MIT •

Donald Trump is one of the most controversial presidential candidates.

(Mirror Daily, United States) – Politics is rough and though, especially when the elections are just around the corner. And there are a lot of people out there who take more pleasure in debating the debates or making fun of some candidates. One Twitter account, in particular, has attracted a lot of attention lately, because it does terrific impersonations of Donald Trump, and the mind behind them is an artificial neural network.

Plenty of people are talking about the man and the ideas behind @DeepDrumpf, and that is also what we are going to do next.

Bradley Hayes is the man that managed to turn the Republican candidate’s speeches into hilarious tweets. And even though the texts are written by artificial intelligence, they sometimes sound exactly like something Trump would say.

The MIT postdoctoral associate who is into robotics research says that he reprogrammed an artificial neural network so that it was able to both study and emulate Trump’s speeches.

Momentarily all of the source material comes from debates, so the robot behind the very popular Twitter account has a bit of a hard time recognizing a speech pattern. This is mostly due to the fact that during the debates, the candidates often tend to interrupt each other, leaving sentences in mid-air.

Hayes said that the idea behind @DeepDrumpf came to him one day while he was having lunch with his colleagues. Scientists can’t just stop speaking science, not even when they’re enjoying a sandwich together. And in one of these instances Hayes and his fellow robotics researchers were talking about different techniques of statistical modeling.

After a while, he read an article about the complexity of the speeches of various political candidates. The piece was talking about the simplistic nature of Donald Trump’s argumentations. The simpler the Republican candidate speaks, the more fans he gathers because people need a representative that communicates in their language.

This was when the idea behind @DeepDrumpf completely formed and Hayes decided to reprogram the AI. What is astonishing is that the robot, which functions extremely well sometimes, is nothing but a weekend project for the MIT postdoctoral associate.

The algorithm behind the tweets is relatively complicated, and the AI still has a lot to learn about the person it is trying to mimic, so sometimes Hayes helps the robot create posts.

He gives the program an incentive, for example, [Hillary is], and then the artificial intelligence generates a series of characters. Hayes then chooses the ones that make the more sense and puts them in a tweet.

A lot of people are talking about the man and the ideas behind @DeepDrumpf. Some members of the social media platform expressed the fact that they would rather vote for Drumpf instead of Trump.

Image source: www.wikimedia.org

"MIT researchers filmed people sneezing and then played the videos on slow motion to study the fluid dispersion."

Researchers filmed sneezes and then played the videos in slow-mo to study the fluid dispersion.

(Mirror Daily, United States) – According to the latest MIT study, sneezing is nastier than you thought. Researchers from the renowned university filmed more than 100 sneezing people and then watched the videos on slow motion. The results? Sneezing is nastier than you thought.

A team of MIT researchers filmed more than 100 sneezing people with high-quality cameras and then replayed the videos in slow-motion. The results of their research were that sneezing is actually way grosser than people thought it is.

Popular belief is that sneezing produces a spray of tiny droplets of mucus, something similar to a spraying a can. But the researchers from MIT busted that myth. It seems that when a person sneezes, the matter that is expelled is in the form of fluid fragments. The image is similar to how paint looks when you throw it from your brush and towards a wall.

The MIT’s head of Disease Transmission and Fluid Dynamics Laboratory, assistant professor Lydia Bourouiba, was the lead researcher of the study. She says that the purpose of the research was to determine how big the droplets of mucus emitted by a sneezing person are.

According to Bourouiba, the first study used healthy patients and determined that the emissions were rather similar for everybody.

It seems that the sneezes vary slightly from person to person and the transformations that affect the fluid after it exists the person’s mouth are dependent on the elasticity of the individual’s saliva. Thus, a normal sneeze ejects the material that firstly takes the shape of small filaments than morph into droplets.

The saliva droplets can either remain suspended in the surrounding environment or they fall to the floor. If the sneezing individual has a more elastic saliva, the sneeze will remain longer in filament form.

Bourouiba and her team documented the dispersion patterns of different sneezes in order to better understand how certain diseases spread. According to a person’s saliva elasticity, or fluid intake, that certain individual could be categorized as a “super spreader”.

The MIT team of researchers will continue with their research, but they will shift their focus on infected individuals. This way they will be able to see if there are any differences between the sneezes.

The images from the study clearly show that sneezing is nastier than you thought, but the brave, iron-stomach people from MIT will use their researcher to better assess the spread of an infectious disease.

Bourouiba said that after watching so many videos and images, she only focuses on the analyses of the fluid’s mechanics and she simply forgets that she is analyzing mucus and saliva.

Image source: www.wikimedia.org

Talking About the Man and the Ideas behind @DeepDrumpf

Sneezing Is Nastier Than You Thought

Categories