Machine Learning is the study of methods for programming computers to learn. Computers are applied to a wide range of tasks, and for most of these it is relatively easy for programmers to design and implement the necessary software. However, there are many tasks for which this is difficult or impossible. These can be divided into four general categories. First, there are problems for which there exist no human experts. For example, in modern automated manufacturing facilities, there is a need to predict machine failures before they occur by analyzing sensor readings. Because the machines are new, there are no human experts who can be interviewed by a programmer to provide the knowledge necessary to build a computer system. A machine learning system can study recorded data and subsequent machine failures and learn prediction rules. Second, there are problems where human experts exist, but where they are unable to explain their expertise. This is the case in many perceptual tasks, such as speech recognition, hand-writing recognition, and natural language understanding. Virtually all humans exhibit expert-level abilities on these tasks, but none of them can describe the detailed steps that they follow as they perform them. Fortunately, humans can provide machines with examples of the inputs and correct outputs for these tasks, so machine learning algorithms can learn to map the inputs to the outputs. Third, there are problems where phenomena are changing rapidly. In finance, for example, people would like to predict the future behavior of the stock market, of consumer purchases, or of exchange rates. These behaviors change frequently, so that even if a programmer could construct a good predictive computer program, it would need to be rewritten frequently. A learning program can relieve the programmer of this burden by constantly modifying and tuning a set of learned prediction rules. Fourth, there are applications that need to be customized for each computer user separately. Consider, for example, a program to filter unwanted electronic mail messages. Different users will need different filters. It is unreasonable to expect each user to program his or her own rules, and it is infeasible to provide every user with a software engineer to keep the rules up-to-date. A machine learning system can learn which mail messages the user rejects and maintain the filtering rules automatically. Machine learning addresses many of the same research questions as the fields of statistics, data mining, and psychology, but with differences of emphasis. Statistics focuses on understanding the phenomena that have generated the data, often with the goal of testing different hypotheses about those phenomena. Data mining seeks to find patterns in the data that are understandable by people. Psychological studies of human learning aspire to understand the mechanisms underlying the various learning behaviors exhibited by people (concept learning, skill acquisition, strategy change, etc.).

Machine learning

A survey on deep learning in medical image analysis

Deep convolutional neural networks have performed remarkably well on many Computer Vision tasks. However, these networks are heavily reliant on big data to avoid overfitting. Overfitting refers to the phenomenon when a network learns a function with very high variance such as to perfectly model the training data. Unfortunately, many application domains do not have access to big data, such as medical image analysis. This survey focuses on Data Augmentation, a data-space solution to the problem of limited data. Data Augmentation encompasses a suite of techniques that enhance the size and quality of training datasets such that better Deep Learning models can be built using them. The image augmentation algorithms discussed in this survey include geometric transformations, color space augmentations, kernel filters, mixing images, random erasing, feature space augmentation, adversarial training, generative adversarial networks, neural style transfer, and meta-learning. The application of augmentation methods based on GANs are heavily covered in this survey. In addition to augmentation techniques, this paper will briefly discuss other characteristics of Data Augmentation such as test-time augmentation, resolution impact, final dataset size, and curriculum learning. This survey will present existing methods for Data Augmentation, promising developments, and meta-level decisions for implementing Data Augmentation. Readers will understand how Data Augmentation can improve the performance of their models and expand limited datasets to take advantage of the capabilities of big data.

/pdf/a-survey-on-image-data-augmentation-for-deep-learning-4ru1bj2hbb.pdf

A survey on Image Data Augmentation for Deep Learning

An intelligent electronic appliance preferably includes a user interface, data input and/or output port, and an intelligent processor. A preferred embodiment comprises a set top box for interacting with broadband media streams, with an adaptive user interface, content-based media processing and/or media metadata processing, and telecommunications integration. An adaptive user interface models the user, by observation, feedback, and/or explicit input, and presents a user interface and/or executes functions based on the user model. A content-based media processing system analyzes media content, for example audio and video, to understand the content, for example to generate content-descriptive metadata. A media metadata processing system operates on locally or remotely generated metadata to process the media in accordance with the metadata, which may be, for example, an electronic program guide, MPEG 7 data, and/or automatically generated format. A set top box preferably includes digital trick play effects, and incorporated digital rights management features.

