Non-blocking P2P communication Buffers MPI distinguishes different types of buffers: variables user-level buffers hardware/system buffers MPI implementations are excellent in tu...

Neural network training Feedforward step Training the neural network A neural network is trained through back-propagation. It allows to adapt the weights in the layer successively. Moving fro...

Variance Bias Trade-off What is Bias and Variance? Bias: The error introduced by approximating a real-world problem, which may be complex, by a much simpler model. High bias can cause an algo...

ROS Learning Objectives What is ROS? Philosophy Features ROS Wiki Structure Robotic Operating System -> Open Source Set of Libraries Let us Develop and Manage A Modular Framework ...

https://miscada-ml-2425.notes.dmaitre.phyip3.dur.ac.uk/assets/presentations/lecture-5/lecture-5.slides.html#/33 Neural networks Neutral networks started as an attempt to model the human brain. It...

MPI—basics and blocking P2P communication MPI 是进程 (process) 级并行，process 间内存不共享，与线程 (thread) 级并行不同 https://mpitutorial.com/tutorials/ https://scc.ustc.edu.cn/_upload/article/files/e0/98/a9f0c...

Lecture: Learning Objectives The aim of this lecture is to design path planning for robot navigation. Objectives: Obstacle avoidance Search algo...

Regularisation We can prevent a too general model from overfitting through regularisation. We modify the loss function to include a penalty for too high parmeter: \[J_{pen}(X,y.\vec w) = CJ(X,y,...

Not all separable problems are linearly separable. A straight line in not the best decision boundary in many cases. Let’s look at a separable problem that can’t be linearly separated: If we try...

Learning Curves Using polynomial features for this dataset we achieved a good separation, but we wondered how well the model might generalise. Testing the model The data was generated accordi...