japanese puns?

Learning a new language is full of discoveries along the way!

As I am spending more time on learning the Japanese language the more different things seem to unlock. One of those things is the apparent fun Japanese companies have with puns/slight writing mismatches.

Like this one – I think (as I can not be 100% sure yet…learning!):

こわはたんの活わだ?or これはたんの活わだ?or こわはたんの汚わだ? or これはたんの汚わだ?

This is an advertisement in a supermarket for a laundry detergent. It is themed to an Anime called “Attack on Titan” – properly because the detergents name is Attack. So when I tried to make sense of the text I first read it wrong, of course.

Let’s look at it step-by-step:

I first started reading the Hiragana portion and make sense of it. There I made my first mistake which is to misread the first second character. For some reason my brain went for (wa) when I should have gone for (re).

Then I typed away further and came to the Kanji. I read a (katsu) when it in fact was a (kitanai).

Given that you’ve typed those into Google Translate you will get very interesting results. I had a good laugh by then:

I am not sure if this is on purpose or not – as I do not yet know if I am just making a mess on this or if this is intentionally done so that, given your level of Japanese reading and attention-spent reading it, you get very different and funny results.

Any Japanese readers that can add some explanations? Am I far off with the thoughts?

I see artificial people

When people think of artificial intelligence, AI, they think of Alexa, Siri, Google Home and self-driving cars.

When an AI dreams of humans it dreams up their faces. No really.

Generative Adversarial Networks (GAN) are a relatively new concept in Machine Learning, introduced for the first time in 2014. Their goal is to synthesize artificial samples, such as images, that are indistinguishable from authentic images. A common example of a GAN application is to generate artificial face images by learning from a dataset of celebrity faces. 

Style-based GANs – Generating and Tuning Realistic Artificial Faces

This is some highly impressive stuff given that the system does produce fairly believable results without lots of distortions. You can see some distortions if you click on the image below and keep refreshing. Evertime it will generate a new face for you…

You definitely will not know this person: