When we are working with relational databases, it is usual for some queries to be extremely slow. This can be due to different reasons, but the most common is that the data access pattern that the query demands is not…
Process Discovery is a family of techniques used to obtain information from event logs and reconstruct the underlying model they have generated. This model can then be used for different tasks such as identifying bottlenecks, monitoring deviations from an ideal…
In the first part of the CUDA Tutorial, we mentioned the GPU’s smallest unit of execution, warps, the equivalent of a kernel on a CPU. However, warps alone cannot harness the full power of the GPU; you need to make…
CUDA is the language used for programming on NVIDIA GPUs, it is vital for a large number of computing tasks, and yet it is a mystery to a large number of programmers. In this post, we attempt to explain some…
Introduction This article assumes that there is a basic knowledge about embeddings of objects, either text or images. In case you don’t have any notions on the subject, the post on Text Embeddings: the basis of modern NLP explains this…
Introduction All graphics in this article are interactive due to the nature of trying to represent text graphically. To see which sentence each point represents, simply hover the mouse over it. The results obtained have been generated on the “STS…
The kernel trick is a typical method in machine learning to transform data from an original space to an arbitrary Hilbert space, usually of higher dimensions, where they are more easily separable (ideally, linearly separable). This technique is the basis…
Nowadays, when talking about neural network models, it is common to mention how many graphics cards they need. For larger ones, there is talk of how many cards are needed, and how many GB of RAM may be required for…