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

Binarized Neural Networks (BNNs) eliminate bitwidth redundancy in Convolutional Neural Networks (CNNs) by using a single bit (-1/+1) for network parameters and intermediate representations. This greatly reduces off-chip data transfer and storage overhead. However, considerable computation redundancy remains in BNN inference. To tackle this problem, we investigate the similarity property in input data and kernel weights. We identify an average of 79% input similarity and 61% kernel similarity measured by our proposed metric across common network architectures. Motivated by this observation, we propose SimBNN, a fast and energy-efficient acceleration framework for BNN inference that leverages similarity properties. SimBNN consists of a set of similarity-aware accelerators, a weight reuse optimization algorithm, and a similarity selection mechanism. SimBNN incorporates two types of BNN accelerators, which exploit the input similarity and kernel similarity, respectively. More specifically, the result from the previous stage is reused if similarity is identified, thus significantly reducing BNN computation overhead. Furthermore, we propose a weight reuse optimization algorithm, which increases the weight similarity by off-line re-ordering weight kernels. Finally, our framework provides a systematic method to determine the optimal strategy between input data and kernel weights reuse, based on the similarity characteristics of input data and pre-trained BNNs.

SimBNN: A Similarity-Aware Binarized Neural Network Acceleration Framework

We propose and analyze a kernelized version of Q-learning. Although a kernel space is typically inﬁnite-dimensional, extensive study has shown that generalization is only aﬀected by the eﬀective dimension of the data. We incorporate such ideas into the Q-learning framework and derive regret bounds for arbitrary kernels. In particular, we provide concrete bounds for linear kernels and Gaussian RBF kernels; notably, the latter bound looks almost identical to the former, only that the actual dimension is replaced by a diﬀerent notion of dimensionality. Finally, we test our algorithm on a suite of classic control tasks; remarkably, under the Gaussian RBF kernel, it achieves reasonably good performance after only 1000 environmental steps, while its neural network counterpart, deep Q-learning, still struggles.

Provably Efficient Kernelized Q-Learning

We consider the non-Lambertian object intrinsic problem of recovering diffuse albedo, shading, and specular highlights from a single image of an object. 
We build a large-scale object intrinsics database based on existing 3D models in the ShapeNet database. Rendered with realistic environment maps, millions of synthetic images of objects and their corresponding albedo, shading, and specular ground-truth images are used to train an encoder-decoder CNN. Once trained, the network can decompose an image into the product of albedo and shading components, along with an additive specular component. 
Our CNN delivers accurate and sharp results in this classical inverse problem of computer vision, sharp details attributed to skip layer connections at corresponding resolutions from the encoder to the decoder. Benchmarked on our ShapeNet and MIT intrinsics datasets, our model consistently outperforms the state-of-the-art by a large margin. 
We train and test our CNN on different object categories. Perhaps surprising especially from the CNN classification perspective, our intrinsics CNN generalizes very well across categories. Our analysis shows that feature learning at the encoder stage is more crucial for developing a universal representation across categories. 
We apply our synthetic data trained model to images and videos downloaded from the internet, and observe robust and realistic intrinsics results. Quality non-Lambertian intrinsics could open up many interesting applications such as image-based albedo and specular editing.

Learning Non-Lambertian Object Intrinsics across ShapeNet Categories

Stereotactic neurosurgery involves a technique that can locate the brain targets using an external coordinate system. With the advancements of magnetic resonance imaging (MRI), numerous studies on frameless stereotaxy and MRI-guided/verified technique have been reported to improve the workflow and surgical outcomes. Intraoperative (intraop) MRI guidance in frameless techniques is an appealing method which could simplify workflow by reducing the coregistration errors in different imaging modalities and monitoring the surgical progress. Manually operated platforms thus emerge for MRI-guided frameless procedures. However, these procedures could still be complicated and time-consuming due to their intensive manual operation. To further simplify the procedure and enhance accuracy, robotics has been introduced. In this chapter, we review the state-of-the-art intraop MRI-guided robotic platforms for stereotactic neurosurgery. To improve surgical workflow and achieve greater clinical penetration, three key enabling techniques are discussed with emphasis on their current status, limitations, and future trends.

Prospective Techniques for Magnetic Resonance Imaging–Guided Robot-Assisted Stereotactic Neurosurgery

Accurate placement of tubular shaped surgical tools is necessary for a large variety of image-guided medical interventions. In this process, localization of the instrument, or a robotic assistant manipulating the instrument, is crucial for successful registration of physical space to medical image space. Various fiducial frames and registration methods have been proposed and discussed in literature. However, these frames are typically bulky in size or otherwise not appropriate for use with MR imaging. In particular, it is impossible or awkward to integrate them with the surgical tools. This paper presents the design a compact Cylindrical Helix Imaging Coordinate Registration Fiducial (CHIC) and algorithm to precisely and robustly localize the frame in 6 degree of freedom (DOF). The mathematical model of the frame is developed and evaluated with simulation. This cylindrical shaped frame is particularly suitable for mounting to the distal end of tubular shape surgical tools, which provides a direct imaging visualization of frame, tool and possibly the surgical sites. The paper uses MRI-guided surgical procedure as a focusing application, although the broader aim is development of a versatile registration frame that is usable for a variety of image-guided procedures ranging from ultrasound to fluoroscopy and computerized tomography (CT). Accuracy and performance were evaluated in three cases: simulated images with artificial noise, arbitrarily re-sliced 3D MRI volume, and real 3T MRI images.

Hao Su

Papers

SimBNN: A Similarity-Aware Binarized Neural Network Acceleration Framework

Provably Efficient Kernelized Q-Learning

Learning Non-Lambertian Object Intrinsics across ShapeNet Categories

Prospective Techniques for Magnetic Resonance Imaging–Guided Robot-Assisted Stereotactic Neurosurgery

CHIC: Cylindrical Helix Imaging Coordinate Registration Fiducial for MRI-guided interventions