0:00:16 | i |
---|
0:00:16 | so my name is patrick bob um from france of a little uh uh and also france |
---|
0:00:22 | and so this talk is gonna be a |
---|
0:00:25 | uh what i'm knocking security a mostly out to design secure well uh and baiting scheme |
---|
0:00:31 | and then will show you how to use a nice mess medic zero record optimal to once per in order |
---|
0:00:36 | to minimize the distortion and and also to guarantee what um |
---|
0:00:41 | security |
---|
0:00:43 | um |
---|
0:00:44 | so the line is very simple a first introduction on the |
---|
0:00:48 | and what are mocking security is and uh i will uh |
---|
0:00:52 | propose requirement in order to achieve secure and baiting |
---|
0:00:56 | uh i would propose so different embedding scheme and compare them using a performance and that is this |
---|
0:01:02 | and then will conclude my work |
---|
0:01:04 | after the what |
---|
0:01:06 | okay uh what's mean security in what are marking so a security as thing to do is do you picked |
---|
0:01:13 | compression uh no what at to share additional like |
---|
0:01:16 | go and uh |
---|
0:01:18 | channel |
---|
0:01:19 | eight |
---|
0:01:20 | uh you have to consider a first and that this ad that's every |
---|
0:01:23 | it's is very important because the are just every as the brain |
---|
0:01:26 | is able to |
---|
0:01:28 | try to hack you'll system to perform at that |
---|
0:01:31 | it can i have a a a really do a an immediate problem in a uh a a where E |
---|
0:01:36 | it is gonna depend of your uh a now you that you're gonna can see there |
---|
0:01:40 | but uh this is some ones that would try so two |
---|
0:01:45 | hard your system |
---|
0:01:47 | that that's sorry you have to consider that yeah as a bunch of my tell you or |
---|
0:01:52 | than that that you can be for example the what tell the code they'll detect detector then you would be |
---|
0:01:56 | able to yeah form more what we code or equal attack |
---|
0:02:00 | um |
---|
0:02:01 | here we go that can see there's that's the |
---|
0:02:03 | at necessary will |
---|
0:02:05 | have i con turns that that what they mapped to read the |
---|
0:02:08 | and a a what a map with the same key |
---|
0:02:11 | you can have all of L scenario where you assume that the just every for example knows the ms that |
---|
0:02:17 | are embedded in so but a map content T pretty depends of the signal you |
---|
0:02:21 | and a strong assumption is that uh you assume that to what am is a um uh |
---|
0:02:26 | that that sorry know everything about |
---|
0:02:29 | the |
---|
0:02:30 | but they marking scheme except the secret key |
---|
0:02:33 | this this is also one the or so as assumption |
---|
0:02:36 | and what are than the is objective of this uh the adversary re |
---|
0:02:40 | in all cases than a be to estimate you secret key and uh if it's it's to make what they're |
---|
0:02:47 | are all copies of sage |
---|
0:02:48 | okay |
---|
0:02:49 | and that got that don't actually a lot about what is going to a a a a what the is |
---|
0:02:53 | that the stories is uh doing to do because it in all case we can that |
---|
0:02:57 | the uh uh design secure embedding scheme |
---|
0:03:01 | so uh |
---|
0:03:03 | here are the assumption that we are are using |
---|
0:03:06 | we can see they are that the almost is good shown |
---|
0:03:09 | uh is C is this makes sense to sends to the some limit though is your M of course |
---|
0:03:15 | um and there's all assumption is the fact that uh as a horse is i I D in the |
---|
0:03:20 | in there |
---|
0:03:22 | i and in |
---|
0:03:23 | plea uh distributed |
---|
0:03:25 | we assume that all the content uh what them up using the same secret key |
---|
0:03:30 | and is this means that there is a and that that board W A a that map content only at |
---|
0:03:36 | that so we have in this frame |
---|
0:03:39 | the um |
---|
0:03:39 | i just necessary |