Intelligent electronic appliance system and method

An adaptive interface for a programmable system, for predicting a desired user function, based on user history, as well as machine internal status and context. The apparatus receives an input from the user and other data. A predicted input is presented for confirmation by the user, and the predictive mechanism is updated based on this feedback. Also provided is a pattern recognition system for a multimedia device, wherein a user input is matched to a video stream on a conceptual basis, allowing inexact programming of a multimedia device. The system analyzes a data stream for correspondence with a data pattern for processing and storage. The data stream is subjected to adaptive pattern recognition to extract features of interest to provide a highly compressed representation which may be efficiently processed to determine correspondence. Applications of the interface and system include a VCR, medical device, vehicle control system, audio device, environmental control system, securities trading terminal, and smart house. The system optionally includes an actuator for effecting the environment of operation, allowing closed-loop feedback operation and automated learning.

Ergonomic man-machine interface incorporating adaptive pattern recognition based control system

Disclosed is a multimedia database for use in distributed network environments. A query processing module transforms user queries into search transformations that can be used for indexing; an attentional selection module records salient information represented in images by a hierarchy of feature maps, saliency maps and combined saliency maps; a declarative memory comprises active modules that update the representations acquired over time for the same images, as well as across images, to cluster, categorize and organize information extracted from different images by the attentional selection module; and an indexing mechanism utilizes the search transformations from the query processing module to search the declarative memory for an answer to a user query. The multimedia database is used in distributed network environments wherein a server can select and coordinate information flow between multiple database sites and user sites, provide effective multimedia representation, create indexes on multimedia content of said database sites to derive a set of target multimedia database sites and, enable database queries descriptions and final search results to be conveyed to a user, records information needed for database selection, generates useful query patterns based on memorization of responses patterns from various databases based on user queries, and directs interactive dialogue between a user and multimedia databases.

Multimedia database for use over networks

/pdf/deep-learning-in-medical-image-analysis-and-multimodal-55sfa0vx66.pdf

Deep Learning in Medical Image Analysis and Multimodal Learning for Clinical Decision Support

There is provided a network server which interfaces a client with selected database sites from a plurality of database sites. The network server comprises a meta-database (including both text information and multimedia information), a search agent, and a refining module. The search agent indexes the meta-database with a user query obtained from the client, and then distributes queries, developed pursuant to such indexing, to the selected ones of the plurality of database sites. In turn, database site information (responsive to the distributed queries) is retrieved from the selected ones of the plurality of database sites. A refining module is used to update the meta-database with the database relevancy information.

System for selecting multimedia databases over networks

With the introduction of fully convolutional neural networks, deep learning has raised the benchmark for medical image segmentation on both speed and accuracy, and different networks have been proposed for 2D and 3D segmentation with promising results. Nevertheless, most networks only handle relatively small numbers of labels (<10), and there are very limited works on handling highly unbalanced object sizes especially in 3D segmentation. In this paper, we propose a network architecture and the corresponding loss function which improve segmentation of very small structures. By combining skip connections and deep supervision with respect to the computational feasibility of 3D segmentation, we propose a fast converging and computationally efficient network architecture for accurate segmentation. Furthermore, inspired by the concept of focal loss, we propose an exponential logarithmic loss which balances the labels not only by their relative sizes but also by their segmentation difficulties. We achieve an average Dice coefficient of 82% on brain segmentation with 20 labels, with the ratio of the smallest to largest object sizes as 0.14%. Less than 100 epochs are required to reach such accuracy, and segmenting a 128x128x128 volume only takes around 0.4 s.

3D Segmentation with Exponential Logarithmic Loss for Highly Unbalanced Object Sizes

In this paper, we propose a novel method to establish temporalcorrespondence between the frames of two videos. 3D epipolargeometry is used to eliminate the distortion generated bythe projection from 3D to 2D. Although the fundamental matrixcontains the extrinsic property of the projective geometrybetween views, it is sensitive to noise. Therefore, wepropose the use of a rank constraint of corresponding pointsin two views to measure the similarity between trajectories.This rank constraint shows more robustness and avoids computationof the fundamental matrix. A dynamic programmingapproach using the similarity measurement is proposed to findthe non-linear time-warping function for videos containinghuman activities. In this way, videos of different individualstaken at different times and from distinct viewpoints canbe synchronized. A temporal pyramid of trajectories is appliedto improve the accuracy of the view-invariant dynamictime-warping approach. We show various applications of thisapproach such as video synthesis, human action recognition,and computer aider training. Compared to state-of-the-arttechniques, our method shows a great improvement.

/pdf/view-invariant-alignment-and-matching-of-video-sequences-1zlqu4nx9p.pdf

Tanveer Syeda-Mahmood

Papers

Multimedia database for use over networks

Deep Learning in Medical Image Analysis and Multimodal Learning for Clinical Decision Support

System for selecting multimedia databases over networks

3D Segmentation with Exponential Logarithmic Loss for Highly Unbalanced Object Sizes

View-invariant Alignment and Matching of Video Sequences