Deeper neural networks are more difficult to train. We present a residual learning framework to ease the training of networks that are substantially deeper than those used previously. We explicitly reformulate the layers as learning residual functions with reference to the layer inputs, instead of learning unreferenced functions. We provide comprehensive empirical evidence showing that these residual networks are easier to optimize, and can gain accuracy from considerably increased depth. On the ImageNet dataset we evaluate residual nets with a depth of up to 152 layers—8× deeper than VGG nets [40] but still having lower complexity. An ensemble of these residual nets achieves 3.57% error on the ImageNet test set. This result won the 1st place on the ILSVRC 2015 classification task. We also present analysis on CIFAR-10 with 100 and 1000 layers. The depth of representations is of central importance for many visual recognition tasks. Solely due to our extremely deep representations, we obtain a 28% relative improvement on the COCO object detection dataset. Deep residual nets are foundations of our submissions to ILSVRC & COCO 2015 competitions1, where we also won the 1st places on the tasks of ImageNet detection, ImageNet localization, COCO detection, and COCO segmentation.

/pdf/deep-residual-learning-for-image-recognition-b75jg61qrq.pdf

Deep Residual Learning for Image Recognition

In this work we investigate the effect of the convolutional network depth on its accuracy in the large-scale image recognition setting. Our main contribution is a thorough evaluation of networks of increasing depth using an architecture with very small (3x3) convolution filters, which shows that a significant improvement on the prior-art configurations can be achieved by pushing the depth to 16-19 weight layers. These findings were the basis of our ImageNet Challenge 2014 submission, where our team secured the first and the second places in the localisation and classification tracks respectively. We also show that our representations generalise well to other datasets, where they achieve state-of-the-art results. We have made our two best-performing ConvNet models publicly available to facilitate further research on the use of deep visual representations in computer vision.

Very Deep Convolutional Networks for Large-Scale Image Recognition

Deeper neural networks are more difficult to train. We present a residual learning framework to ease the training of networks that are substantially deeper than those used previously. We explicitly reformulate the layers as learning residual functions with reference to the layer inputs, instead of learning unreferenced functions. We provide comprehensive empirical evidence showing that these residual networks are easier to optimize, and can gain accuracy from considerably increased depth. On the ImageNet dataset we evaluate residual nets with a depth of up to 152 layers---8x deeper than VGG nets but still having lower complexity. An ensemble of these residual nets achieves 3.57% error on the ImageNet test set. This result won the 1st place on the ILSVRC 2015 classification task. We also present analysis on CIFAR-10 with 100 and 1000 layers. 
The depth of representations is of central importance for many visual recognition tasks. Solely due to our extremely deep representations, we obtain a 28% relative improvement on the COCO object detection dataset. Deep residual nets are foundations of our submissions to ILSVRC & COCO 2015 competitions, where we also won the 1st places on the tasks of ImageNet detection, ImageNet localization, COCO detection, and COCO segmentation.

Deep learning is a form of machine learning that enables computers to learn from experience and understand the world in terms of a hierarchy of concepts. Because the computer gathers knowledge from experience, there is no need for a human computer operator to formally specify all the knowledge that the computer needs. The hierarchy of concepts allows the computer to learn complicated concepts by building them out of simpler ones; a graph of these hierarchies would be many layers deep. This book introduces a broad range of topics in deep learning. The text offers mathematical and conceptual background, covering relevant concepts in linear algebra, probability theory and information theory, numerical computation, and machine learning. It describes deep learning techniques used by practitioners in industry, including deep feedforward networks, regularization, optimization algorithms, convolutional networks, sequence modeling, and practical methodology; and it surveys such applications as natural language processing, speech recognition, computer vision, online recommendation systems, bioinformatics, and videogames. Finally, the book offers research perspectives, covering such theoretical topics as linear factor models, autoencoders, representation learning, structured probabilistic models, Monte Carlo methods, the partition function, approximate inference, and deep generative models. Deep Learning can be used by undergraduate or graduate students planning careers in either industry or research, and by software engineers who want to begin using deep learning in their products or platforms. A website offers supplementary material for both readers and instructors.