---|
0:03:41 | uh as only what am i can turn to does not know about them but didn't nice H for example |
---|
0:03:49 | or was the or here out to design uh secure scheme |
---|
0:03:53 | uh and also sorts all to play with a to was also a very strong can trends and what that |
---|
0:03:58 | masking meanings the |
---|
0:04:00 | and weighting distortions that has to be minimised |
---|
0:04:03 | and also to maximise the robustness |
---|
0:04:06 | okay |
---|
0:04:09 | i to minimize the or uh let's and the robustness so yeah we use um |
---|
0:04:15 | the rationale already use since a long time in what they're marking |
---|
0:04:19 | it's to use in it coding and you set side information from the almost |
---|
0:04:23 | is come from the famous close that's paper but one dove T paper or uh uh |
---|
0:04:28 | a what that marking |
---|
0:04:29 | and uh the I Ds and an I D to generate a different coding rage and that we'll code for |
---|
0:04:35 | the same may say |
---|
0:04:37 | and |
---|
0:04:38 | is an baiting in very simple in this case since you have different and dating regions in a to embed |
---|
0:04:43 | them a said you go to the close |
---|
0:04:45 | decoding regions and you |
---|
0:04:47 | uh uh you're able to minimize the distortion okay |
---|
0:04:51 | so no i will a some up the stream main we got crime i'm and fall secure them baiting |
---|
0:04:57 | so the first one that i called it distribution splitting |
---|
0:05:02 | you are have uh |
---|
0:05:03 | you want to achieve a |
---|
0:05:05 | a very strong |
---|
0:05:07 | statistical property |
---|
0:05:09 | chord is a perfect secrecy uh in steak and the refugee of take was security in what they marking |
---|
0:05:15 | it's means that the distribution of the content |
---|
0:05:18 | uh is the original content so P of X and the distribution of the what um not content i it |
---|
0:05:24 | E E Y given the key a exactly the same |
---|
0:05:28 | if you are of these properties and so just a re can not do anything about a uh i can |
---|
0:05:34 | that that that you'll system then because it doesn't know is a coefficient are what all marked or even |
---|
0:05:39 | uh is the carrier may age |
---|
0:05:42 | okay |
---|
0:05:44 | uh |
---|
0:05:44 | here we assume that we are doing a binary and baiting in means that actually you can't really |
---|
0:05:50 | you're your uh distribution of what that map the content |
---|
0:05:54 | P of Y into two distribution |
---|
0:05:57 | the one uh for of the ms a a zero and the one for the said one |
---|
0:06:02 | and of course you have to divide |
---|
0:06:04 | a two in order to normalize the distribution |
---|
0:06:07 | uh in a bill to achieve this predicting we go now use a partitioning function uh J G |
---|
0:06:15 | and uh you will see i would show you an example but basically once you have G A a a |
---|
0:06:20 | uh G |
---|
0:06:21 | you can compute the |
---|
0:06:24 | distribution and fall zero according to this partitioning function then you can compute the D |
---|
0:06:30 | be shown the what are not good content and bidding one |
---|
0:06:33 | also according |
---|
0:06:34 | to this partitioning function |
---|
0:06:37 | you're is an example of the J function so it's |
---|
0:06:39 | this |
---|
0:06:40 | lou |
---|
0:06:41 | uh |
---|
0:06:43 | that a wise uh function |
---|
0:06:46 | and so you see that the and function your is divided so it |
---|
0:06:50 | in in two parts |
---|
0:06:52 | here i just |
---|
0:06:54 | are you |
---|
0:06:55 | the part for one dating zero but you have also as you a part one bit one so the blue |
---|
0:07:00 | K here |
---|
0:07:01 | is the distribution |
---|
0:07:02 | one you want but you want but you |
---|
0:07:05 | if you add the two distributions of the chrome complain military terry one for dating one you of the now |
---|
0:07:10 | have |
---|
0:07:12 | the |
---|
