Basics of Deep learning
Download as PPTX, PDF3 likes291 views
The document discusses artificial intelligence, particularly focusing on machine learning and deep learning, which enable machines to mimic human cognitive functions like learning and problem-solving. It outlines various architectures in deep learning, including convolutional and recurrent neural networks, and highlights applications such as image recognition and natural language processing. Additionally, it addresses challenges in training these systems and the importance of large datasets and computational power.
Basics of Deep learning
- 2. Colloquially, the term "artificial intelligence" is applied when a machine mimics "cognitive" functions that humans associate with other human minds, such as "learning" and "problem solving".
- 3. Machine learning is a field of computer science that gives computers the ability to learn without being explicitly programmed.
- 4. Artificial neural networks (ANNs) or Connectionist systems are computing systems inspired by the Biological neural networks that constitute animal brains. Such systems learn (progressively improve performance on) tasks by considering examples, generally without task-specific programming.
- 5. Deep learning is also known as Deep structured learning or Hierarchical learning
- 6. Commerce websites, and it is increasingly present in consumer products such as cameras and smartphones. Machine-learning systems are used to identify objects in images, transcribe speech into text, match news items, posts or products with usersā interests, and select relevant results of search. Increasingly, these applications make use of a class of techniques
- 7. Deep learning is a subset of a more general field of artificial intelligence called machine learning. āneural network with more than two layers.ā More neurons than previous networks More complex ways of connecting layers/neurons in NNs Explosion in the amount of computing power available to train Automatic feature extraction Deep learning as neural networks with a large number of parameters and layers in one of four fundamental network architectures: ļ§Unsupervised pretrained networks ļ§Convolutional neural networks ļ§Recurrent neural networks ļ§Recursive neural networks
- 8. Inspired by Bio model Artificial Intelligence which contains Machine Learning Artificial Neural network
- 12. Deep learning as neural networks with a large number of parameters and layers in one of four fundamental network architectures: ļ§Unsupervised pretrained networks ļ§Convolutional neural networks ļ§Recurrent neural networks ļ§Recursive neural networks āneural network with more than two layers.ā More neurons than previous networks More complex ways of connecting layers/neurons in NNs Explosion in the amount of computing power available to train Automatic feature extraction
- 13. Generative deep architectures, which are intended to characterize the high-order correlation properties of the observed or visible data for pattern analysis or synthesis purposes, and/or characterize the joint statistical distributions of the visible data and their associated classes. In the latter case, the use of Bayes rule can turn this type of architecture into a discriminative one. Discriminative deep architectures, which are intended to directly provide discriminative power for pattern classification, often by characterizing the posterior distributions of classes conditioned on the visible data; and Hybrid deep architectures, where the goal is discrimination but is assisted (often in a significant way) with the outcomes of generative architectures via better optimization or/and regularization, or discriminative criteria are used to learn the parameters in any of the deep generative models in category 1) above. Despite the complex categorization of the deep learning architectures,theoneās that are in practice are deep feed-forward networks, Convolution networks and Recurrent Networks.
- 14. Low level feature Mid level feature High level feature
- 15. Low level Mid level High level Line Curve Bird Color Shape Image Feather Beak Bird
- 16. ļ§Large layer of ANN, ļ§High computational spped, ļ§Large memory size. ļ§Input and output layers ļ§Analog signal
- 17. Machine learning tasks are typically classified into two broad categories, depending on whether there is a learning "signal" or "feedback" available to a learning system: ā¢Supervised learning ā¢Semi-supervised learning ā¢Active learning ā¢Reinforcement learning ā¢Unsupervised learning ā¢Feature learning ā¢Support vector machine ā¢Classification ā¢Clustering ā¢Density estimation ā¢Dimensionality reduction
- 18. Deep learning as neural networks with a large number of parameters and layers in one of four fundamental network architectures: ļ§Unsupervised pretrained networks ļ§Convolutional neural networks ļ§Recurrent neural networks ļ§Recursive neural networks
- 19. ā¢Examine the foundations of machine learning and neural networks ā¢Learn how to train feed-forward neural networks ā¢Use software to implement your first neural network ā¢Manage problems that arise as you begin to make networks deeper ā¢Build neural networks that analyze complex images ā¢Perform effective dimensionality reduction using autoencoders ā¢Dive deep into sequence analysis to examine language ā¢Understand the fundamentals of reinforcement learning
- 20. Three set of neurons: ļ§Receive an input signal ļ§Send an output signal ļ§Processing many layers linear and non-linear transformations..
- 21. ļ§Pylearn2 ļ§Theano ļ§Caffe ļ§Torch ļ§Cuda-convnet ļ§Deeplearning4j ļ§Other Deep Learning Tools ļ§Python ļ§Tensorflow
- 22. ļ§Colorization of Black and White Images ļ§Automatic Speech Recognition ļ§Machine Translations ļ§Object Classification and Detection in Photographs ļ§Automatic Hand writing Generation ļ§Automatic Game playing ļ§Generative Model Chatbots ļ§Natural language processing ļ§Customer relation management ļ§Recommendation system ļ§bioinformatics ļ§Mobile advertising, Autonomous driving driving ļ§visual object recognition, object detection and many other domains such as drug discovery and genomics.
- 23. Common issues are ļ§Over fitting ļ§Learning Rate and ļ§Computation time.
- 24. Deep Learning = Data Science + cloud storage + Soft Computing
- 25. Deep Learning is Large network Large Data Large Model Efficient Algorithm Large Computation Large insights Improved Performance Nearer to Human Brain