Neural Networks and Deep Learning
Neural Networks and Deep Learning
Columbia University course ECBM E4040
Zoran Kostic, Ph.D., Dipl. Ing., Professor of Professional Practice, zk2172(at)columbia.edu
Electrical Engineering Department, Data Sciences Institute, Â Columbia University in the City of New York
Course in a nutshell:Â
 Theoretical underpinnings and practical aspects of Neural Networks and Deep Learning. Convolutional and Recurrent Neural Networks. Focus on applications and projects.
Bulletin Description:
Developing features & internal representations of the world, artificial neural networks, classifying handwritten digits with logistics regression, feedforward deep networks, back propagation in multilayer perceptrons, regularization of deep or distributed models, optimization for training deep models, convolutional neural networks, recurrent and recursive neural networks, deep learning in speech and object recognition.
Time:
Fall 2021: To register into the course: (i) students need to get onto the Columbia SSOL waitlist, and (ii) populate this form (using CU UNI email). Those students who satisfy the requirements will be moved from the waitlist into the registered list by early September. The course can be taken by undergraduate and graduate student. See additional course information here.
Spring 2021, Fall 2020, Fall 2019, Fall 2018, Fall 2017, Fall 2016.
Prerequisites:
Required prerequisites: knowledge of linear algebra, probability and statistics, programming. Strongly recommended: machine learning.
Suggested prerequisite courses: (BMEB W4020) or (BMEE E4030) or (ECBM E4090) or (EECS E4750) or (COMS W4771) or similar.
Organization
Lectures:
Presentation of material by instructors and guest lecturers
Homeworks:
Combination of analytical and programming assignments
Projects:
Team-based
Students with complementary backgrounds
Significant design
Reports and presentations to Columbia and NYC community
Best could qualify for publications and/or funding
Industry participation:
Project definition and sponsoring
Weekly presentations
Interaction with students through mentoring
Content
Introduction to neural networks.
Convolutional and recurrent neural networks.
Focuses on the intuitive understanding of deep learning.
Review of underpinning theory - linear algebra, statistics, machine learning.
Analytical study and software design.
Three-four assignments in Python and one DL framework (Tensorflow or PyTorch)
Significant project.
Enables further exploration of key concepts in deep learning.
Typical Syllabus
Introduction to Course E4040
Introduction to Deep Learning (DL)
Introduction to DL computing Resources
Machine Learning Algorithms
t-SNE Data Visualization
Universal approximation theorem - Visual proof
Algebra reviewÂ
Deep Feed-forward Networks
Back Propagation
Optimization
Convolutional Neural Nets (CNNs)
CNN applications
CNN Examples
Regularization
Practical Methodology for Deep Learning
Recurrent Neural Nets (RNNs)
RNN applications
Deel Learning Applications
Autoencoders
Generative Networks GANs, Variational Encoders
Generative Models
Guest Lectures
Deep Learning Trends
Books, Tools and Resources
BOOKS:
2017-2023 software platform:
Google TensorFlow, Google Cloud, Python, Github
2016 software platform:
Amazon AWS cloud tools, Code development on github, bitbucket
Project Areas
Medical
Autonomous cars
Environmental
Smart cities
Physical data analytics
2022 Fall Projects
BinaryConnect: Training Deep Neural Networks with binary weights during propagations
Application of transformer model for the time series forecast
Deep Learning for Symbolic Mathematics
Image Demoireing with Learnable Bandpass Filters
IoU comparison: DIoU, CIoU
Joint Face Detection and Alightment using Multi-task Cascaded CNN
MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications
Optical Coherence Tomography Enabled Classification of the Pulmonary Vein
Registration of Optical Coherence Tomography Volumes Subject to Rotation, Translation and Occlusion
Singing Voice Separation from Monaural Recordings Using Deep Recurrent Neural Networks
Single-Image Depth Perception in the Wild
Spectral Representations for Convolutional Neural Networks
Swin Transformer: Hierarchical Vision Transformer Using Shifted Windows
Swin Transformer: Hierarchical Vision Transformer Using Shifted Windows
Uncertainty Estimation Using a Single Deep Deterministic Neural Network
Using LSTMs to Predict Stock Prices
Vision and Deep Learning-Based Algorithms to Detect and Quantify Cracks on Concrete
Weight Normalization: A Simple Reparameterization to Accelerate Training of Deep Neural Networks
2021 Fall Projects
Composing Music With Recurrent Neural Networks
A New Backbone for Hyperspectral Image Reconstruction
A Recurrent Latent Variable Model for Sequential Data
Custom DenseNet models for Tiny ImageNet Classification
Deep compressive autoencoder for action potential compression in large-scale neural recording
Deep Networks with Stochastic Depth
DeepPainter: Painter Classification Using Deep Convolutional Autoencoders
Fairness without Demographics through Adversarially Reweighted Learning
FishNet: A Versatile Backbone for Image, Region, and Pixel Level Prediction
MoEL: Mixture of Empathetic Listeners
Neural Distance Embeddings for Biological Sequences
Neural Distance Embeddings for Biological Sequences
Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations
QANet: Combining Local Convolution with Global Self-Attention for Reading Comprehension
Residual Attention Network for Image Classification
Social GAN for Human Driving Trajectories Prediction
Spectral Representations for Convolutional Neural Networks
Squeeze-and-Excitation Networks Code avaiable
Towards Accurate Binary Convolutional Neural Network Xiaofan Lin, Cong Zhao, Wei Pan, (2017NIPS)
Video-based human emotion recognition
2021 Spring Projects
A deep learning framework for financial time series using stacked autoencoders and long-short term memory
A Neural Algorithm of Artistic Style
A neural attention model for speech command recognition
BinaryConnect: Training Deep Neural Networks with Binary Weights during Propagations
Composing Music With Recurrent Neural Networks
Conditional Generative Adversarial Net
