University of Copenhagen
data science with r
The course is based on RStudio and a collection of modern R packages. The main focus will be on learning to exploit the full potential of these tools, which can serve as an infrastructure for almost any conceivable data analysis in R. Generalized additive models will be treated as non-trivial examples of how to build a predictive regression model in R.
RStudio: An integrated development environment for R, which supports interactive data analysis, building of data analysis pipelines, and R software development
Tidyverse: A framework and collection of R packages centered on the concept of tidy data
Generalized additive models: A flexible but interpretable and easy-to-use prediction model
Visualization: High-quality figures created from structured specifications using the R package ggplot2
Reproducible analysis: Automatic and reproducible reports are written and generated using R Markdown
Interactive communication: Reactive web-applications for interactive presentations of data and analyses written using Shiny
R as a programming language
Organisation of R code
Predictive modeling and model assessment with R
What you will learn
You will learn to build a complete data analysis pipeline in R. This includes learning R programming techniques for:
Data import from multiple sources
Data manipulation and visualization
Automatic and interactive report generation
In addition to the technical programming skills, you will also be introduced to a conceptual framework for data analysis, where all steps involved in data analysis are automatized via a programmatic pipeline.
Big data analysis - tools and methods
This course will bring you to the forefront of the field by introducing you to the newest tools and methods in large-scale data analysis based on cutting-edge research and extensive experience.
What you will learn
After the course, you will:
Be able to set up basic Big Data Analysis from beginning to end: from retrieving and cleaning the data, to establishing the information level, extracting patterns and finding outliers, to curating the necessary data
Be acquainted with a number of advanced tools like: Data cleaning, statistical methods for very large datasets, data stream analysis and finding patterns and outliers in Big Data, collecting data from instruments and devices (e.g. internet of things (IoT)) and hardware systems design for efficient BDA
Throughout the course, we will focus on using a few structured datasets which illustrate a commercial context and which will be used to demonstrate the different steps in Big Data Analysis.
Data cleaning: Detecting and correcting (or removing) corrupt or inaccurate records
Statistical methods: Robust methods for very large datasets and data with very large variance and outliers
Finding patterns and outliers in Big Data: Which methods can be used to identify sparse patterns in very large datasets, and how to identify data that does not follow the overall pattern for a dataset?
Collecting data from instruments and devices: How to collect, store, and analyze data from a multitude of sources (e.g. apparatus, IoT, etc.)
Systems for Big Data Analysis: Common systems for BDA; Hadoop, PyDisco, etc., and hardware systems design for efficient BDA.
Increasingly, it is not only leading players such as Google and Facebook, but also small and mediumsized companies that are successfully applying deep learning techniques to solve commercially relevant problems in a broad variety of areas as diverse as drug design, customer relation management, and mortgage risk estimation. This course will give you detailed insight into deep learning, introducing you to the basics as well as to the latest tools and methods in this emerging field.
What you will learn
By completing the course, you will be able to set up and use basic deep learning techniques. You will learn how to use deep convolutional neural networks and recurrent neural networks for image, text, and time series analysis tasks. You will also become acquainted with advanced tools and become familiar with using appropriate computational resources to train and apply deep learning models.
The course will also teach you the theoretical foundations of deep neural networks, which will provide you with the understanding necessary for adapting and successfully applying deep learning in your own applications.
Deep learning refers to machine learning algorithms that process data in multiple stages, each stage working on a different representation of the data. These representations are learned and enable data to be analyzed at different levels of abstraction.
Thorough introduction to the basics of neural networks including how to train them (e.g. back propagation)
Introduction to convolutional neural networks
Introduction to recurrent neural networks, for example long short-term memory networks (LSTMs) and gated recurrent units (GRUs), for time series modelling and predicting
Training and applying convolutional and recurrent neural networks for text- and image analysis
Utilizing data augmentation and other preprocessing steps to further improve the generalization
Introduction to generative adversarial networks (GANs)
Using modern software tools for deep learning, in particular TensorFlow (used by DeepMind, Google Brain, Ebay, Twitter, Qualcomm, SAP, and many more) as well as Keras
Application examples presented by experts with first-hand experience in applying deep learning in scientific and commercial applications
Exploiting appropriate hardware systems to speed up the computeintensive process of generating complex deep learning models, e.g. via graphics processing units
Artificial Intelligence and legal Disruption
This seminar is divided into three broad sections. Introducing the emerging field of RoboLaw, the aim of the first section is to critically consider the competing views proffered by commentators about both the extent of the challenge and the nature of the necessary responses. Thus, this first section explores the prominent positions taken in the literature and sketches the contours of the existing debates surrounding the use of autonomous weapons systems in armed conflict. This sets the stage for the second section, which delves into the nature of the legal challenges posed by artificial intelligence and robotic technologies. These revolve largely around issues of rights and responsibilities, but this section will also discuss appropriate and necessary regulatory responses drawn from experience in technology regulation more broadly.
The final section grounds these more theoretical discussions to specific case studies where artificial intelligence and robotics may be societally disruptive, allowing us to consider the role the law might play in more concrete scenarios. This last section is left intentionally flexible to accommodate for the interests of the students who are actually enrolled in the course and will be agreed upon in our first meeting, with a reading list to follow shortly after. This last section of the course is geared towards student-driven content and peer-to-peer teaching and learning, and the class is expected to be clustered into topic-centred groups which will lead the particular session for the given topic. This means that you will self-organise into groups according to the topic area of common interest, and will as a group will be responsible for conducting a full seminar on this topic. This includes finding suitable readings, disseminating these to the rest of the class through Absalon, and actually conducting the class. It should be noted that, since it is designed along the peer-to-peer learning and teaching framework, that attention must be paid to mutual respect: continued engagement and participation is therefore mandatory and may affect your final grade.
Digitization has transformed the legal and contractual relationships between undertakings, citizens and governments. As more data is being collected, stored, exchanged and used electronically, business opportunities and concomitant risks and duties arise, with respect to data privacy and security. The digital revolution is changing the concepts of property, liability,
commerce, public space, currency and dispute resolution. At the same time it is creating new ethical concerns and requires specifically designed regulations and remedies. Althogh its consequences are far reaching – there is still no holistic approach to the subject matter.
This course aims to fill in that gap and to provide students with the foundation for the understanding of the legal implications of the multiple aspects of digitalization and automatization. The course is case law and research based.
Among others, the course deals with topics such as:
digital privacy and security,
digital property and intellectual property,
digital public space,
digital remedies and
Being an introductory course, the course does not require prior knowledge of the topic.