Search Results for author: Kai Xu

Found 132 papers, 54 papers with code

InterFusion: Text-Driven Generation of 3D Human-Object Interaction

no code implementations22 Mar 2024 Sisi Dai, Wenhao Li, Haowen Sun, Haibin Huang, Chongyang Ma, Hui Huang, Kai Xu, Ruizhen Hu

In this study, we tackle the complex task of generating 3D human-object interactions (HOI) from textual descriptions in a zero-shot text-to-3D manner.

Human-Object Interaction Detection Object +1

Surface Reconstruction from Point Clouds via Grid-based Intersection Prediction

no code implementations21 Mar 2024 Hui Tian, Kai Xu

Surface reconstruction from point clouds is a crucial task in the fields of computer vision and computer graphics.

Surface Reconstruction

LAB: Large-Scale Alignment for ChatBots

no code implementations2 Mar 2024 Shivchander Sudalairaj, Abhishek Bhandwaldar, Aldo Pareja, Kai Xu, David D. Cox, Akash Srivastava

This work introduces LAB (Large-scale Alignment for chatBots), a novel methodology designed to overcome the scalability challenges in the instruction-tuning phase of large language model (LLM) training.

Instruction Following Language Modelling +2

Learning Dual-arm Object Rearrangement for Cartesian Robots

no code implementations21 Feb 2024 Shishun Zhang, Qijin She, Wenhao Li, Chenyang Zhu, Yongjun Wang, Ruizhen Hu, Kai Xu

To achieve the goal, the core idea is to develop an effective object-to-arm task assignment strategy for minimizing the cumulative task execution time and maximizing the dual-arm cooperation efficiency.

Computational Efficiency Object +1

Conversational Crowdsensing: A Parallel Intelligence Powered Novel Sensing Approach

no code implementations4 Feb 2024 Zhengqiu Zhu, Yong Zhao, Bin Chen, Sihang Qiu, Kai Xu, Quanjun Yin, Jincai Huang, Zhong Liu, Fei-Yue Wang

The transition from CPS-based Industry 4. 0 to CPSS-based Industry 5. 0 brings new requirements and opportunities to current sensing approaches, especially in light of recent progress in Chatbots and Large Language Models (LLMs).

Scheduling

GliDe with a CaPE: A Low-Hassle Method to Accelerate Speculative Decoding

no code implementations3 Feb 2024 Cunxiao Du, Jing Jiang, Xu Yuanchen, Jiawei Wu, Sicheng Yu, Yongqi Li, Shenggui Li, Kai Xu, Liqiang Nie, Zhaopeng Tu, Yang You

Speculative decoding is a relatively new decoding framework that leverages small and efficient draft models to reduce the latency of LLMs.

The Risk of Federated Learning to Skew Fine-Tuning Features and Underperform Out-of-Distribution Robustness

no code implementations25 Jan 2024 Mengyao Du, Miao Zhang, Yuwen Pu, Kai Xu, Shouling Ji, Quanjun Yin

To tackle the scarcity and privacy issues associated with domain-specific datasets, the integration of federated learning in conjunction with fine-tuning has emerged as a practical solution.

Federated Learning

DiffusionEdge: Diffusion Probabilistic Model for Crisp Edge Detection

1 code implementation4 Jan 2024 Yunfan Ye, Kai Xu, Yuhang Huang, Renjiao Yi, Zhiping Cai

With the recent success of the diffusion probabilistic model (DPM), we found it is especially suitable for accurate and crisp edge detection since the denoising process is directly applied to the original image size.

Denoising Edge Detection

Latent Space Explorer: Visual Analytics for Multimodal Latent Space Exploration

no code implementations1 Dec 2023 Bum Chul Kwon, Samuel Friedman, Kai Xu, Steven A Lubitz, Anthony Philippakis, Puneet Batra, Patrick T Ellinor, Kenney Ng

Machine learning models built on training data with multiple modalities can reveal new insights that are not accessible through unimodal datasets.

Polyhedral Surface: Self-supervised Point Cloud Reconstruction Based on Polyhedral Surface

no code implementations23 Oct 2023 Hui Tian, Kai Xu

This method provides more flexible to represent sharp feature and surface boundary on open surface.

Point cloud reconstruction

Accurate and Fast Compressed Video Captioning

1 code implementation ICCV 2023 Yaojie Shen, Xin Gu, Kai Xu, Heng Fan, Longyin Wen, Libo Zhang

Addressing this, we study video captioning from a different perspective in compressed domain, which brings multi-fold advantages over the existing pipeline: 1) Compared to raw images from the decoded video, the compressed video, consisting of I-frames, motion vectors and residuals, is highly distinguishable, which allows us to leverage the entire video for learning without manual sampling through a specialized model design; 2) The captioning model is more efficient in inference as smaller and less redundant information is processed.

Video Captioning

Prompt-based Context- and Domain-aware Pretraining for Vision and Language Navigation

no code implementations7 Sep 2023 Ting Liu, Yue Hu, Wansen Wu, Youkai Wang, Kai Xu, Quanjun Yin

In the indoor-aware stage, we apply an efficient tuning paradigm to learn deep visual prompts from an indoor dataset, in order to augment pretrained models with inductive biases towards indoor environments.

