11 Comments

Quite insightful and profoundly argumented topic, even if still at its beginnings. I'm eager to read all the episodes of this series!

I also dare to submit to your attention a visualization I published a while ago on La Lettura. Basically I tried to describe an LLM generative process against natural (human) language, namely the language we speak everyday (no matter what language, but, as an Italian speaker, you already know that English translates “lingua” and “linguaggio” as an unique word, and I'm referring to the latter here).

Arguably this work takes on the matter from a slightly different angle than yours. It's a journalistic, divulgative, and simplified visual description, and it relies on my background in General Linguistics much more than on my capacities as a designer.

Let me know if you find anything of interest for your further studies.

https://www.behance.net/gallery/182883483/Chat-GPT-and-natural-language-a-comparison

Expand full comment

Ops, I forgot to write that visualization is in Italian! Currently I haven't translated it to English, and I shall do that as soon as I can. Sorry for this

Expand full comment

I like the fact that it's in Italian!

Expand full comment

Very interesting and little-researched topic in general (perhaps only in academic research). I'm very curious to read the next issues.

Expand full comment

☆☆☆☆☆

Well written and insightful! Looking forward to your next dispatch!

Expand full comment

It's great to hear you are interested! I am also looking forward to publishing more.

Expand full comment

Erico - this is a very important topic for everyone in the data science world, as it is very poorly covered in general. We only see the basic graphs that are used over and over (and which are difficult to interpret for those who are not specialists). I look forward to viewing / reading your series.

Expand full comment

Let's see how this is going to develop! Often these series have a life on their own! :)

Expand full comment

Enrico, this is really very interesting and timely. I come from the ML domain and for several reasons approached DV as a consequence of it. What happened in my research experience is DV is an essential component of the ML pipeline, both as a preliminary tool and as a way to interact with domain experts, who should always be involved in all steps of data analysis process (I am thinking, for instance, of medical applications). Can't wait to read the next contribution.

Expand full comment

We should talk! I have done this kind of research for a few years and now I am looking around to take slightly different directions. I'd be curious to learn more about the way you integrate ML and DV.

Expand full comment

Let’s chat soon over meet!

Expand full comment