К содержимому
learnspaceYOUR NEXT CHAPTER
ПРОСТРАНСТВО ОБУЧЕНИЯ
ГлавнаяКаталог курсовМоё обучениеCoursera

Знания без границ

Учитесь у лучших университетов и компаний мира.

Открыть Coursera
Интеграция
Пространство университета
Моё пространствоСтраница курса
↵
ЯЛичный кабинетСтудент
© 2026 LearnSpaceКаждый день — возможность узнать больше.Помощь
Deep Learning for AI Part 1 · LearnSpace
Назад в каталог
courseraАнализ данных

Deep Learning for AI Part 1

Курс от Northeastern University
Средний≈ 22.8 чАнглийский
О курсеНавыкиПрограммаПреподаватели

О курсе

This is Part 1 of a two-part graduate sequence in deep learning. It establishes the foundations of modern deep learning and the core neural architectures behind today's AI systems. You will build from how neural networks learn—through forward propagation and backpropagation—to convolutional networks for computer vision, recurrent networks for sequence data, and the first generative architectures: variational autoencoders, generative adversarial networks, and Transformers. The course emphasizes both conceptual understanding and hands-on implementation in TensorFlow/Keras and PyTorch. Part 2 continues with advanced generative modeling.

Навыки, которые вы освоите

Deep LearningConvolutional Neural NetworksGenerative Adversarial Networks (GANs)Computer VisionAutoencodersRecurrent Neural Networks (RNNs)Model TrainingArtificial Neural NetworksGenerative Model ArchitecturesImage AnalysisApplied Machine LearningGenerative AIKeras (Neural Network Library)

Программа курса

7 модулей · 108 учебных материалов

01Overview of Neural Networks and Deep Learning20 материалов

Course Introduction

Course IntroductionЧтениеSyllabus - Deep Learning for AI Part 1ЧтениеMeet Your FacultyЧтениеAcademic IntegrityЧтение

Why Deep Learning? Modern AI Applications

Учитесь у экспертов

Xuemin Jin

Преподаватель курса

Deep Learning for AI Part 1
В каталоге вашей программы

Инвестируйте в себя

Новые знания — в удобное для вас время.

Начать на Coursera

Обучение откроется на Coursera
в новой вкладке

Обучение на Coursera

≈ 22.8 ч

7 модулей

Язык: Английский

Часть программы вашего университета
Why Deep Learning? Modern AI ApplicationsВидео
Deep Learning Overview and MotivationЧтение
Real-World Applications Across Vision, Language, and ScienceЧтение

Neural Network Architecture and Core Terminology

Neural NetworksВидеоNeurons, Layers, and the Network StructureЧтениеWeights, Biases, and Learned ParametersЧтениеAssess Your Learning: Why Deep Learning and Neural Network ArchitectureЗадание

Forward Propagation and Loss Functions

The Forward Pass: Computing PredictionsЧтениеLoss Functions for Classification and ReconstructionЧтение

Backpropagation and Gradient-Based Optimization

Backpropagation and the Chain RuleЧтениеOptimization Algorithms: SGD, Momentum, AdaGrad, and AdamЧтениеAssess Your Learning: Forward Propagation and BackpropagationЗадание

TensorFlow, PyTorch, and Tensor Fundamentals

Choosing a Framework: TensorFlow vs. PyTorchЧтениеTensor Fundamentals: Scalars Through 3D+ Tensors and Tensor AttributesЧтение

Discriminative vs. Generative Models: A Course Preview

Discriminative vs. Generative Models: A Course PreviewЧтениеAssess Your Learning: Frameworks, Tensors, and Discriminative vs. Generative ModelsЗадание
02Convolutional Neural Networks13 материалов

Why Convolutional Neural Networks?

Why Convolutional Neural Networks?Чтение

Convolution: Filters, Padding, and Stride

Convolution Mechanics and Filter VisualizationЧтениеPadding Modes and StrideЧтение

Pooling Operations and Feature Maps

Max Pooling and Average PoolingЧтениеStrided Convolution and Feature Map InterpretationЧтениеAssess Your Learning: Why CNNs, Convolution, and PoolingЗадание

Batch Normalization, Dropout, and Activation Functions

Batch Normalization, Dropout, and Activation FunctionsВидеоBatch Normalization and Internal Covariate ShiftЧтениеDropout Ratios and Activation Function ChoicesЧтение

CNN Layer Flow and Architecture Patterns

CNN Layer Flow and Architecture PatternsЧтениеAssess Your Learning: Batch Normalization, Dropout, and CNN Layer FlowЗадание

A Simple CNN Example: MNIST and CIFAR-10

A Simple CNN Example: MNIST and CIFAR-10ЧтениеAssess Your Learning: CNN Worked ExampleЗадание
03Introduction to Computer Vision13 материалов

Core Tasks in Computer Vision

What Is Computer Vision? Goals, Scope, and Task TaxonomyЧтениеImage and Video Data Types and ApplicationsЧтение

Object Detection: R-CNN Family and Two-Stage Detectors

R-CNN and the Region Proposal ApproachЧтениеFast R-CNN, Faster R-CNN, and Two-Stage DetectionЧтениеAssess Your Learning: Computer Vision Tasks and R-CNNЗадание

YOLOv8 Architecture: Backbone, Neck, and Detection Head

The YOLO Concept and Architecture OverviewЧтениеYOLOv8: Backbone, FPN Neck, and Detection HeadЧтениеYOLOv8: Loss Function and Non-Maximum SuppressionЧтение

