no code implementations • 22 Jan 2024 • Philip Amortila, Tongyi Cao, Akshay Krishnamurthy
A pervasive phenomenon in machine learning applications is distribution shift, where training and deployment conditions for a machine learning model differ.
no code implementations • 14 Nov 2023 • Zhili Chen, Maosheng Ye, Shuangjie Xu, Tongyi Cao, Qifeng Chen
Unlike existing end-to-end autonomous driving frameworks, PPAD models the interactions among ego, agents, and the dynamic environment in an autoregressive manner by interleaving the Prediction and Planning processes at every timestep, instead of a single sequential process of prediction followed by planning.
no code implementations • ICCV 2023 • Xuechao Chen, Shuangjie Xu, Xiaoyi Zou, Tongyi Cao, Dit-yan Yeung, Lu Fang
To take full advantage of the historical frames high-efficiently, we shunt the historical points into two groups with reference to the current points.
no code implementations • 25 Sep 2022 • Rui Wan, Shuangjie Xu, Wei Wu, Xiaoyi Zou, Tongyi Cao
The whole fusion architecture named Dynamic Cross Attention Network (DCAN) exploits multi-level image features and adapts to multiple representations of point clouds, which allows DCA to serve as a plug-in fusion module.
no code implementations • ICCV 2023 • Maosheng Ye, Jiamiao Xu, Xunnong Xu, Tengfei Wang, Tongyi Cao, Qifeng Chen
Also, to model the multi-modality in motion forecasting, we design a novel self-ensembling scheme to obtain accurate teacher targets to enforce the self-constraints with multi-modality supervision.
Ranked #9 on Motion Forecasting on Argoverse CVPR 2020
no code implementations • 16 Jan 2022 • Shuangjie Xu, Rui Wan, Maosheng Ye, Xiaoyi Zou, Tongyi Cao
Two major challenges of 3D LiDAR Panoptic Segmentation (PS) are that point clouds of an object are surface-aggregated and thus hard to model the long-range dependency especially for large instances, and that objects are too close to separate each other.
no code implementations • 16 Nov 2021 • Maosheng Ye, Rui Wan, Shuangjie Xu, Tongyi Cao, Qifeng Chen
The Sparse Feature Encoder extracts the local context information for each point, and the Sparse Geometry Feature Enhancement enhances the geometric properties of a sparse point cloud via multi-scale sparse projection and attentive multi-scale fusion.
no code implementations • ICCV 2021 • Maosheng Ye, Shuangjie Xu, Tongyi Cao, Qifeng Chen
By utilizing these two modules iteratively, features can be propagated between two different representations.
no code implementations • CVPR 2021 • Maosheng Ye, Tongyi Cao, Qifeng Chen
We propose the Temporal Point Cloud Networks (TPCN), a novel and flexible framework with joint spatial and temporal learning for trajectory prediction.
Ranked #52 on Motion Forecasting on Argoverse CVPR 2020
no code implementations • NeurIPS 2020 • Tongyi Cao, Akshay Krishnamurthy
We study reinforcement learning in continuous state and action spaces endowed with a metric.
no code implementations • CVPR 2020 • Maosheng Ye, Shuangjie Xu, Tongyi Cao
We present a Hybrid Voxel network that solves this problem by fusing voxel feature encoder (VFE) of different scales at point-wise level and project into multiple pseudo-image feature maps.
no code implementations • 21 Nov 2017 • Tongyi Cao, Akshay Krishnamurthy
We design new algorithms for the combinatorial pure exploration problem in the multi-arm bandit framework.