Contrastive Learning Vision and Language Navigation +1

Stage-by-stage Wavelet Optimization Refinement Diffusion Model for Sparse-View CT Reconstruction

1 code implementation30 Aug 2023 Kai Xu, Shiyu Lu, Bin Huang, Weiwen Wu, Qiegen Liu

Diffusion models have emerged as potential tools to tackle the challenge of sparse-view CT reconstruction, displaying superior performance compared to conventional methods.

SuperUDF: Self-supervised UDF Estimation for Surface Reconstruction

1 code implementation28 Aug 2023 Hui Tian, Chenyang Zhu, Yifei Shi, Kai Xu

The key insight is that if the UDF is estimated correctly, the 3D points should be locally projected onto the underlying surface following the gradient of the UDF.

Surface Reconstruction

MIPS-Fusion: Multi-Implicit-Submaps for Scalable and Robust Online Neural RGB-D Reconstruction

no code implementations17 Aug 2023 Yijie Tang, Jiazhao Zhang, Zhinan Yu, He Wang, Kai Xu

For the first time, randomized optimization is made possible in neural tracking with several key designs to the learning process, enabling efficient and robust tracking even under fast camera motions.

RGB-D Reconstruction

2D3D-MATR: 2D-3D Matching Transformer for Detection-free Registration between Images and Point Clouds

1 code implementation ICCV 2023 Minhao Li, Zheng Qin, Zhirui Gao, Renjiao Yi, Chenyang Zhu, Yulan Guo, Kai Xu

The commonly adopted detect-then-match approach to registration finds difficulties in the cross-modality cases due to the incompatible keypoint detection and inconsistent feature description.

Keypoint Detection Patch Matching

PlaneRecTR: Unified Query Learning for 3D Plane Recovery from a Single View

1 code implementation ICCV 2023 Jingjia Shi, Shuaifeng Zhi, Kai Xu

3D plane recovery from a single image can usually be divided into several subtasks of plane detection, segmentation, parameter estimation and possibly depth estimation.

Depth Estimation Segmentation

RayMVSNet++: Learning Ray-based 1D Implicit Fields for Accurate Multi-View Stereo

no code implementations16 Jul 2023 Yifei Shi, Junhua Xi, Dewen Hu, Zhiping Cai, Kai Xu

In particular, we propose RayMVSNet which learns sequential prediction of a 1D implicit field along each camera ray with the zero-crossing point indicating scene depth.

Depth Estimation Multi-Task Learning +1

Tensorformer: Normalized Matrix Attention Transformer for High-quality Point Cloud Reconstruction

1 code implementation28 Jun 2023 Hui Tian, Zheng Qin, Renjiao Yi, Chenyang Zhu, Kai Xu

Surface reconstruction from raw point clouds has been studied for decades in the computer graphics community, which is highly demanded by modeling and rendering applications nowadays.

Point cloud reconstruction Surface Reconstruction

Decoupled Diffusion Models: Image to Zero and Zero to Noise

no code implementations23 Jun 2023 Yuhang Huang, Zheng Qin, Xinwang Liu, Kai Xu

We find that decoupling the diffusion process reduces the learning difficulty and the explicit transition probability improves the generative speed significantly.

Denoising Image Generation

Understanding how Differentially Private Generative Models Spend their Privacy Budget

no code implementations18 May 2023 Georgi Ganev, Kai Xu, Emiliano De Cristofaro

Generative models trained with Differential Privacy (DP) are increasingly used to produce synthetic data while reducing privacy risks.

Estimating the Density Ratio between Distributions with High Discrepancy using Multinomial Logistic Regression

no code implementations1 May 2023 Akash Srivastava, Seungwook Han, Kai Xu, Benjamin Rhodes, Michael U. Gutmann

We show that if these auxiliary densities are constructed such that they overlap with $p$ and $q$, then a multi-class logistic regression allows for estimating $\log p/q$ on the domain of any of the $K+2$ distributions and resolves the distribution shift problems of the current state-of-the-art methods.

Binary Classification Density Ratio Estimation +4

Enhancing Video Super-Resolution via Implicit Resampling-based Alignment

1 code implementation arXiv 2024 Kai Xu, Ziwei Yu, Xin Wang, Michael Bi Mi, Angela Yao

We show that bilinear interpolation inherently attenuates high-frequency information while an MLP-based coordinate network can approximate more frequencies.

Video Super-Resolution

Weakly-supervised Single-view Image Relighting

no code implementations CVPR 2023 Renjiao Yi, Chenyang Zhu, Kai Xu

For re-rendering, we propose a differentiable specular rendering layer to render low-frequency non-Lambertian materials under various illuminations of spherical harmonics.

Image Relighting Inverse Rendering

SOCS: Semantically-aware Object Coordinate Space for Category-Level 6D Object Pose Estimation under Large Shape Variations

no code implementations ICCV 2023 Boyan Wan, Yifei Shi, Kai Xu

We propose Semantically-aware Object Coordinate Space (SOCS) built by warping-and-aligning the objects guided by a sparse set of keypoints with semantically meaningful correspondence.

6D Pose Estimation 6D Pose Estimation using RGB +2

Deep Graph-based Spatial Consistency for Robust Non-rigid Point Cloud Registration

1 code implementation CVPR 2023 Zheng Qin, Hao Yu, Changjian Wang, Yuxing Peng, Kai Xu

We first design a local spatial consistency measure over the deformation graph of the point cloud, which evaluates the spatial compatibility only between the correspondences in the vicinity of a graph node.

