I am pleased to announce that we have improved our insult detector!
It is still the first generation of the detector, which operates on the principles of deep neural networks, LTSM, embedding, and word2vec. Now with two enhanced datasets – one for Czech and another for English...
It is still the first generation of the detector, which operates on the principles of deep neural networks, LTSM, embedding, and word2vec. We have now enhanced it with two improved datasets – one for Czech and the other for English. Additionally, we have implemented a remarkably fast API hosted on AWS, allowing you to track results online.
We are already working on the second generation, which will hybridly combine the power of Deep Learning with the ability of expert systems to explain! You can also look forward to a mobile application for Android.
P.S.: The neural network surprised us by being able to distinguish between sentences like: “The pig has piglets.” = 9% and “You are a pig.” = 87%! 
Původní zdroj: wordpress
Související články
November 2018
This Week, Google Unveiled Its Latest Technological Toy – Bidirectional Encoder Representations Transformers, or BERT
ReadJanuary 2019
The End of Politicians' Lies? Artificial Intelligence Will Soon Be Able to Verify Their Statements in Real Time
ReadSeptember 2020