When things go south: Anomaly Detection in Time-Series Data.

Zapraszamy do zapisów na warsztat, na którym poznasz tajniki machine learning - które przekaże Ci ekspert z firmy PCSS - z departamentu: Machine Learning Engineer at Advanced Visualization and Interaction.

Brief:

The complexity of today's systems requires the introduction of automated monitoring to support humans. The vast amount of unclear metrics from thousands of machines is impossible for us to process. How can we automate this approach? We will discuss the step-by-step data exploration, modeling phase and evaluation of our solutions to help us detect anomalies in time series data.

Expected Outcome:

- Preprocessing and exploration of time-series data

- Defining what is an anomaly

- Annotation of time-series data

- Detecting anomalies in dataset using clustering with DTW, autoencoders (NN, CNN, LSTM-CNN)

About Wojciech:

Wojciech Stefaniak specializes in Time-Series and Medical Imaging analysis putting a great emphasis on explainability aspect of chosen solutions. He also performed post-hoc interpretability analysis of a Reinforcement Learning system in Air Traffic Management system. He passes his knowledge to high-school students and teachers in Intel coaching programmes conducting lectures both in Poland and Europe.

Uwaga: Język warsztatu zostanie dopasowany pod uczestników.

Warsztaty są facylitowane przez Agile Akita ®