Point Cloud Registration

Learning Accurate Template Matching with Differentiable Coarse-to-Fine Correspondence Refinement

1 code implementation15 Mar 2023 Zhirui Gao, Renjiao Yi, Zheng Qin, Yunfan Ye, Chenyang Zhu, Kai Xu

To tackle the challenges, we propose an accurate template matching method based on differentiable coarse-to-fine correspondence refinement.

Robotic Grasping Template Matching

Multi-Symmetry Ensembles: Improving Diversity and Generalization via Opposing Symmetries

1 code implementation4 Mar 2023 Charlotte Loh, Seungwook Han, Shivchander Sudalairaj, Rumen Dangovski, Kai Xu, Florian Wenzel, Marin Soljacic, Akash Srivastava

In this work, we present Multi-Symmetry Ensembles (MSE), a framework for constructing diverse ensembles by capturing the multiplicity of hypotheses along symmetry axes, which explore the hypothesis space beyond stochastic perturbations of model weights and hyperparameters.

Representation Learning Uncertainty Quantification

SSR-2D: Semantic 3D Scene Reconstruction from 2D Images

no code implementations7 Feb 2023 Junwen Huang, Alexey Artemov, Yujin Chen, Shuaifeng Zhi, Kai Xu, Matthias Nießner

In this work, we explore a central 3D scene modeling task, namely, semantic scene reconstruction without using any 3D annotations.

3D Scene Reconstruction Colorization +1

Edge Preserving Implicit Surface Representation of Point Clouds

no code implementations12 Jan 2023 Xiaogang Wang, Yuhang Cheng, Liang Wang, Jiangbo Lu, Kai Xu, GuoQiang Xiao

Among them, the differential Laplican regularizer can effectively alleviate the implicit surface unsmoothness caused by the point cloud quality deteriorates; Meanwhile, in order to reduce the excessive smoothing at the edge regions of implicit suface, we proposed a dynamic edge extract strategy for sampling near the sharp edge of point cloud, which can effectively avoid the Laplacian regularizer from smoothing all regions.

3D Reconstruction Surface Reconstruction

BUFFER: Balancing Accuracy, Efficiency, and Generalizability in Point Cloud Registration

1 code implementation CVPR 2023 Sheng Ao, Qingyong Hu, Hanyun Wang, Kai Xu, Yulan Guo

Extensive experiments on real-world scenarios demonstrate that our method achieves the best of both worlds in accuracy, efficiency, and generalization.

Computational Efficiency Open-Ended Question Answering +1

Learning Physically Realizable Skills for Online Packing of General 3D Shapes

1 code implementation5 Dec 2022 Hang Zhao, Zherong Pan, Yang Yu, Kai Xu

We study the problem of learning online packing skills for irregular 3D shapes, which is arguably the most challenging setting of bin packing problems.

Action Generation Reinforcement Learning (RL)

Multi-resolution Monocular Depth Map Fusion by Self-supervised Gradient-based Composition

1 code implementation3 Dec 2022 Yaqiao Dai, Renjiao Yi, Chenyang Zhu, Hongjun He, Kai Xu

Therefore, we propose a novel depth map fusion module to combine the advantages of estimations with multi-resolution inputs.

Monocular Depth Estimation

3D-Aware Object Goal Navigation via Simultaneous Exploration and Identification

no code implementations CVPR 2023 Jiazhao Zhang, Liu Dai, Fanpeng Meng, Qingnan Fan, Xuelin Chen, Kai Xu, He Wang

However, leveraging 3D scene representation can be prohibitively unpractical for policy learning in this floor-level task, due to low sample efficiency and expensive computational cost.

NIFT: Neural Interaction Field and Template for Object Manipulation

no code implementations20 Oct 2022 Zeyu Huang, Juzhan Xu, Sisi Dai, Kai Xu, Hao Zhang, Hui Huang, Ruizhen Hu

Given a few object manipulation demos, NIFT guides the generation of the interaction imitation for a new object instance by matching the Neural Interaction Template (NIT) extracted from the demos in the target Neural Interaction Field (NIF) defined for the new object.

Descriptive Imitation Learning +1

6DOF Pose Estimation of a 3D Rigid Object based on Edge-enhanced Point Pair Features

no code implementations17 Sep 2022 Chenyi Liu, Fei Chen, Lu Deng, Renjiao Yi, Lintao Zheng, Chenyang Zhu, Jia Wang, Kai Xu

We introduce a well-targeted down-sampling strategy that focuses more on edge area for efficient feature extraction of complex geometry.

6D Pose Estimation

HybridGNN: Learning Hybrid Representation in Multiplex Heterogeneous Networks

no code implementations3 Aug 2022 Tiankai Gu, Chaokun Wang, Cheng Wu, Jingcao Xu, Yunkai Lou, Changping Wang, Kai Xu, Can Ye, Yang song

One of the most important tasks in recommender systems is to predict the potential connection between two nodes under a specific edge type (i. e., relationship).

Recommendation Systems

AutoTransition: Learning to Recommend Video Transition Effects

1 code implementation27 Jul 2022 Yaojie Shen, Libo Zhang, Kai Xu, Xiaojie Jin

First we learn the embedding of video transitions through a video transition classification task.

Retrieval Video Editing

Repairing Systematic Outliers by Learning Clean Subspaces in VAEs

1 code implementation17 Jul 2022 Simao Eduardo, Kai Xu, Alfredo Nazabal, Charles Sutton

Seeing as a systematic outlier is a combination of patterns of a clean instance and systematic error patterns, our main insight is that inliers can be modelled by a smaller representation (subspace) in a model than outliers.

