Deep learning, a chief driver of the AI revolution, can achieve state-of-the-art accuracy in many cognitive or perceptual tasks such as naming objects in a scene or recognizing optimal paths in an environment.
It involves assembling large data sets, creating a neural network, and training, visualizing, and evaluating different models, using specialized hardware - often requiring unique programming knowledge. These tasks are frequently even more challenging because of the complex theory behind them.
In this seminar, we’ll demonstrate new MATLAB features that simplify these tasks and eliminate the low-level programming. We’ll build and train neural networks that recognize handwriting, categorize foods, classify signals, and control machines.
Potential topics within this talk include the following:
Manage large data sets (images, signals, text, etc.)
Create, analyze, and visualize networks, and gain insight into the black box nature of deep learning models
Automatically label ground truth or generate synthetic data
Build or edit deep learning models with a drag-and-drop interface
Perform classification, regression, and semantic segmentation with images or signals
Leverage pre-trained models (e.g. GoogLeNet and ResNet) for transfer learning
Import models from Keras-TensorFlow, Caffe, and the ONNX Model format
Speed up network training with parallel computing on a cluster
About the Presenter
Abhijit Bhattacharjee is a Senior Application Engineer based in Los Angeles who specializes in the areas of computer vision, audio signal processing, machine learning, and deep learning. He holds a MSEE degree from the University of Southern California and works with clients in all industries, including consumer devices, semiconductors, government, and academic. Prior to MathWorks, Abhijit was a researcher at USC Information Sciences Institute working in programs funded by NASA and DARPA.
Senior Application Engineer, Mathworks