ML Conference 2019 - Data to the Rescue! Predicting and Preventing Accidents at Sea

- Dr. Simon Potter

Discover how machine learning and data science can predict and prevent accidents at sea, using innovative models and algorithms to identify risky ships and prevent loss.

Key takeaways

We can model one ship at a time and use deep learning networks to predict accidents.
The why behind accidents can be due to geography, navigational, and time.
We use XGBoost models and TCN networks to predict accidents.
We have a large dataset of 100 million transmissions a day.
We can use clustering algorithms to identify risky ships and create a program to help them change their behavior.
Insurance companies are interested in loss prevention and we can help them with that.
We can gain trust in machine learning models by giving users the ability to interpret the results.
SHAP values can be used to understand the contribution of each feature to the final score.
We can use reinforcement learning algorithms to train models.
We can use machine learning and data science to help predict and prevent accidents at sea.
The severity of an accident is important and we need to consider it when modeling accidents.
We can use transfer learning to improve the accuracy of our models.
We can apply machine learning and data science to solve complex problems in the maritime domain.
Data science tools can help build trust in models and we need to make sure that our models are transparent and explainable.
We can use data science to help detect and prevent accidents at sea.
The use of machine learning and data science in the maritime domain can help prevent accidents and save lives.
We need to consider the context of the data when modeling accidents.
We can use machine learning and data science to help solve the problem of predicting and preventing accidents at sea.

ML Conference 2019 - Data to the Rescue! Predicting and Preventing Accidents at Sea

More talks