Outlier Detection

dpart: Differentially Private Autoregressive Tabular, a General Framework for Synthetic Data Generation

2 code implementations12 Jul 2022 Sofiane Mahiou, Kai Xu, Georgi Ganev

We propose a general, flexible, and scalable framework dpart, an open source Python library for differentially private synthetic data generation.

Synthetic Data Generation

RayMVSNet: Learning Ray-based 1D Implicit Fields for Accurate Multi-View Stereo

no code implementations CVPR 2022 Junhua Xi, Yifei Shi, Yijie Wang, Yulan Guo, Kai Xu

In particular, we propose RayMVSNet which learns sequential prediction of a 1D implicit field along each camera ray with the zero-crossing point indicating scene depth.

Multi-Task Learning

Learning High-DOF Reaching-and-Grasping via Dynamic Representation of Gripper-Object Interaction

no code implementations3 Apr 2022 Qijin She, Ruizhen Hu, Juzhan Xu, Min Liu, Kai Xu, Hui Huang

To resolve the sample efficiency issue in learning the high-dimensional and complex control of dexterous grasping, we propose an effective representation of grasping state characterizing the spatial interaction between the gripper and the target object.

Object

DisARM: Displacement Aware Relation Module for 3D Detection

no code implementations CVPR 2022 Yao Duan, Chenyang Zhu, Yuqing Lan, Renjiao Yi, Xinwang Liu, Kai Xu

However, adopting relations between all the object or patch proposals for detection is inefficient, and an imbalanced combination of local and global relations brings extra noise that could mislead the training.

3D Object Detection object-detection +1

3DRM:Pair-wise relation module for 3D object detection

1 code implementation20 Feb 2022 Yuqing Lan, Yao Duan, Yifei Shi, Hui Huang, Kai Xu

Context has proven to be one of the most important factors in object layout reasoning for 3D scene understanding.

3D Object Detection Object +3

ARM3D: Attention-based relation module for indoor 3D object detection

1 code implementation20 Feb 2022 Yuqing Lan, Yao Duan, Chenyi Liu, Chenyang Zhu, Yueshan Xiong, Hui Huang, Kai Xu

In the field of 3D object detection, previous methods have been taking the advantage of context encoding, graph embedding, or explicit relation reasoning to extract relation context.

3D Object Detection Graph Embedding +3

RIM-Net: Recursive Implicit Fields for Unsupervised Learning of Hierarchical Shape Structures

1 code implementation CVPR 2022 Chengjie Niu, Manyi Li, Kai Xu, Hao Zhang

Each level of the tree corresponds to an assembly of shape parts, represented as implicit functions, to reconstruct the input shape.

STEdge: Self-training Edge Detection with Multi-layer Teaching and Regularization

1 code implementation13 Jan 2022 Yunfan Ye, Renjiao Yi, Zhiping Cai, Kai Xu

In particular, we impose a consistency regularization which enforces the outputs from each of the multiple layers to be consistent for the input image and its perturbed counterpart.

Edge Detection Self-Supervised Learning

Box2Seg: Learning Semantics of 3D Point Clouds with Box-Level Supervision

no code implementations9 Jan 2022 Yan Liu, Qingyong Hu, Yinjie Lei, Kai Xu, Jonathan Li, Yulan Guo

In this paper, we introduce a neural architecture, termed Box2Seg, to learn point-level semantics of 3D point clouds with bounding box-level supervision.

Semantic Segmentation

Decoupling Makes Weakly Supervised Local Feature Better

1 code implementation CVPR 2022 Kunhong Li, Longguang Wang, Li Liu, Qing Ran, Kai Xu, Yulan Guo

Weakly supervised learning can help local feature methods to overcome the obstacle of acquiring a large-scale dataset with densely labeled correspondences.

Camera Localization Image Matching +1

A Bayesian-Symbolic Approach to Reasoning and Learning in Intuitive Physics

no code implementations NeurIPS 2021 Kai Xu, Akash Srivastava, Dan Gutfreund, Felix Sosa, Tomer Ullman, Josh Tenenbaum, Charles Sutton

In this paper, we propose a Bayesian-symbolic framework (BSP) for physical reasoning and learning that is close to human-level sample-efficiency and accuracy.

Bayesian Inference Bilevel Optimization +3

Targeted Neural Dynamical Modeling

2 code implementations NeurIPS 2021 Cole Hurwitz, Akash Srivastava, Kai Xu, Justin Jude, Matthew G. Perich, Lee E. Miller, Matthias H. Hennig

These approaches, however, are limited in their ability to capture the underlying neural dynamics (e. g. linear) and in their ability to relate the learned dynamics back to the observed behaviour (e. g. no time lag).

Learning Efficient Online 3D Bin Packing on Packing Configuration Trees

1 code implementation ICLR 2022 Hang Zhao, Yang Yu, Kai Xu

PCT is a full-fledged description of the state and action space of bin packing which can support packing policy learning based on deep reinforcement learning (DRL).

3D Bin Packing

Scaling Densities For Improved Density Ratio Estimation

no code implementations29 Sep 2021 Akash Srivastava, Seungwook Han, Benjamin Rhodes, Kai Xu, Michael U. Gutmann

As such, estimating density ratios accurately using only samples from $p$ and $q$ is of high significance and has led to a flurry of recent work in this direction.

