Deep Learning School "Basic Concepts"

Aug 8, 2022, 6:00 PM → Aug 12, 2022, 2:00 PM Europe/Berlin

Haus Nordhelle

Martin Erdmann (RWTH Aachen University)

The Deep Learning School "Basic Concepts" is hosted by the community organization DIG-UM with support from the BMBF-funded ErUM-Data-Hub. The event serves the professional education of young scientists belonging to the ErUM Community.

In August 2022, we will host the first Deep Learning School at Landhaus Nordhelle in Meinerzhagen. In a comfortable ambience, intensive classes on Neural Network Building Blocks, Mastering Model Building, Convolutional Neural Networks, Recurrent Neural Networks and Introspection of Neural Networks will be held over 4 days. The workshop is aimed at deep-learning starters from all ErUM communities (Research on Universe and Matter) who have a basic knowledge of physics. 

A fee of 300€ will be charged for participation in the school. The workshop fee includes the cost of the workshop, accommodation and basic catering.

The workshop is full as the maximum number of 59 possible participants has been reached.

Please take note of the current Corona measures linked here.

 

 

Bulletin-1-20220805.pdf

Deep Learning School "Basic Concepts".pdf

ListOfParticipants-20220805.pdf

Monday, August 8

6:00 PM → 7:00 PM Supper

7:30 PM → 9:00 PM Neural Network Building Blocks

7:30 PM Introduction to the Summer School

Speaker: Martin Erdmann (RWTH Aachen University)

Introduction+Deep+Learning+Basic+Concepts+v5.pdf

7:40 PM Neural Network Building Blocks: Lecture

affine transformation, activation, loss function, mean-square-error, backpropagation, Universal Approximation Theorem, ...

Speaker: Dennis Noll (RWTH Aachen University (DE))

220808_noll-deep_learning_1.pdf

Tuesday, August 9

9:00 AM → 10:30 AM Neural Network Building Blocks

9:00 AM Neural Network Building Blocks: Libraries

Introduction to deep learning software libraries: how to build and train a neural network

Speaker: Dennis Noll (RWTH Aachen University (DE))

220809_noll-deep_learning_2.pdf

11:00 AM → 12:30 PM Neural Network Building Blocks

11:00 AM Neural Network Building Blocks: Hands-on

Hands-on: constructing, training, evaluating neural networks

Speaker: Dennis Noll (RWTH Aachen University (DE))

220809_noll-deep_learning_3.pdf

2:30 PM → 4:00 PM Mastering Model Building

2:30 PM Mastering Model Building:Lecture

Input preprocessing, cross entropy, penalty term, stochastic gradient decent, learning strategies, batch normalization, residual learning,...)

Speaker: Marcel Rieger (University of Hamburg)

Hands-on 1: gradients

Hands-on 2: Keras API

mastering_model_building.pdf

Solution 1: gradients

Solution 2: Keras API

4:30 PM → 6:00 PM Mastering Model Building

4:30 PM Mastering Model Building: Hands-on (1)

Speaker: Marcel Rieger (University of Hamburg)

Hands-on 3: TensorFlow Playground

Hands-on 4: Particle mass regression

Slides in first block

Solution 4: Particle mass regression

Wednesday, August 10

9:00 AM → 10:30 AM Mastering Model Building

9:00 AM Mastering Model Building: Hands-on (2)

Speaker: Marcel Rieger (University of Hamburg)

Slides in first block

Solution: Jet flavor identification

11:00 AM → 12:30 PM Convolutional Neural Networks

11:00 AM Convolutional Neural Networks: Lecture

Convolutional layers, filters, parameter sharing, translational invariance, padding, pooling, striding, advanced concepts: inception, xception, unet, ...)

Speaker: Judith Reindl (UniBW Munich)

Convolutional Neural Networks_Reindl_2022_v2.pdf

2:30 PM → 4:00 PM Convolutional Neural Networks

2:30 PM Convolutional Neural Networks: Hands-on (1)

Speaker: Judith Reindl (UniBW Munich)

4:30 PM → 6:00 PM Convolutional Neural Networks

4:30 PM Convolutional Neural Networks: Hands-on (2)

Speaker: Judith Reindl (UniBW Munich)

Thursday, August 11

9:00 AM → 10:30 AM Recurrent Neural Networks

9:00 AM Recurrent Neural Networks : Lecture

Sequential relations, cell architectures, (bi-)directional analysis, long-short-term-memory, gated-recurrent-unit,...

Speaker: Nikolai Hartmann (LMU Munich)

nikolai_rnn_deep_learning_school-18.08.2022_final.pdf

11:00 AM → 12:30 PM Recurrent Neural Networks

11:00 AM Recurrent Neural Networks: Hands-on (1)

Speaker: Nikolai Hartmann (LMU Munich)

Hands-on Link

2:30 PM → 4:00 PM Recurrent Neural Networks

2:30 PM Recurrent Neural Networks: Hands-on (2)

Speaker: Nikolai Hartmann (LMU Munich)

4:30 PM → 6:00 PM Introspection of Neural Networks

4:30 PM Introspection of Neural Networks: Lecture & Hands-on (1)

Activation maximization, saliency maps, deconvolutional network, transposed convolutions, attribution-based predictions, discriminative localization, layer-wise-relevance-propagation...)

Speaker: Caroline Heneka (University of Hamburg)

DL-basics-interpretation-Heneka.pdf

Friday, August 12

9:00 AM → 10:30 AM Introspection of Neural Networks

9:00 AM Introspection of Neural Networks: Lecture & Hands-on (2)

Speaker: Caroline Heneka (University of Hamburg)

11:00 AM → 12:00 PM Introspection of Neural Networks

11:00 AM Introspection of Neural Networks: Lecture & Hands-on (3)

Speaker: Caroline Heneka (University of Hamburg)

11:50 AM Farewell

Speaker: Martin Erdmann (RWTH Aachen University)

12:00 PM → 1:00 PM Lunch