Training YOLOv8 on Custom Datasets

Data Preparation and Training WalkthroughЧтениеAssess Your Learning: YOLOv8 Architecture and TrainingЗадание

Transfer Learning for Vision Applications

Feature Extraction vs. Fine-Tuning for VisionЧтениеThe Keras Pretrained Model API for VisionЧтениеAssess Your Learning: Transfer Learning for VisionЗадание
04Recurrent Neural Networks15 материалов

Introduction to Recurrent Neural Networks

Why Recurrent Networks? Sequence Modeling ApplicationsЧтениеRNN vs. CNN: Handling Temporal DataЧтение

RNN Architecture: Hidden State and Backprop Through Time

The Hidden State Update and RNN UnrollingЧтениеBackprop Through Time and Vanishing/Exploding GradientsЧтениеAssess Your Learning: Introduction to RNNs and Backprop Through TimeЗадание

LSTM: Gating and Long-Range Memory

LSTM Architecture: Forget, Input, and Output GatesЧтениеThe Cell State and Long-Range Memory in LSTMsЧтение

GRU and Architectural Comparison

GRU Architecture: Reset and Update GatesЧтениеGRU vs. LSTM: Trade-offs and Selection CriteriaЧтениеAssess Your Learning: LSTM and GRUЗадание

Handling Text Data: Tokenization and Word Embeddings

The IMDB Dataset and One-Hot EncodingЧтениеWord Embeddings and Embedding LayersЧтение

RNN Worked Example: LSTM Sentiment Classification

Building the LSTM Model for Sentiment AnalysisЧтениеTraining, Evaluation, and ResultsЧтениеAssess Your Learning: Text Data Handling and LSTM Sentiment ClassificationЗадание
05Variational Autoencoders19 материалов

Introduction to Generative Modeling

Generative vs. Discriminative ModelsЧтениеChallenges in Generative Modeling and a Toy Generative ModelЧтение

Representation Learning and the Generative Landscape

Representation Learning and Probability Theory ReviewЧтениеGenerative Model Taxonomy: VAEs, GANs, Flows, Diffusion, EBMsЧтениеAssess Your Learning: Introduction to Generative Modeling and Representation LearningЗадание

Autoencoders: Architecture and Applications

Autoencoder Motivation and Architecture OverviewЧтениеBuilding the Encoder and DecoderЧтение

Transposed Convolution

Transposed ConvolutionВидеоTransposed Convolution for DecodingЧтениеAssess Your Learning: Autoencoders and Latent Space ExplorationЗадание

Variational Autoencoders: Probabilistic Framework

The Probabilistic Extension: From AE to VAEЧтениеThe Reparameterization TrickЧтение

VAE Loss: Reconstruction Error and KL Divergence

The ELBO Loss FunctionЧтениеKL Divergence and Regularizing the Latent SpaceЧтениеAssess Your Learning: VAE Probabilistic Framework and LossЗадание

VAE vs. Autoencoder: Key Differences

VAE vs. Autoencoder: Key DifferencesЧтение

A VAE Example: Latent Space and Image Morphing

Face Generation and Latent Space ArithmeticЧтениеInterpolating and Morphing in Latent SpaceЧтениеAssess Your Learning: VAE vs. Autoencoder and Worked ExampleЗадание
06Generative Adversarial Networks13 материалов

What Are Generative Adversarial Networks?

The Adversarial Framework: Generator and DiscriminatorЧтениеGAN Types, Applications, and Ethical ConsiderationsЧтение

Deep Convolutional GANs: Architecture and Training

DCGAN Architecture and Design PrinciplesЧтениеDCGAN Training: Fashion MNIST and Lego Bricks ExamplesЧтениеAssess Your Learning: What Are GANs and Deep Convolutional GANsЗадание

GAN Training Tips and Stabilization

GAN Training Instability and Mode CollapseЧтениеStabilization Techniques: Normalization, Learning Rate, Label SmoothingЧтение

Wasserstein GAN and Gradient Penalty (WGAN-GP)

Wasserstein Distance and the WGAN ObjectiveЧтениеGradient Penalty and WGAN-GP Training ResultsЧтениеAssess Your Learning: GAN Training Tips and WGAN-GPЗадание

Conditional GANs and CycleGAN

Conditional GAN Architecture and Class ConditioningЧтениеCycleGAN and Unpaired Domain TranslationЧтениеAssess Your Learning: Conditional GANs and CycleGANЗадание
07Transformers15 материалов

What Is a Transformer?

Why Transformers? Advantages Over RNNs and CNNsЧтениеGPT Overview and Key Transformer ApplicationsЧтение

The Attention Mechanism

The Attention MechanismВидеоWhat Is Attention? The Concept and IntuitionЧтениеAttention ContinuedЧтениеAssess Your Learning: What Is a Transformer and the Attention MechanismЗадание

Self-Attention, Multi-Head Attention, and Scaling

Self-Attention and Network ParametersЧтениеMulti-Head Attention and Parallel Representation LearningЧтение

Transformer Layers: Positional Encoding and Causal Masking

Positional Encoding: Sinusoidal and LearnedЧтениеCausal Masking for Autoregressive GenerationЧтениеAssess Your Learning: Multi-Head Attention and Positional EncodingЗадание

A GPT-Style Application: Wine Review Generation

Building a GPT-Style Language ModelЧтениеTraining, Generating, and Evaluating TextЧтениеAssess Your Learning: GPT-Style ApplicationЗадание

Course Conclusion

Congratulations! Чтение