Binary Classification Density Ratio Estimation

Learning Practically Feasible Policies for Online 3D Bin Packing

2 code implementations31 Aug 2021 Hang Zhao, Chenyang Zhu, Xin Xu, Hui Huang, Kai Xu

In this problem, the items are delivered to the agent without informing the full sequence information.

3D Bin Packing Collision Avoidance

Accelerating Video Object Segmentation with Compressed Video

2 code implementations CVPR 2022 Kai Xu, Angela Yao

We propose an efficient plug-and-play acceleration framework for semi-supervised video object segmentation by exploiting the temporal redundancies in videos presented by the compressed bitstream.

Object Segmentation +3

Deep Tiny Network for Recognition-Oriented Face Image Quality Assessment

no code implementations9 Jun 2021 Baoyun Peng, Min Liu, Zhaoning Zhang, Kai Xu, Dongsheng Li

Based on the proposed quality measurement, we propose a deep Tiny Face Quality network (tinyFQnet) to learn a quality prediction function from data.

Face Image Quality Face Image Quality Assessment +1

ROSEFusion: Random Optimization for Online Dense Reconstruction under Fast Camera Motion

no code implementations12 May 2021 Jiazhao Zhang, Chenyang Zhu, Lintao Zheng, Kai Xu

We propose to tackle the difficulties of fast-motion camera tracking in the absence of inertial measurements using random optimization, in particular, the Particle Filter Optimization (PFO).

Pose Tracking

Potential Convolution: Embedding Point Clouds into Potential Fields

no code implementations5 Apr 2021 Dengsheng Chen, Haowen Deng, Jun Li, Duo Li, Yao Duan, Kai Xu

In this work, rather than defining a continuous or discrete kernel, we directly embed convolutional kernels into the learnable potential fields, giving rise to potential convolution.

3D Shape Classification Scene Segmentation

Learning Fine-Grained Segmentation of 3D Shapes without Part Labels

no code implementations CVPR 2021 Xiaogang Wang, Xun Sun, Xinyu Cao, Kai Xu, Bin Zhou

Learning-based 3D shape segmentation is usually formulated as a semantic labeling problem, assuming that all parts of training shapes are annotated with a given set of tags.

Clustering Deep Clustering +1

StablePose: Learning 6D Object Poses from Geometrically Stable Patches

no code implementations CVPR 2021 Yifei Shi, Junwen Huang, Xin Xu, Yifan Zhang, Kai Xu

According to the theory of geometric stability analysis, a minimal set of three planar/cylindrical patches are geometrically stable and determine the full 6DoFs of the object pose.

6D Pose Estimation using RGB Object +1

Selective Sensing: A Data-driven Nonuniform Subsampling Approach for Computation-free On-Sensor Data Dimensionality Reduction

no code implementations1 Jan 2021 Zhikang Zhang, Kai Xu, Fengbo Ren

In this paper, we propose a selective sensing framework that adopts the novel concept of data-driven nonuniform subsampling to reduce the dimensionality of acquired signals while retaining the information of interest in a computation-free fashion.

Compressive Sensing Dimensionality Reduction +1

A Bayesian-Symbolic Approach to Learning and Reasoning for Intuitive Physics

no code implementations1 Jan 2021 Kai Xu, Akash Srivastava, Dan Gutfreund, Felix Sosa, Tomer Ullman, Joshua B. Tenenbaum, Charles Sutton

As such, learning the laws is then reduced to symbolic regression and Bayesian inference methods are used to obtain the distribution of unobserved properties.

Bayesian Inference Common Sense Reasoning +2

Study of charmonium-like and fully-charm tetraquark spectroscopy

no code implementations31 Dec 2020 Zheng Zhao, Kai Xu, Attaphon Kaewsnod, Xuyang Liu, Ayut Limphirat, Yupeng Yan

The masses of tetraquark states of all $qc\bar q \bar c$ and $cc\bar c \bar c$ quark configurations are evaluated in a constituent quark model, where the Cornell-like potential and one-gluon exchange spin-spin coupling are employed.

High Energy Physics - Phenomenology

Hausdorff Point Convolution with Geometric Priors

no code implementations24 Dec 2020 Pengdi Huang, Liqiang Lin, Fuyou Xue, Kai Xu, Danny Cohen-Or, Hui Huang

We show that HPC constitutes a powerful point feature learning with a rather compact set of only four types of geometric priors as kernels.

Semantic Segmentation

On Learning the Right Attention Point for Feature Enhancement

no code implementations11 Dec 2020 Liqiang Lin, Pengdi Huang, Chi-Wing Fu, Kai Xu, Hao Zhang, Hui Huang

We present a novel attention-based mechanism to learn enhanced point features for point cloud processing tasks, e. g., classification and segmentation.

Classification Point Cloud Classification +1

A geometry-based relaxation algorithm for equilibrating a trivalent polygonal network in two dimensions and its implications

no code implementations5 Nov 2020 Kai Xu

The succeed of simulation strongly supports the ellipse packing hypothesis that was proposed to explain the dynamic behaviors of a trivalent 2D structure.

