DynamicPPL: Stan-like Speed for Dynamic Probabilistic Models

We present the preliminary high-level design and features of DynamicPPL.jl, a modular library providing a lightning-fast infrastructure for probabilistic programming. Besides a computational performance that is often close to or better than Stan, DynamicPPL provides an intuitive DSL that allows the rapid development of complex dynamic probabilistic programs. Being entirely written in Julia, a high-level dynamic programming language for numerical computing, DynamicPPL inherits a rich set of features available through the Julia ecosystem. 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. The main features of DynamicPPL are: 1) a meta-programming based DSL for specifying dynamic models using an intuitive tilde-based notation; 2) a tracing data-structure for tracking RVs in dynamic probabilistic models; 3) a rich contextual dispatch system allowing tailored behaviour during model execution; and 4) a user-friendly syntax for probabilistic queries. Finally, we show in a variety of experiments that DynamicPPL, in combination with Turing.jl, achieves computational performance that is often close to or better than Stan.

Deep Differentiable Grasp Planner for High-DOF Grippers

We present an end-to-end algorithm for training deep neural networks to grasp novel objects. Our algorithm builds all the essential components of a grasping system using a forward-backward automatic differentiation approach, including the forward kinematics of the gripper, the collision between the gripper and the target object, and the metric of grasp poses. In particular, we show that a generalized Q1 grasp metric is defined and differentiable for inexact grasps generated by a neural network, and the derivatives of our generalized Q1 metric can be computed from a sensitivity analysis of the induced optimization problem. We show that the derivatives of the (self-)collision terms can be efficiently computed from a watertight triangle mesh of low-quality. Put together, our algorithm allows the computation of grasp poses for high-DOF grippers in unsupervised mode with no ground truth data or improves the results in supervised mode using a small dataset. Our new learning algorithm significantly simplifies the data preparation for learning-based grasping systems and leads to higher qualities of learned grasps on common 3D shape datasets, achieving a 22% higher success rate on physical hardware and 0.12 higher value of the Q1 grasp quality metric.

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

We present a novel approach to category-level 6D object pose and size estimation. To tackle intra-class shape variation, we learn canonical shape space (CASS), a unified representation for a large variety of instances of a certain object category. In particular, CASS is modeled as the latent space of a deep generative model of canonical 3D shapes with normalized pose and size. We train a variational auto-encoder (VAE) for generating 3D point clouds in the canonical space from an RGBD image. The VAE is trained in a cross-category fashion, exploiting the publicly available large 3D shape repositories. Since the 3D point cloud is generated in normalized pose and size, the encoder of the VAE learns view-factorized RGBD embedding. It maps an RGBD image in arbitrary view into a pose-independent 3D shape representation. Object pose and size are then estimated via contrasting it with a pose-dependent feature of the input RGBD extracted with a separate deep neural networks. We integrate the learning of CASS and pose and size estimation into an end-to-end trainable network, achieving the state-of-the-art pose and size accuracy.

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

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. Then, take the absolute value of each element of the coefficient matrix, and retain the k largest coefficients per column, set the other elements to 0, to get a new coefficient matrix. Finally, FSSC performs spectral clustering over the new coefficient matrix. The experimental results on face clustering and motion segmentation show FSSC algorithm can not only obviously reduce the running time, but also achieve higher accuracy compared with the state-of-the-art representation-based subspace clustering algorithms, which verifies that the FSSC algorithm is efficacious and feasible.

Reinforcement Learning based Visual Navigation with Information-Theoretic Regularization

We present a target-driven navigation approach for improving the cross-target and cross-scene generalization for visual navigation. Our approach incorporates an information-theoretic regularization into a deep reinforcement learning (RL) framework. First, we present a supervised generative model to constrain the intermediate process of the RL policy, which is used to generate a future observation from a current observation and a target. Next, we predict a navigation action by analyzing the difference between the generated future and the current. Our approach takes into account the connection between current observations and targets, and the interrelation between actions and visual transformations. This results in a compact and generalizable navigation model. We perform experiments on the AI2-THOR framework and the Active Vision Dataset (AVD) and show at least 7.8% improvement in navigation success rate and 5.7% in SPL, compared to the supervised baseline, in unexplored environments.

