6 ms·
This seems pretty click-baity to me. Especially "no one knows how it works." Alice generates ciphertext based on plain text and key. Bob generates plaintext
by JackFr 10y ago
This seems pretty click-baity to me. Especially "no one knows how it works."
Alice generates ciphertext based on plain text and key. Bob generates plaintext based on ciphertext and key. Eve generates plaintext based on only ciphertext. Train the three networks across a variety of plaintetxts,
No doubt it's cool, but I would be very surprised if it offered any insight into, or represented an advance for strong cryptography.
- bkin 10y agoJinxed! :-)
- cheez 10y agoThis is less about cryptography and very much about neural networks making stuff.
- pohl 10y agoI might be reading your comment incorrectly, but it sounds like you're interpreting that phrase to mean that no one knows how training the neural network works. I think, instead, it's in reference to the resulting function not having been analyzed. From the article: The researchers didn't perform an exhaustive analysis of the encryption methods devised by Alice and Bob...
- rjbwork 10y ago>I think, instead, it's in reference to the resulting function not having been analyzed. I think it may be that it is "unanalyzable" if you will. Thousands of matrices multiplied together seems as if it would be hard to come to an actual understanding of...
- patrickmn 10y agoThis right here is why machine learning--although very exciting--is an incredibly frustrating endeavor compared to software engineering. It's statistics at its "worst." At the risk of glorifying AI, a good analogy might be that understanding a complex neural network is similar in nature to trying to understand how neurons in a human brain are wired, and why they cause some things to happen (or not happen) with some probability. It's more like the practice of medicine than it is engineering.
- smallnamespace 10y agoWell, the brain is also a very complex network, and we can often understand each other and predict future outputs just fine. The difference though is that humans have evolved to be able to explain themselves to others and to themselves, even to the point of confabulating false motivations. Until we work on making neural nets explain themselves, they'll continue to be black boxes.
- solipsism 10y agoOur reach shouldn't be limited by our brains' capacity to comprehend, should it?
- patrickmn 10y agoThat reminds me of http://multivax.com/last_question.html http://multivax.com/last_question.html
- ars 10y ago> Thousands of matrices multiplied together seems as if it would be hard to come to an actual understanding of... Most of the operations will cancel out. If I multiply a number by a fixed series of numbers, even if that list is millions long, I can just pre-multiply them and have a single operation. The same is likely happening here. Especially if all the operations are algerbraic it would not be hard to simplify and then analyze.
- rjbwork 10y agoYeah that makes sense I suppose. Just turning it all into a single matrix or an order of magnitude of smaller matrices would certainly simplify analysis.
- Houshalter 10y agoThe problem with neural networks is they use nonlinear transformations after every layer to prevent this. Usually something like max(0, x). However it is sometimes possible to train a much smaller neural network to imitate a larger network, or to prune irrelevant nodes and connections from a big network.
- rjbwork 10y agoDo they do it TO prevent it, or does it provide some additional benefit?
- Houshalter 10y agoWell sort of both. Because of the reasons mentioned above, a purely linear neural net wouldn't be able to learn anything more complicated than simple linear functions. Adding nonlinearities vastly increases the space of possible functions it can learn.
- rjbwork 10y agoAh right. Makes perfect sense. Cheers!
- posterboy 10y agoIndeed, a proper link to an official announcement would be better.
- agumonkey 10y agoI was just watching Westworld (1973), first meeting, I quote: "In some cases they've been designed by other computers. We don't know exactly how they work." Timely.
- qubex 10y agoQuite coincidentally I watched that yesterday. Other great quote “I find it hard to believe in a ‘disease of machinery’!”
- agumonkey 10y agoConsequence of the tv show ~reboot I suppose ?
- qubex 10y agoCould have very plausibly been so, but actually I was just watching through a playlist of “Murderous AI”-trope films. (Live in Italy, Westworld not released here yet.)
- agumonkey 10y agoOh, I'd take that playlist if possible :) ps: Westworld 2016 seems very ambitious, and so far well receive by U.S. audience. I wonder if it will accelerate or slow the distribution.
- lesolorzanova 10y agoI completely agree. Also, they may not know but because they haven't made a detailed analysis, it's not like they couldn't know even if they wanted to. So yeah, click-bait.
- ecesena 10y agoThe paper [1] is by Abadi [2], which is a pretty big name in crypto. So I can assume it's something with value, wether the press can or can't convey that is another story. [1] https://arxiv.org/pdf/1610.06918v1.pdf https://arxiv.org/pdf/1610.06918v1.pdf [2] https://en.wikipedia.org/wiki/Mart%C3%ADn_Abadi https://en.wikipedia.org/wiki/Mart%C3%ADn_Abadi
- meganvito 10y agoI see a trend to use arxiv for FLAG publichmonde. Not sure this is another BIG thing.
- typicalmaybe 10y agoPretty big name is a bit of an overstatement. A pretty big name in cryptography would be someone like Goldwasser/Goldreich/Micali/Bellare/Rivest/Bernstein/... and probably about 200 more before we get to Abadi.
- eanzenberg 10y agoAt this point in time, when I hear 'no one knows/understands the process' or some variation relating to neural networks or deep learning, I know their maturity on the subject.
- Houshalter 10y agoThere's nothing clickbait about it. No one knows how the cryptographic algorithm it invented works. It's very, very difficult, perhaps impossible, to reverse engineer neural networks. They also don't claim it's an advance for strong cryptography, it's just interesting experiment with NNs.