Biological Physics Adaptation and Self-Organizing Systems Cell Behavior

Improving the Reconstruction of Disentangled Representation Learners via Multi-Stage Modelling

no code implementations25 Oct 2020 Akash Srivastava, Yamini Bansal, Yukun Ding, Cole Hurwitz, Kai Xu, Bernhard Egger, Prasanna Sattigeri, Josh Tenenbaum, David D. Cox, Dan Gutfreund

Current autoencoder-based disentangled representation learning methods achieve disentanglement by penalizing the (aggregate) posterior to encourage statistical independence of the latent factors.

Disentanglement

WarpCore: A Library for fast Hash Tables on GPUs

1 code implementation16 Sep 2020 Daniel Jünger, Robin Kobus, André Müller, Christian Hundt, Kai Xu, Weiguo Liu, Bertil Schmidt

The rapidly growing amount of data emerging in many fields motivated the need for accelerated hash tables designed for modern parallel architectures.

Distributed, Parallel, and Cluster Computing

SymmetryNet: Learning to Predict Reflectional and Rotational Symmetries of 3D Shapes from Single-View RGB-D Images

no code implementations2 Aug 2020 Yifei Shi, Junwen Huang, Hongjia Zhang, Xin Xu, Szymon Rusinkiewicz, Kai Xu

We propose an end-to-end deep neural network which is able to predict both reflectional and rotational symmetries of 3D objects present in the input RGB-D image.

Multi-Task Learning Symmetry Detection

AReLU: Attention-based Rectified Linear Unit

1 code implementation24 Jun 2020 Dengsheng Chen, Jun Li, Kai Xu

Adding the attention module with a rectified linear unit (ReLU) results in an amplification of positive elements and a suppression of negative ones, both with learned, data-adaptive parameters.

Meta-Learning Transfer Learning

Fusion-Aware Point Convolution for Online Semantic 3D Scene Segmentation

1 code implementation CVPR 2020 Jiazhao Zhang, Chenyang Zhu, Lintao Zheng, Kai Xu

Online semantic 3D segmentation in company with real-time RGB-D reconstruction poses special challenges such as how to perform 3D convolution directly over the progressively fused 3D geometric data, and how to smartly fuse information from frame to frame.

RGB-D Reconstruction Scene Segmentation

Learning in the Frequency Domain

4 code implementations CVPR 2020 Kai Xu, Minghai Qin, Fei Sun, Yuhao Wang, Yen-Kuang Chen, Fengbo Ren

Experiment results show that learning in the frequency domain with static channel selection can achieve higher accuracy than the conventional spatial downsampling approach and meanwhile further reduce the input data size.

Instance Segmentation Semantic Segmentation

DynamicPPL: Stan-like Speed for Dynamic Probabilistic Models

2 code implementations7 Feb 2020 Mohamed Tarek, Kai Xu, Martin Trapp, Hong Ge, Zoubin Ghahramani

Since DynamicPPL is a modular, stand-alone library, any probabilistic programming system written in Julia, such as Turing. jl, can use DynamicPPL to specify models and trace their model parameters.

Probabilistic Programming

Deep Differentiable Grasp Planner for High-DOF Grippers

no code implementations4 Feb 2020 Min Liu, Zherong Pan, Kai Xu, Kanishka Ganguly, Dinesh Manocha

We present an end-to-end algorithm for training deep neural networks to grasp novel objects.

Robotics

Learning Canonical Shape Space for Category-Level 6D Object Pose and Size Estimation

no code implementations CVPR 2020 Dengsheng Chen, Jun Li, Zheng Wang, Kai Xu

To tackle intra-class shape variations, we learn canonical shape space (CASS), a unified representation for a large variety of instances of a certain object category.

3D Shape Representation Generating 3D Point Clouds +1

Constructing the F-Graph with a Symmetric Constraint for Subspace Clustering

no code implementations17 Dec 2019 Kai Xu, Xiao-Jun Wu, Wen-Bo Hu

Based on further studying the low-rank subspace clustering (LRSC) and L2-graph subspace clustering algorithms, we propose a F-graph subspace clustering algorithm with a symmetric constraint (FSSC), which constructs a new objective function with a symmetric constraint basing on F-norm, whose the most significant advantage is to obtain a closed-form solution of the coefficient matrix.

Clustering Face Clustering +1

Reinforcement Learning-based Visual Navigation with Information-Theoretic Regularization

1 code implementation9 Dec 2019 Qiaoyun Wu, Kai Xu, Jun Wang, Mingliang Xu, Dinesh Manocha

The regularization maximizes the mutual information between navigation actions and visual observation transforms of an agent, thus promoting more informed navigation decisions.

Robotics

Decoupling Features and Coordinates for Few-shot RGB Relocalization

no code implementations26 Nov 2019 Siyan Dong, Songyin Wu, Yixin Zhuang, Kai Xu, Shanghang Zhang, Baoquan Chen

To address this issue, we approach camera relocalization with a decoupled solution where feature extraction, coordinate regression, and pose estimation are performed separately.

Camera Relocalization Pose Estimation +1

AdvancedHMC.jl: A robust, modular and efficient implementation of advanced HMC algorithms

1 code implementation pproximateinference AABI Symposium 2019 Kai Xu, Hong Ge, Will Tebbutt, Mohamed Tarek, Martin Trapp, Zoubin Ghahramani

Stan's Hamilton Monte Carlo (HMC) has demonstrated remarkable sampling robustness and efficiency in a wide range of Bayesian inference problems through carefully crafted adaption schemes to the celebrated No-U-Turn sampler (NUTS) algorithm.

