality. We compare to several segmentation strategies and find that our approach improves BLEU score on three languages by an average of 2.7 BLEU overall compared to an automatic punctuation baseline. Further, we demonstrate the effectiveness of two constrained decoding strategies to improve well-formedness of the model output from above 99% to 100%.
We introduce anchored radial observations (ARO), a novel shape encoding for learning neural field representation of shapes that is category-agnostic and generalizable amid significant shape variations. The main idea behind our work is to reason about shapes through partial observations from a set of viewpoints, called anchors. We develop a general and unified shape representation by employing a fixed set of anchors, via Fibonacci sampling, and designing a coordinate-based deep neural network to predict the occupancy value of a query point in space. Differently from prior neural implicit models, that use global shape feature, our shape encoder operates on contextual, query-specific features. To predict point occupancy, locally observed shape information from the perspective of the anchors surrounding the input query point are encoded and aggregated through an attention module, before implicit decoding is performed. We demonstrate the quality and generality of our network, coined ARO-Net, on surface reconstruction from sparse point clouds, with tests on novel and unseen object categories, "one-shape" training, and comparisons to state-of-the-art neural and classical methods for reconstruction and tessellation.
In this paper, we study the problem of visual grounding by considering both phrase extraction and grounding (PEG). In contrast to the previous phrase-known-at-test setting, PEG requires a model to extract phrases from text and locate objects from images simultaneously, which is a more practical setting in real applications. As phrase extraction can be regarded as a $1$D text segmentation problem, we formulate PEG as a dual detection problem and propose a novel DQ-DETR model, which introduces dual queries to probe different features from image and text for object prediction and phrase mask prediction. Each pair of dual queries is designed to have shared positional parts but different content parts. Such a design effectively alleviates the difficulty of modality alignment between image and text (in contrast to a single query design) and empowers Transformer decoder to leverage phrase mask-guided attention to improve performance. To evaluate the performance of PEG, we also propose a new metric CMAP (cross-modal average precision), analogous to the AP metric in object detection. The new metric overcomes the ambiguity of Recall@1 in many-box-to-one-phrase cases in phrase grounding. As a result, our PEG pre-trained DQ-DETR establishes new state-of-the-art results on all visual grounding benchmarks with a ResNet-101 backbone. For example, it achieves $91.04\%$ and $83.51\%$ in terms of recall rate on RefCOCO testA and testB with a ResNet-101 backbone. Code will be availabl at \url{https://github.com/IDEA-Research/DQ-DETR}.
In recent years, molecular graph representation learning (GRL) has drawn much more attention in molecular property prediction (MPP) problems. The existing graph methods have demonstrated that 3D geometric information is significant for better performance in MPP. However, accurate 3D structures are often costly and time-consuming to obtain, limiting the large-scale application of GRL. It is an intuitive solution to train with 3D to 2D knowledge distillation and predict with only 2D inputs. But some challenging problems remain open for 3D to 2D distillation. One is that the 3D view is quite distinct from the 2D view, and the other is that the gradient magnitudes of atoms in distillation are discrepant and unstable due to the variable molecular size. To address these challenging problems, we exclusively propose a distillation framework that contains global molecular distillation and local atom distillation. We also provide a theoretical insight to justify how to coordinate atom and molecular information, which tackles the drawback of variable molecular size for atom information distillation. Experimental results on two popular molecular datasets demonstrate that our proposed model achieves superior performance over other methods. Specifically, on the largest MPP dataset PCQM4Mv2 served as an "ImageNet Large Scale Visual Recognition Challenge" in the field of graph ML, the proposed method achieved a 6.9% improvement compared with the best works. And we obtained fourth place with the MAE of 0.0734 on the test-challenge set for OGB-LSC 2022 Graph Regression Task. We will release the code soon.
This paper presents the first significant work on directly predicting 3D face landmarks on neural radiance fields (NeRFs), without using intermediate representations such as 2D images, depth maps, or point clouds. Our 3D coarse-to-fine Face Landmarks NeRF (FLNeRF) model efficiently samples from the NeRF on the whole face with individual facial features for accurate landmarks. To mitigate the limited number of facial expressions in the available data, local and non-linear NeRF warp is applied at facial features in fine scale to simulate large emotions range, including exaggerated facial expressions (e.g., cheek blowing, wide opening mouth, eye blinking), for training FLNeRF. With such expression augmentation, our model can predict 3D landmarks not limited to the 20 discrete expressions given in the data. Robust 3D NeRF facial landmarks contribute to many downstream tasks. As an example, we modify MoFaNeRF to enable high-quality face editing and swapping using face landmarks on NeRF, allowing more direct control and wider range of complex expressions. Experiments show that the improved model using landmarks achieves comparable to better results.
Reference Expression Segmentation (RES) and Reference Expression Generation (REG) are mutually inverse tasks that can be naturally jointly trained. Though recent work has explored such joint training, the mechanism of how RES and REG can benefit each other is still unclear. In this paper, we propose a unified mutual supervision framework that enables two tasks to improve each other. Our mutual supervision contains two directions. On the one hand, Disambiguation Supervision leverages the expression unambiguity measurement provided by RES to enhance the language generation of REG. On the other hand, Generation Supervision uses expressions automatically generated by REG to scale up the training of RES. Such unified mutual supervision effectively improves two tasks by solving their bottleneck problems. Extensive experiments show that our approach significantly outperforms all existing methods on REG and RES tasks under the same setting, and detailed ablation studies demonstrate the effectiveness of all components in our framework.
Zero-shot transfer learning for document understanding is a crucial yet under-investigated scenario to help reduce the high cost involved in annotating document entities. We present a novel query-based framework, QueryForm, that extracts entity values from form-like documents in a zero-shot fashion. QueryForm contains a dual prompting mechanism that composes both the document schema and a specific entity type into a query, which is used to prompt a Transformer model to perform a single entity extraction task. Furthermore, we propose to leverage large-scale query-entity pairs generated from form-like webpages with weak HTML annotations to pre-train QueryForm. By unifying pre-training and fine-tuning into the same query-based framework, QueryForm enables models to learn from structured documents containing various entities and layouts, leading to better generalization to target document types without the need for target-specific training data. QueryForm sets new state-of-the-art average F1 score on both the XFUND (+4.6%~10.1%) and the Payment (+3.2%~9.5%) zero-shot benchmark, with a smaller model size and no additional image input.