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

We propose a deep convolutional neural network architecture codenamed Inception that achieves the new state of the art for classification and detection in the ImageNet Large-Scale Visual Recognition Challenge 2014 (ILSVRC14). The main hallmark of this architecture is the improved utilization of the computing resources inside the network. By a carefully crafted design, we increased the depth and width of the network while keeping the computational budget constant. To optimize quality, the architectural decisions were based on the Hebbian principle and the intuition of multi-scale processing. One particular incarnation used in our submission for ILSVRC14 is called GoogLeNet, a 22 layers deep network, the quality of which is assessed in the context of classification and detection.

/pdf/going-deeper-with-convolutions-1yobw2o2ds.pdf

Going deeper with convolutions

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

Training Deep Neural Networks is complicated by the fact that the distribution of each layer's inputs changes during training, as the parameters of the previous layers change. This slows down the training by requiring lower learning rates and careful parameter initialization, and makes it notoriously hard to train models with saturating nonlinearities. We refer to this phenomenon as internal covariate shift, and address the problem by normalizing layer inputs. Our method draws its strength from making normalization a part of the model architecture and performing the normalization for each training mini-batch. Batch Normalization allows us to use much higher learning rates and be less careful about initialization, and in some cases eliminates the need for Dropout. Applied to a state-of-the-art image classification model, Batch Normalization achieves the same accuracy with 14 times fewer training steps, and beats the original model by a significant margin. Using an ensemble of batch-normalized networks, we improve upon the best published result on ImageNet classification: reaching 4.82% top-5 test error, exceeding the accuracy of human raters.

/pdf/batch-normalization-accelerating-deep-network-training-by-3dkqye6twt.pdf

Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift

Communications between entities who may share common interests. For entities determined to be sharing common interests (e.g., searching using the same terms or topics, browsing a page, a site or a groups of topically related sites), options for communication among the entities are provided. For example, a chat room may be dynamically created for persons who are currently searching or browsing the same or related information. As another example, a “homepage” may be created for each query and contain various types of information related to the query. A permission module controls which entities may participate, what types of information (and from what sources) an entity can (or desires to) receive, what types of information the entity may (or desires to) share.

System and method for impromptu shared communication spaces

An apparatus is provided for sampling instructions in a processor pipeline of a system. The pipeline has a plurality of processing stages. The apparatus includes a fetch unit for fetching instructions into a first stage of the pipeline. Certain randomly selected instructions are identified, and state information of the system is sampled while a particular selected instruction is in any stage of the pipeline. Software is informed when the particular selected instruction leaves the pipeline so that the software can read any of the sampled state information.

Apparatus for randomly sampling instructions in a processor pipeline

A multi-stage query processing system and method enables multi-stage query scoring, including “snippet” generation, through incremental document reconstruction facilitated by a multi-tiered mapping scheme. At one or more stages of a multi-stage query processing system a set of relevancy scores are used to select a subset of documents for presentation as an ordered list to a user. The set of relevancy scores can be derived in part from one or more sets of relevancy scores determined in prior stages of the multi-stage query processing system. In some embodiments, the multi-stage query processing system is capable of executing one or more passes on a user query, and using information from each pass to expand the user query for use in a subsequent pass to improve the relevancy of documents in the ordered list.

Multi-stage query processing system and method for use with tokenspace repository

We compare several algorithms for identifying mirrored hosts on the World Wide Web. The algorithms operate on the basis of URL strings and linkage data: the type of information about Web pages easily available from Web proxies and crawlers. Identification of mirrored hosts can improve Web-based information retrieval in several ways: first, by identifying mirrored hosts, search engines can avoid storing and returning duplicate documents. Second, several new information retrieval techniques for the Web make inferences based on the explicit links among hypertext documents—mirroring perturbs their graph model and degrades performance. Third, mirroring information can be used to redirect users to alternate mirror sites to compensate for various failures, and can thus improve the performance of Web browsers and proxies. We evaluated four classes of “top-down” algorithms for detecting mirrored host pairs (that is, algorithms that are based on page attributes such as URL, IP address, and hyperlinks between pages, and not on the page content) on a collection of 140 million URLs (on 230,000 hosts) and their associated connectivity information. Our best approach is one which combines five algorithms and achieved a precision of 0.57 for a recall of 0.86 considering 100,000 ranked host pairs.

A Comparison of Techniques to Find Mirrored Hosts on the WWW.

A method is described for identifying pages that are near duplicates in a linked database. In the linked database, pages can have incoming links and outgoing links. Two pages are selected, a first page and a second page. For each selected page, the number of outgoing links is determined. The two pages are marked as near duplicates based on the number of common outgoing links for the two pages.

Jeffrey Dean

Papers

System and method for impromptu shared communication spaces

Apparatus for randomly sampling instructions in a processor pipeline

Multi-stage query processing system and method for use with tokenspace repository

A Comparison of Techniques to Find Mirrored Hosts on the WWW.

Method for identifying near duplicate pages in a hyperlinked database