Decoupling Features and Coordinates for Few-shot RGB Relocalization

Cross-scene model adaption is a crucial feature for camera relocalization applied in real scenarios. It is preferable that a pre-learned model can be quickly deployed in a novel scene with as little training as possible. The existing state-of-the-art approaches, however, can hardly support few-shot scene adaption due to the entangling of image feature extraction and 3D coordinate regression, which requires a large-scale of training data. To address this issue, inspired by how humans relocalize, we approach camera relocalization with a decoupled solution where feature extraction, coordinate regression and pose estimation are performed separately. Our key insight is that robust and discriminative image features used for coordinate regression should be learned by removing the distracting factor of camera views, because coordinates in the world reference frame are obviously independent of local views. In particular, we employ a deep neural network to learn view-factorized pixel-wise features using several training scenes. Given a new scene, we train a view-dependent per-pixel 3D coordinate regressor while keeping the feature extractor fixed. Such a decoupled design allows us to adapt the entire model to novel scene and achieve accurate camera pose estimation with only few-shot training samples and two orders of magnitude less training time than the state-of-the-arts.

PQ-NET: A Generative Part Seq2Seq Network for 3D Shapes

We introduce PQ-NET, a deep neural network which represents and generates 3D shapes via sequential part assembly. The input to our network is a 3D shape segmented into parts, where each part is first encoded into a feature representation using a part autoencoder. The core component of PQ-NET is a sequence-to-sequence or Seq2Seq autoencoder which encodes a sequence of part features into a latent vector of fixed size, and the decoder reconstructs the 3D shape, one part at a time, resulting in a sequential assembly. The latent space formed by the Seq2Seq encoder encodes both part structure and fine part geometry. The decoder can be adapted to perform several generative tasks including shape autoencoding, interpolation, novel shape generation, and single-view 3D reconstruction, where the generated shapes are all composed of meaningful parts.

Rescan: Inductive Instance Segmentation for Indoor RGBD Scans

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). We propose an algorithm that analyzes these "rescans" to infer a temporal model of a scene with semantic instance information. Our algorithm operates inductively by using the temporal model resulting from past observations to infer an instance segmentation of a new scan, which is then used to update the temporal model. The model contains object instance associations across time and thus can be used to track individual objects, even though there are only sparse observations. During experiments with a new benchmark for the new task, our algorithm outperforms alternate approaches based on state-of-the-art networks for semantic instance segmentation.

New Formulation of Mixed-Integer Conic Programming for Globally Optimal Grasp Planning

We present a two-level branch-and-bound (BB) algorithm to compute the globally optimal gripper pose that maximizes a grasp metric. Our method can take the gripper's kinematics feasibility into consideration to ensure that a given gripper can reach the set of grasp points without collisions or predict infeasibility with finite-time termination when no pose exists for a given set of grasp points. Our main technical contribution is a novel mixed-integer conic programming (MICP) formulation for the inverse kinematics of the gripper that uses a small number of binary variables and tightened constraints, which can be efficiently solved via a low-level BB algorithm. Our experiments show that globally optimal gripper poses for various target objects can be computed taking 20-180 minutes of computation on a desktop machine and the computed grasp quality, in terms of the Q1 metric, is better than those generated using sampling-based planners.

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

Building change detection is of great significance in high resolution remote sensing applications. Multi-index learning, one of the state-of-the-art building change detection methods, still has drawbacks like incapability to find change types directly and heavy computation consumption of MBI. In this paper, a two-stage building change detection method is proposed to address these problems. In the first stage, a multi-scale filtering building index (MFBI) is calculated to detect building areas in each temporal with fast speed and moderate accuracy. In the second stage, images and the corresponding building maps are partitioned into grids. In each grid, the ratio of building areas in time T2 and time T1 is calculated. Each grid is classified into one of the three change patterns, i.e., significantly increase, significantly decrease and approximately unchanged. Exhaustive experiments indicate that the proposed method can detect building change types directly and outperform the current multi-index learning method.