Bayesian Inference Benchmarking

Rescan: Inductive Instance Segmentation for Indoor RGBD Scans

no code implementations ICCV 2019 Maciej Halber, Yifei Shi, Kai Xu, Thomas Funkhouser

In depth-sensing applications ranging from home robotics to AR/VR, it will be common to acquire 3D scans of interior spaces repeatedly at sparse time intervals (e. g., as part of regular daily use).

Instance Segmentation Segmentation +1

Multiple instance dense connected convolution neural network for aerial image scene classification

no code implementations22 Aug 2019 Qi Bi, Kun Qin, Zhili Li, Han Zhang, Kai Xu

While the current convolution neural network tends to extract global features and global semantic information in a scene, the geo-spatial objects can be located at anywhere in an aerial image scene and their spatial arrangement tends to be more complicated.

General Classification Scene Classification

Building change detection based on multi-scale filtering and grid partition

no code implementations22 Aug 2019 Qi Bi, Kun Qin, Han Zhang, Wenjun Han, Zhili Li, Kai Xu

Exhaustive experiments indicate that the proposed method can detect building change types directly and outperform the current multi-index learning method.

Change Detection

Active Scene Understanding via Online Semantic Reconstruction

no code implementations18 Jun 2019 Lintao Zheng, Chenyang Zhu, Jiazhao Zhang, Hang Zhao, Hui Huang, Matthias Niessner, Kai Xu

In our method, the exploratory robot scanning is both driven by and targeting at the recognition and segmentation of semantic objects from the scene.

Scene Understanding Semantic Segmentation

NeoNav: Improving the Generalization of Visual Navigation via Generating Next Expected Observations

1 code implementation17 Jun 2019 Qiaoyun Wu, Dinesh Manocha, Jun Wang, Kai Xu

First, the latent distribution is conditioned on current observations and the target view, leading to a model-based, target-driven navigation.

Visual Navigation

Learning Part Generation and Assembly for Structure-aware Shape Synthesis

no code implementations16 Jun 2019 Jun Li, Chengjie Niu, Kai Xu

Enlightened by the fact that 3D shape structure is characterized as part composition and placement, we propose to model 3D shape variations with a part-aware deep generative network, coined as PAGENet.

Learning Phase Competition for Traffic Signal Control

1 code implementation12 May 2019 Guanjie Zheng, Yuanhao Xiong, Xinshi Zang, Jie Feng, Hua Wei, Huichu Zhang, Yong Li, Kai Xu, Zhenhui Li

Increasingly available city data and advanced learning techniques have empowered people to improve the efficiency of our city functions.

Reinforcement Learning (RL)

CoLight: Learning Network-level Cooperation for Traffic Signal Control

4 code implementations11 May 2019 Hua Wei, Nan Xu, Huichu Zhang, Guanjie Zheng, Xinshi Zang, Chacha Chen, Wei-Nan Zhang, Yanmin Zhu, Kai Xu, Zhenhui Li

To enable cooperation of traffic signals, in this paper, we propose a model, CoLight, which uses graph attentional networks to facilitate communication.

Multi-agent Reinforcement Learning

Spatiotemporal CNN for Video Object Segmentation

1 code implementation CVPR 2019 Kai Xu, Longyin Wen, Guorong Li, Liefeng Bo, Qingming Huang

Specifically, the temporal coherence branch pretrained in an adversarial fashion from unlabeled video data, is designed to capture the dynamic appearance and motion cues of video sequences to guide object segmentation.

Object Segmentation +5

AdaCoSeg: Adaptive Shape Co-Segmentation with Group Consistency Loss

no code implementations CVPR 2020 Chenyang Zhu, Kai Xu, Siddhartha Chaudhuri, Li Yi, Leonidas Guibas, Hao Zhang

While the part prior network can be trained with noisy and inconsistently segmented shapes, the final output of AdaCoSeg is a consistent part labeling for the input set, with each shape segmented into up to (a user-specified) K parts.

Instance Segmentation Segmentation +1

Shape2Motion: Joint Analysis of Motion Parts and Attributes from 3D Shapes

1 code implementation CVPR 2019 Xiaogang Wang, Bin Zhou, Yahao Shi, Xiaowu Chen, Qinping Zhao, Kai Xu

For the task of mobility analysis of 3D shapes, we propose joint analysis for simultaneous motion part segmentation and motion attribute estimation, taking a single 3D model as input.

Attribute Segmentation

PartNet: A Recursive Part Decomposition Network for Fine-grained and Hierarchical Shape Segmentation

no code implementations CVPR 2019 Fenggen Yu, Kun Liu, Yan Zhang, Chenyang Zhu, Kai Xu

Meanwhile, to increase the segmentation accuracy at each node, we enhance the recursive contextual feature with the shape feature extracted for the corresponding part.

Ranked #14 on 3D Part Segmentation on ShapeNet-Part (Class Average IoU metric)

3D Instance Segmentation 3D Part Segmentation +1

Generating Grasp Poses for a High-DOF Gripper Using Neural Networks

no code implementations1 Mar 2019 Min Liu, Zherong Pan, Kai Xu, Kanishka Ganguly, Dinesh Manocha

The quality of the grasp poses is on par with the groundtruth poses in the dataset.