Deep Learning

Intra-operative medical imaging enables incorporation of human experience and intelligence in a controlled, closed-loop fashion. Magnetic resonance imaging (MRI) is an ideal modality for surgical guidance of diagnostic and therapeutic procedures, with its ability to perform high resolution, real-time, high soft tissue contrast imaging without ionizing radiation. However, for most current image-guided approaches only static pre-operative images are accessible for guidance, which are unable to provide updated information during a surgical procedure. The high magnetic field, electrical interference, and limited access of closed-bore MRI render great challenges to developing robotic systems that can perform inside a diagnostic high-field MRI while obtaining interactively updated MR images. To overcome these limitations, we are developing a piezoelectrically actuated robotic assistant for actuated percutaneous prostate interventions under real-time MRI guidance. Utilizing a modular design, the system enables coherent and straight forward workflow for various percutaneous interventions, including prostate biopsy sampling and brachytherapy seed placement, using various needle driver configurations. The unified workflow compromises: 1) system hardware and software initialization, 2) fiducial frame registration, 3) target selection and motion planning, 4) moving to the target and performing the intervention (e.g. taking a biopsy sample) under live imaging, and 5) visualization and verification. Phantom experiments of prostate biopsy and brachytherapy were executed under MRI-guidance to evaluate the feasibility of the workflow. The robot successfully performed fully actuated biopsy sampling and delivery of simulated brachytherapy seeds under live MR imaging, as well as precise delivery of a prostate brachytherapy seed distribution with an RMS accuracy of 0.98mm.

/pdf/a-fully-actuated-robotic-assistant-for-mri-guided-prostate-3qlpef74u1.pdf

A Fully Actuated Robotic Assistant for MRI-Guided Prostate Biopsy and Brachytherapy.

Nanostructured supramolecular polymers (SPs) are filamentous assemblies possessing a high degree of internal order and have important uses in regenerative medicine, drug delivery, and soft matter electronics. Despite recent advances in functional SPs, a challenging topic is the development of robust assembly protocols enabling the incorporation of various functional units without altering its supramolecular architecture. We report here the robust tubular assembly of camptothecin (CPT) analogues into functional SPs. Covalent linkage of two CPT moieties to various short hydrophilic segments (e.g., nonionic, cationic, anionic, and zwitterionic) leads to a class of CPT analogues that self-assemble in water into tubular SPs. Systemic administration of nonionic SPs effectively suppresses tumor growth. Furthermore, these tubular SPs act as universal dispersing agents in water for low-molecular-weight hydrophobes.

Macrocyclization of a Class of Camptothecin Analogues into Tubular Supramolecular Polymers.

The wide deployment of GPS devices has generated gigantic datasets of pedestrian and vehicular trajectories. These datasets offer great opportunities for enhancing our understanding of human mobility patterns, thus benefiting many applications ranging from location-based services (LBS) to transportation system planning. In this work, we introduce the notion of pathlet for the purpose of compressing and planning trajectories. Given a collection of trajectories on a roadmap as input, we seek to compute a compact dictionary of pathlets so that the number of pathlets that are used to represent each trajectory is minimized. We propose an effective approach whose complexity is linear in the number of trajectories. Experimental results show that our approach is able to extract a compact pathlet dictionary such that all trajectories can be represented by the concatenations of a few pathlets from the dictionary. We demonstrate the usefulness of the learned pathlet dictionary in route planning.

/pdf/pathlet-learning-for-compressing-and-planning-trajectories-2vmuxq7ffq.pdf

Pathlet learning for compressing and planning trajectories

/pdf/towards-scale-invariant-graph-related-problem-solving-by-4e206i8ox4.pdf

Towards Scale-Invariant Graph-related Problem Solving by Iterative Homogeneous GNNs.

3D shape completion for real data is important but challenging, since partial point clouds acquired by real-world sensors are usually sparse, noisy and unaligned Different from previous methods, we address the problem of learning 3D complete shape from unaligned and real-world partial point clouds To this end, we propose a weakly-supervised method to estimate both 3D canonical shape and 6-DoF pose for alignment, given multiple partial observations associated with the same instance The network jointly optimizes canonical shapes and poses with multi-view geometry constraints during training, and can infer the complete shape given a single partial point cloud Moreover, learned pose estimation can facilitate partial point cloud registration Experiments on both synthetic and real data show that it is feasible and promising to learn 3D shape completion through large-scale data without shape and pose supervision

Hao Su

Papers

A Fully Actuated Robotic Assistant for MRI-Guided Prostate Biopsy and Brachytherapy.

Macrocyclization of a Class of Camptothecin Analogues into Tubular Supramolecular Polymers.

Pathlet learning for compressing and planning trajectories

Towards Scale-Invariant Graph-related Problem Solving by Iterative Homogeneous GNNs.

Weakly-supervised 3D Shape Completion in the Wild