Communications - person-to-person, local, global, mass, analog, digital... we have devised an endless collection of techniques, technologies, devices, languages, protocols... to COMMUNICATE with each other.

Given all this, data science is bound to play a HUGE role in communications.

Earlier, we looked at how ML complements traditional NLP techniques, to understand human language, including speech - this has given rise to Skype Translate, Alexa, phone IVAs, etc.

Now, we'll look at other interesting/uncommon communication-related DS applications.

Various ML approaches (all based on having preclassified data) have been, and are being used, to filter out junk emails.

Hoaxy is a tool that can graph the spread of tweets and fake news stories.

A Twitter user's tweets can be analyzed and style-learned, to produce fake tweets in the same style ["it can build an n-gram language model for any public twitter account and then generate new tweets sounding like that account; it can also generate a word cloud of the words most frequently used by that twitter account"].

WaPo's 'Heliograf' system cranks out news articles, but only just-the-facts-m'am ones.

Wordsmith is another system that also creates automated articles - one potential its creators see, is hyper-focused ('targeted') news streams.

The BIGGEST generator of all, is GPT-3 - a pre-trained language model running as a GAN...

A person's audio can be analyzed, to enable new audio creation. This is a YouTube version.

Face detection, using the Viola-Jones algorithm [an ML-based one]... Here is another writeup.

Just FYI, there are other face detection algorithms (not all are ML-based): Feature Point Detection, Bag-of-Words Models, Histogram-of-Oriented Gradients (HOG), Deformable Parts Models, Exemplar Models...).

'Deepfake' - deep learning to TRANSPLANT a face (using a source video plus a source set of IMAGES, and create a target video with swapped out face) exists, which is WORRISOME. THIS IS THE 'REAL' "FAKE NEWS"!

How exactly does deepfake work? Watch this, and this.

Fake videos... are disturbing, and are a genuine threat to society.