The musings of a transformer

Earlier last week I posted a poll on twitter asking If my readers would like me to post a GPT generated article. The votes were very evenly distributed:

The remainder of this article is generated using GPT-2 network (using this site) primed on bits of my other articles to covey some of the style. The images were generated by https://app.generative.photos/ from RosebudAI - a recent hot startup in the AI space. When done reading, please consider future historians analyzing the outburst of AI in 2010-2020 and decide if they'd be impressed or will they be like "WTF were they thinking back then!?". 

 


 

The study was done in the summer of 2014, but there have been so many recent news stories about Uber (and similar companies) and the impact it has had on public safety, ”We're very happy” to add to the body of knowledge we've accumulated.


What can we learn about the state of public transportation?

Our findings indicate that if public transportation is to be made safe, “we have to build the systems on a much higher level”, and that this will require substantial change from the traditional public-sector perspective. We've discussed the problems in the above graphic:



In a move to improve safety in space, SpaceX will begin launching small cubesats. This will be a test bed for sending smaller satellites that will help us understand the Earth. It seems this is a new way to get things done that will allow us to do more research and not have to pay for more people.

It seems that the artists have used the videos as a way to present their robots to the public, in a way that makes them seem as if they are living animals.  I think this is part of the point: this is not an "art piece".   It is a very practical way to present something to the public, in a way that is clearly "art" in the sense that it is highly artistic and artistic in the sense that it is very human.  In the video above, there is also a little boy running across the floor.


The problem of vision/computer vision has received a lot of attention since the publication of the seminal paper "Learning to distinguish images of cars from images of other objects" by LeCun et al. (2004).

The first difficulty in solving this problem is that the camera-independent image data is too small. In other words, the image data from a single camera is too noisy and is very different from the image data from a large number of cameras.

A great deal of experimentation continues to be required. One way that can be used to improve the reliability of data mining models is through a process called supervised data generation, in which a data scientist or computer science PhD can produce a model for a given problem using a database, a database of images, and a process that will eventually produce a visual representation of the data. A model can be made to do anything from detecting specific patterns in a large set of data, to learning a set of new images or identifying faces, to producing a map of a given place, or even understanding a given language.


How can we make sense of this new data?

If we consider the way that our brains work, we can think of data as representing information.

For example, we can predict the value of the probability of a coin toss, but we do not know the value of the value of the coin (a value of 1 is not really a number but the probability of winning a coin toss - 1-coin-toss-p) - but we can predict the number of ways that a word "feels" in English (a value of 0 is "not" a number). This gives rise to a certain amount of semantic ambiguity.


So, in the real world, we expect that humans are often wrong. Even if a deep neural network could be trained to understand language, we would expect it to produce gibberish.

And that is precisely what has happened.

If you found an error, highlight it and press Shift + Enter or click here to inform us.

Comments

comments