Deep Double Descent: Where Bigger Models and More Data Hurt
Deep Koalarization: Image Colorization using CNNs and Inception-ResNet-v2
Deep Learning for Price Prediction of Cryptocurrencies
DeepPaint -- Classification of Paintings using Convolutional Autoencoder
Densely Connected Convolutional Networks
Enhancing Detection of Steady-State Visual Evoked Potentials Using Deep Learning
From 2D to 3D: Kidney Tumor Segmentation Challenge
GRUV: Algorithmic Music Generation using Recurrent Neural Networks
Highway Networks
MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications
Neural Networks for Automated Essay Grading
Recurrent Neural Networks to Create Comprehensive 3D Models of Granular Media in YADE
Residual Attention Network for Image Classification
Singing Voice Separation from Monaural Recordings Using Deep Recurrent Neural Networks
U-Net: Convolutional Networks for Biomedical Image Segmentation
Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networks
2020 Fall Projects
Custom DenseNet models for Tiny ImageNet Classification
Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networks
A neural attention model for speech command recognition
CNN-Generated Images Are Surprisingly Easy to Spot... for Now
Composing Music With Recurrent Neural Networks
Custom DenseNet models for Tiny ImageNet Classification
Deep Compression: Compressing Deep Neural Networks with Pruning, Trained Quantization and Huffman Coding
Deep Learning and the Cross-Section of Stock Returns: Neural Networks Combining Price and Fundamental Information
Image Super-Resolution Using Deep Convolutional Networks
Machine learning methods for crop yield prediction and climate change impact assessment in agriculture
MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications
Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations
PoseCNN: A convolutional Neural Network for 6D Object Pose Estimation in Cluttered Scenes
Random Erasing Data Augumentation
Residual Attention Network for Image Classification
Rethinking Model Scaling for Convolutional Neural Networks - EfficientNet
Semi-Supervised Learning with Graph Convolutional Neural Networks
Time series forecasting of petroleum production using deep LSTM recurrent networks
Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networks
2019 Projects
A deep learning framework for financial time series using stacked autoencoders and long-short term memory
Adversarial Autoencoder Assisted Artifact Reduction of Ballistocardiogram in Simultaneous EEG-fMRI Recordings
Adversarial Variational Bayes: Unifying Variational Autoencoders and Generative Adversarial Networks
Automated Gleason Grading of Prostate Cancer Tissue Microarrays via Deep Learning
Deep learning-based feature engineering for stock price movement prediction
Deformable Convolutional Networks:(Expanding the receptive field through Deformable Convolution)
Development and validation a RNN model to teach machines to read and comprehend
Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks
FishNet: A Versatile Backbone for Image, Region, and Pixel Level Prediction
MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications
Multi-Digit Number Recognition from Street View Imagery Using Deep Convolutional Neural Networks
One-Shot Video Object Segmentation for Mobile Vision Applications
Pelee: A Real-Time Object Detection System on Mobile Devices
Residual Attention Network for Image Classification
Short Term Electricity Consumption Forecasting in Residential Building with a Selected Auto-regressive Features&ConvLSTM Neural Network Method
Show and Tell: A Neural Image Caption Generator
SSD: Single Shot MultiBox Detector
Stock market's price movement prediction with LSTM neural networks
Towards Accurate Binary Convolutional Neural Network
Weight Normalization: A Simple Reparameterization to Accelerate Training of Deep Neural Networks
Why Should I Trust You? Explaining the Predictions of Any Classifier
World Models
2018 Projects
A deep learning framework for relationship extraction from articles using long-short term memory and named entity recognition
A Neural Algorithm of Artistic Style
A Neural Representation of Sketch Drawings
Adversarial Variational Bayes: Unifying Variational Autoencoders and Generative Adversarial Networks
Backprop KF: Learning Discriminative Deterministic State Estimators
Deep contextualized word representations
Dynamic Routing Between Capsules
Gesture Recognition
Learned in Translation: Contextualized Word Vectors
Learning a Probabilistic Latent Space of Object Shapes via 3D Generative-Adversarial Modeling
Maximum Classifier Discrepancy for Unsupervised Domain Adaptation
Multi-Digit Number Recognition from Street View Imagery Using Deep Convolutional Neural Networks
Neural Networks for Automated Essay Grading
Parallel Multi-Dimensional LSTM,With Application to Fast Biomedical Volumetric Image Segmentation
PixelGAN Autoencoders
Prevention of catastrophic forgetting in Neural Networks for lifelong learning
Semantic Image Inpainting with Deep Generative Models
Towards Accurate Binary Convolutional Neural Network
Universal Style Transfer via Feature Transforms
Unsupervised Image-to-Image Translation Networks
2016 Projects
Striving for Simplicity: The All Convolutional Net
A Combined Semi-supervised Learning mechanism for Video Data via Deep Learning
A Neural Algorithm of Artistic Style
Adieu features? End-to-end speech emotion recognition using a deep convolutional recurrent network
Colorful Image Colorization
Deep Networks with Stochastic Depth
Highway Networks
Image Super-Resolution Using Deep Convolutional Networks
Learning to Protect Communications with Adversarial Neural Cryptography
Singing Voice Separation from Monaural Recordings Using Deep
Recurrent Neural Networks
Spatial Transformer Networks
Spoken Language Understanding Using Long-Short Term Memory Neural Networks
Striving for Simplicity: The All Convolutional Net
Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networks
Course sponsored by equipment and financial contributions of:
NVidia GPU Education Center, Google Cloud, IBM Bluemix, AWS Educate, Atmel, Broadcom (Wiced platform); Intel (Edison IoT platform), Silicon Labs.