Data Science Masterclass

Course description


Our Data Science Masterclasses are intended for those working as data scientists or data analysts, and want to learn more about specific topics or particular techniques in the field of data science. These courses are an excellent way to connect with fellow data science practitioners and pick up the latest techniques. Courses typically are a full day at the ASI office, and you will get free access to our in-house SherlockML platform throughout the duration of the course and for 90 days after.

Example topics for our Data Science Masterclass are:

  • Introduction to Advanced Machine Learning
  • Advanced Data Acquisition
  • Fuzzy Matching
  • Recommender and next-best-action systems
  • Data visualisation
  • Tensor Flow
  • Spark & Hadoop
  • Algorithm Selection & Design
  • Performance optimisation
  • Recurrent & Convolutional Neural Networks
  • Deploying predictive models into live environments
  • Adversarial model testing
  • AI Safety and Security

Teachers


At ASI Data Science you receive expert tuition from experienced practitioners. Our courses are developed by a team of experienced Machine Learning practitioners, who hold Doctorate Degrees in Science or Computer Science. The ASI has a wealth of experience across different industries, having delivered over 150 commercial data science projects and provided corporate training for three years. Our masterclasses are taught by our in-house experts in the corresponding topic.

Prerequisites


For all of our Data Science masterclasses we will assume familiarity with Python and its standard libraries. For some courses we will assume a basic level of knowledge on machine learning.

Upcoming Event: Neural Networks Masterclass


Our next upcoming masterclass will be about Neural Networks.This course will take you through everything you need to know to understand neural networks. This course is designed for novices in neural nets, and will take you step by step along the way of building neural nets. We’ll demonstrate how to go from fitting linear regressions, to fitting highly non-linear functions to generic datasets. During this course, you will build your own learning algorithm rather than using pre-packaged libraries. This allows you to develop a deep understanding of optimisation of neural nets, rather than learning how to operate a black box. The workshop consists of a mixture of lectures and practical exercises in Jupyter notebooks. You will get plenty of chance to ask and learn from our experts in the field of deep learning.

Neural nets are used for:

  • Regression analysis and (non-linear) time series prediction

  • Classification problems, including pattern and sequence recognition, novelty detection and sequential decision making

  • Data processing, including filtering and clustering.


And are currently applied in across a wide range of areas such as object recognition, speech recognition, natural language processing, medical diagnosis, automated trading systems, vehicle control, game playing etc.

Learning objectives

At the end of this course, participants will be able to:

  • Set-up a basic neural network from scratch
  • Understand how to train a neural network and optimise its parameters
  • Implement neural networks on a broad set of problems

Course outline

  1. Comparing linear, logistic regression frameworks to neural nets
  2. Neural networks as a framework for fitting data with general nonlinear functions
  3. Working with higher dimensional data
  4. Learning and training neural networks
  5. Advanced topics
  6. Q&A with experts

Any questions?