Deepfakes do not give up and take more strength …
This is a review of some artificial intelligence news that happened throughout the week.
The videos are not real but the threat is …
The Internet already has an excess of information. As if that were not enough, there are people who are trying to influence the thinking of others, either for an election or to spread misinformation, deepfakes are being considered one of the biggest information threats online. You can now edit anyone’s face in a video with tools like machine learning and this is concerning. Deepfakes has become synonymous with misinformation. This time there are organizations that are using deepfakes in a campaign to say that America’s democracy is too fragile.
On the other hand, they also “resurrected” a person who was a victim of a Florida shooting in 2018 to encourage voting. The users did not take it with pleasure and condemned the parents for acting in memory of their son.
Artificial intelligence helping scientists see molecular movements
By applying Natural Language Processing to molecular movements, they create a language that describes the shapes that protein molecules adapt as they change shape. These are defined by their structure and understanding how they control movements helps to understand something like the causes of a disease to the most optimal way to create therapies.
Small alteration in deepfakes makes them invisible to their detection
Deepfakes are becoming easier to generate and more difficult to detect as technology advances. These are the machine learning coronavirus, said Professor Bart Kosko. The most recent deepfake detectors are made based on convolutional neural networks, at first these are precise, they are vulnerable to adverse disturbances, such as small strategic changes of few pixels in an image.
Artificial intelligence helping to create drugs
The Metabolite Traslator tool, created in the Rice laboratory, predicts metabolites that are the product of interactions between small molecules such as drugs and enzymes. They take advantage of deep learning methods and the availability of reaction data that helps them know what a drug will do.
Thanks for reading, I hope you have informed yourself of something new. See you in the next edition.