3 ms·

I wonder if in the future we'll have matrix-style "blobs" of knowledge that we can plug into compatible AI systems. That way it has to just be trained once on a

I wonder if in the future we'll have matrix-style "blobs" of knowledge that we can plug into compatible AI systems. That way it has to just be trained once on a strong system, and then other systems can take advantage of the learning by forking the state and importing it. It would definitely speed up training AI, and possibly even enable it on lower powered hardware. Imagine downloading "english teenager slang 2016 v2.0" to your home AI, so it can understand what the hell your kids are saying :)

jknz 10y ago
If the blob of knowledge consists of the weights of some neural network and if this blob is public... Then an attacker could easily perform imperceptible perturbation to the input in order to make the network believe that the yogurt is an Eiffel tower or vice versa. (Can't find the related publications right now but it appeared several times on hn before). So if you don't want the system to be gameable, such public blobs of weights may need to be avoided.
- kmicklas 10y ago
  Likely by that time we will have overcome this problem.
heimatau 10y ago
I think this is too linear. My two cents are that AI, when it cracks one language, many languages will be cracked after that. In other words, exponential growth. With Open Source still being prevalent, I just think the libraries will be shared as it will bring in a new era for humanity. Think about all those languages we don't even have translators for (namely those countries that have only 20 people speaking a specific derivation of a derivation of a .....of a language humanity can translate). There are almost 7,000 languages in the world [1]. Language is a tough challenge but when we find a breakthrough, it's my strong opinion, that we'll bring everyone to the table of communication. [1] http://www.linguisticsociety.org/content/how-many-languages-are-there-world http://www.linguisticsociety.org/content/how-many-languages-...

jknz 10y ago

If the blob of knowledge consists of the weights of some neural network and if this blob is public... Then an attacker could easily perform imperceptible perturbation to the input in order to make the network believe that the yogurt is an Eiffel tower or vice versa. (Can't find the related publications right now but it appeared several times on hn before). So if you don't want the system to be gameable, such public blobs of weights may need to be avoided.

kmicklas 10y ago
Likely by that time we will have overcome this problem.

heimatau 10y ago

I think this is too linear. My two cents are that AI, when it cracks one language, many languages will be cracked after that. In other words, exponential growth. With Open Source still being prevalent, I just think the libraries will be shared as it will bring in a new era for humanity. Think about all those languages we don't even have translators for (namely those countries that have only 20 people speaking a specific derivation of a derivation of a .....of a language humanity can translate). There are almost 7,000 languages in the world [1]. Language is a tough challenge but when we find a breakthrough, it's my strong opinion, that we'll bring everyone to the table of communication. [1] http://www.linguisticsociety.org/content/how-many-languages-are-there-world http://www.linguisticsociety.org/content/how-many-languages-...