Ask HN: What are the most exciting areas of research in CS right now? | Hacker News Reader

Ask HN: What are the most exciting areas of research in CS right now?

47 pointsaharbi17 comments

Explainable AI (XAI) https://www.darpa.mil/program/explainable-artificial-intelli...

This will be a critical component of AI adoption in high risk and/or highly regulated industries like health care, banking, utilities, etc.

I've seen Keenan Crane of CMU give some great applications of differential geometry to computational geometry and manufacturing/fabrication. He has a book on applied discrete differential geometry available here [1] and also a good youtube channel [2] with talks showing applications. A good nontechnical talk giving an overview of the basics and the motivation for the formalism is [3]. A nontechnical talk showing an application to fabrication is [4].

[1] https://www.cs.cmu.edu/~kmcrane/Projects/DGPDEC/

[2] https://www.youtube.com/user/keenancrane/videos

[3] https://youtu.be/Mcal5Cy7r4E?t=408

[4] https://www.youtube.com/watch?v=KiekJeTEQ-U

Ternary computing (fashionable again because of IoTA platform), n-computing, quantum computing (this waiting for hardware breakthrough).

Do you have any links where I could read about these, well not the quantum. But the other 2 I have never heard of.

Sure, you’re welcome. A gentle intro about IoTA’s reasons for ternary computing here: https://iota.stackexchange.com/questions/8/why-does-iota-use... and here: https://medium.com/@iotasuppoter/iota-the-hardware-part-a610... . A more general and formal discussion here (.pdf file): https://res.mdpi.com/cryptography/cryptography-02-00006/arti...

Ternary is fashionable again because of IOTA? As in the cryptocurrency/network?

...since when? To put it bluntly I don't think any mainstream academic community takes IOTA seriously. Its stated justifications for using ternary computing are pretty weak, theoretical benefits notwithstanding.

You’re welcome, mainstream academic communities do a different job, though. I can see the long-term benefits in trying something technologically novel to accommodate a fully interconnected IoT planet but I am not paid by IOTA to evangelize, so, again, we will see what happens.

Sure, but we can agree IOTA hasn't made anything "fashionable" again right? If it's not being taken seriously by the research community and it hasn't otherwise driven material innovation, it's not setting any trends.

There has always been research interest in ternary computing. But I don't know a single serious research institution or group (even if not academic) which concurs with IOTA's approach.

Well, their vision is shared by industrial partners indeed. I presume we are seeing this from different perspectives (mine: industrial ability within some limitations; yours: full theoretical validation) and maybe we need to agree which is which for “exciting areas of research” and “fashionable”?

Can you name another project besides IOTA using Ternary? I don't think it's becoming a trend.

A couple of Italian high school teachers in electronics are building dedicated hardware from Raspberry / Arduino and writing firmware, details here: https://giuseppetalarico.wordpress.com . Still in its infancy and possibly amateurish but sound, they even got sensible feedback last month while exhibiting at the European Maker Faire in Rome, so we will see.

This was asked recently and some really good topics mentioned:

https://news.ycombinator.com/item?id=17696271

http://www.arxiv-sanity.com/toptwtr

Machine Learning is very marketable right now.

very insightful tyvm

algorithmic game theory

Reply on news.ycombinator.com