Robotics

SCORES: Shape Composition with Recursive Substructure Priors

no code implementations14 Sep 2018 Chenyang Zhu, Kai Xu, Siddhartha Chaudhuri, Renjiao Yi, Hao Zhang

The network may significantly alter the geometry and structure of the input parts and synthesize a novel shape structure based on the inputs, while adding or removing parts to minimize a structure plausibility loss.

Learning to Group and Label Fine-Grained Shape Components

no code implementations13 Sep 2018 Xiaogang Wang, Bin Zhou, Haiyue Fang, Xiaowu Chen, Qinping Zhao, Kai Xu

We propose to generate part hypotheses from the components based on a hierarchical grouping strategy, and perform labeling on those part groups instead of directly on the components.

Segmentation

Triangle Lasso for Simultaneous Clustering and Optimization in Graph Datasets

no code implementations20 Aug 2018 Yawei Zhao, Kai Xu, Xinwang Liu, En Zhu, Xinzhong Zhu, Jianping Yin

The reason is that it finds the similar instances according to their features directly, which is usually impacted by the imperfect data, and thus returns sub-optimal results.

Clustering

Learning Discriminative 3D Shape Representations by View Discerning Networks

2 code implementations11 Aug 2018 Biao Leng, Cheng Zhang, Xiaocheng Zhou, Cheng Xu, Kai Xu

In this network, a Score Generation Unit is devised to evaluate the quality of each projected image with score vectors.

3D Shape Recognition 3D Shape Representation +1

GRAINS: Generative Recursive Autoencoders for INdoor Scenes

no code implementations24 Jul 2018 Manyi Li, Akshay Gadi Patil, Kai Xu, Siddhartha Chaudhuri, Owais Khan, Ariel Shamir, Changhe Tu, Baoquan Chen, Daniel Cohen-Or, Hao Zhang

We present a generative neural network which enables us to generate plausible 3D indoor scenes in large quantities and varieties, easily and highly efficiently.

Graphics

LAPRAN: A Scalable Laplacian Pyramid Reconstructive Adversarial Network for Flexible Compressive Sensing Reconstruction

1 code implementation ECCV 2018 Kai Xu, Zhikang Zhang, Fengbo Ren

We propose a scalable Laplacian pyramid reconstructive adversarial network (LAPRAN) that enables high-fidelity, flexible and fast CS images reconstruction.

Compressive Sensing SSIM

Generative Ratio Matching Networks

no code implementations ICLR 2020 Akash Srivastava, Kai Xu, Michael U. Gutmann, Charles Sutton

In this work, we take their insight of using kernels as fixed adversaries further and present a novel method for training deep generative models that does not involve saddlepoint optimization.

Semi-Supervised Co-Analysis of 3D Shape Styles from Projected Lines

no code implementations18 Apr 2018 Fenggen Yu, Yan Zhang, Kai Xu, Ali Mahdavi-Amiri, Hao Zhang

We present a semi-supervised co-analysis method for learning 3D shape styles from projected feature lines, achieving style patch localization with only weak supervision.

Clustering

Im2Struct: Recovering 3D Shape Structure from a Single RGB Image

no code implementations CVPR 2018 Chengjie Niu, Jun Li, Kai Xu

We propose to recover 3D shape structures from single RGB images, where structure refers to shape parts represented by cuboids and part relations encompassing connectivity and symmetry.

Object

Interpreting Deep Classifier by Visual Distillation of Dark Knowledge

no code implementations11 Mar 2018 Kai Xu, Dae Hoon Park, Chang Yi, Charles Sutton

Interpreting black box classifiers, such as deep networks, allows an analyst to validate a classifier before it is deployed in a high-stakes setting.

Dimensionality Reduction Model Compression

GRASS: Generative Recursive Autoencoders for Shape Structures

no code implementations5 May 2017 Jun Li, Kai Xu, Siddhartha Chaudhuri, Ersin Yumer, Hao Zhang, Leonidas Guibas

We introduce a novel neural network architecture for encoding and synthesis of 3D shapes, particularly their structures.

A GPU-Outperforming FPGA Accelerator Architecture for Binary Convolutional Neural Networks

no code implementations20 Feb 2017 Yixing Li, Zichuan Liu, Kai Xu, Hao Yu, Fengbo Ren

For processing static data in large batch sizes, the proposed solution is on a par with a Titan X GPU in terms of throughput while delivering 9. 5x higher energy efficiency.

A Data-Driven Compressive Sensing Framework Tailored For Energy-Efficient Wearable Sensing

no code implementations15 Dec 2016 Kai Xu, Yixing Li, Fengbo Ren

Compressive sensing (CS) is a promising technology for realizing energy-efficient wireless sensors for long-term health monitoring.

Compressive Sensing

CSVideoNet: A Real-time End-to-end Learning Framework for High-frame-rate Video Compressive Sensing

2 code implementations15 Dec 2016 Kai Xu, Fengbo Ren

This paper addresses the real-time encoding-decoding problem for high-frame-rate video compressive sensing (CS).

Compressive Sensing Video Compressive Sensing

Recurrent 3D Attentional Networks for End-to-End Active Object Recognition

no code implementations14 Oct 2016 Min Liu, Yifei Shi, Lintao Zheng, Kai Xu, Hui Huang, Dinesh Manocha

Active vision is inherently attention-driven: The agent actively selects views to attend in order to fast achieve the vision task while improving its internal representation of the scene being observed.

Object Recognition

Cannot find the paper you are looking for? You can Submit a new open access paper.