141–150 of 183 posts
Adversarial Machine Learning. Fake news detection using ML. Integrating 'good old fashioned AI' ideas with modern ML techniques - to some extent Alpha Go went in this direction. While I am glad AI has moved far more towards the machine learning direction, i suspect the decades of AI research that preceded it may come back in a form that is combined with more modern techniques in some way. I see Alpha Go (and Alpha Ze…
Secure Multi-party Computation. The basic idea is developing methods for two (or more) parties with sensitive data to be able to compute some function of their data without having to reveal the data to one another. The classic example is developing an algorithm that allows two people to figure out who is paid more without either revealing what their salary is. Such algorithms get significantly more complicated if the…
Is there actually any way to do that without leaking the salary? If Bob wants to know Alice's salary couldn't he just run the calculation multiple times with different numbers as his salary?
Not so much science in this list, more personal and practical view: As a web developer: WebAssembly As a DevOp: Kubernetes As a backend engineer: headless (CMS) API systems like Strapi or Wagtail
What is the meaning of headless anyway ?
Secure Multi-party Computation. The basic idea is developing methods for two (or more) parties with sensitive data to be able to compute some function of their data without having to reveal the data to one another. The classic example is developing an algorithm that allows two people to figure out who is paid more without either revealing what their salary is. Such algorithms get significantly more complicated if the…
> The classic example is developing an algorithm that allows two people to figure out who is paid more without either revealing what their salary is. Is there actually any way to do that without leaking the salary? If Bob wants to know Alice's salary couldn't he just run the calculation multiple times with different numbers as his salary?
Earlier quoted context omitted.
One field where CS might be put to interesting use is history. Utilizing machine learning to process and analyze historical texts could shine a light on patterns that have gone unnoticed thus far.
I was looking for a MS in CS (looked in EU; as that's where I want to go for my MS) that would've help me with application of CS in liberal arts. I am specifically interested in fields like literature, history, and archaeology. Couldn't find anything. I would want something where it doesn't just enabled me use CS tools and algo to apply to works in those fields (a crude example would be: use of some ML also on Shakes…
DNA computing might be an interesting new domain [1]. The idea is to use DNA as a memory, while using proteins/RNA as logic operators. This can provide massive speed and efficiency gains, especially for optimisation problems that need parallelization. Just consider that 4bits of information on DNA take only about 1nm^3 of volume, where solid state memory has about 3Tb/in^2 which is roughly equivalent to 10^7 nm^3. To…
Secure Multi-party Computation. The basic idea is developing methods for two (or more) parties with sensitive data to be able to compute some function of their data without having to reveal the data to one another. The classic example is developing an algorithm that allows two people to figure out who is paid more without either revealing what their salary is. Such algorithms get significantly more complicated if the…
Almost all of the answers on this list are not fields that are "emerging" but fields that "have already emerged". The ideal emerging field is one that's so obscure we haven't heard of it yet, but so important that we will. If there are widely disseminated books on Amazon about your field, it's not emerging. If there are hundreds of professionals cranking out papers about your field, it's also not emerging. Emerging f…
On the philosophical side, I recently published a paper which could potentially lead to a whole new genre: making actual scientific (=falsifiable) progress on the previously-ineffable question, "Do we live in a simulation?" "A type of simulation which some experimental evidence suggests we don't live in" https://philpapers.org/archive/ALEATO-6.pdf
Nevertheless, it's an interesting observation that we can now easily do experiments that demonstrate correct behavior of logic to the 10^-15 level. If Descartes were looking for evidence of the fallibility of a daemon creating his sense data, it would have been hard to demonstrate better than 10^-3 or 10^-4.
Some problem domains are killer apps, like self driving cars, and on smartphones, personal digital assistants and verbal interfaces. There's no stopping these initiatives. They only question is how far each can go without moon shot levels of investment. But between the economic interests of especially Google and Apple to advance their mobile devices, and the military to make weapons and intel as smart as possible, I'm convinced there's enough critical mass for AI's pile to stay hot for a couple of decades or more.
The trick is to avoid the roadblocks that today's academic agenda inflicts on researchers by demanding they publish frequent shallow incremental novelties. What's needed is 5-10 years to develop the infrastructure that's needed to enable the fielding of robust general reasoning w/ causation and rich knowledgebases.