0:07:13 | exactly the send distributions and the distribution of the also |
---|
0:07:16 | okay |
---|
0:07:19 | so it is is the first requirement distributions splitting |
---|
0:07:22 | now we have to find a map being in of the two |
---|
0:07:26 | go from |
---|
0:07:27 | the distribution of |
---|
0:07:28 | well can which is gauche shown |
---|
0:07:31 | with a assumption |
---|
0:07:32 | to is the new distribution when you want for example to embed bet one okay |
---|
0:07:38 | we can find plenty of different mappings of course |
---|
0:07:41 | so |
---|
0:07:42 | C |
---|
0:07:43 | we can that holds this mapping being uh function T get beat L T |
---|
0:07:47 | and so the requirement |
---|
0:07:49 | as a already say so |
---|
0:07:51 | you have to be to to this |
---|
0:07:52 | cribbage |
---|
0:07:54 | so the requirement is that you want to find this because the mapping that will minimize embedding dating distortion |
---|
0:08:01 | would do we so |
---|
0:08:02 | so this is important because you have |
---|
0:08:04 | but you can invent of different mapping but is only one that's |
---|
0:08:09 | will minimize the average at two T distribution |
---|
0:08:13 | and here we use optimal once spot in order to do this |
---|
0:08:17 | so the optimal tons pulse you re give a uh implicit form we'd out of the being in the is |
---|
0:08:23 | the case of a scan a distribution so in in one D |
---|
0:08:27 | it's a given by the |
---|
0:08:29 | mean it |
---|
0:08:30 | relative distribution function cat be tied at all of the density of |
---|
0:08:34 | the what that mark content |
---|
0:08:36 | uh |
---|
0:08:37 | i |
---|
0:08:38 | that is applied on the cumulative distribution function of the host content get beat that F four S |
---|
0:08:45 | this is the that |
---|
0:08:46 | the mapping from your |
---|
0:08:48 | you can also derive a is the and bidding being distortion it's right then here |
---|
0:08:54 | and so you can |
---|
0:08:55 | once you know the distribution and you can compute the mapping for |
---|
0:08:59 | he distribution uh you you shows them |
---|
0:09:03 | uh we're is an example so still a my oldest below |
---|
0:09:07 | the be shown than the target distribution and one dating is you were |
---|
0:09:11 | you a a one example of such a mapping here |
---|
0:09:15 | so this mapping being will and able to minimize at this |
---|
0:09:18 | the the distortion on average |
---|
0:09:22 | now |
---|
0:09:23 | is that everything is done you you you you you you can choose different partitioning function of your distribution you |
---|
0:09:30 | know either |
---|
0:09:31 | to uh that form baiting |
---|
0:09:33 | this is what was called the a once upon that you are what our marking |
---|
0:09:38 | so you the distribution is very simple if you want to one bad you will all the cool quite if |
---|
0:09:43 | what are not |
---|
0:09:44 | a coefficient of an i on the left side of the potion distribution |
---|
0:09:48 | if you want to embed one |
---|
0:09:50 | it gonna be on the right side |
---|
0:09:52 | this is not in at because you have only one a coding region of dictionary region |
---|
0:09:58 | uh and it was a it comes from a use what |
---|
0:10:02 | uh no or you can |
---|
0:10:04 | play with is the distribution and design new distribution here is an example where a or coding and as the |
---|
0:10:11 | set an uh are and we also the same probability |
---|
0:10:16 | so the the a here is the same than this one and so long |
---|
0:10:19 | so i called E |
---|
0:10:21 | he not you are what that marking be close all |
---|
0:10:25 | the distance is the |
---|
0:10:27 | region at the same probability P |
---|
0:10:29 | in all the two |
---|
0:10:32 | and as its and probably T P yes |
---|
0:10:35 | uh you can also see me try a seem that tries this uh D |
---|
0:10:39 | the and then you are |
---|
0:10:41 | P bar |
---|
0:10:42 | that you are what they marking it exactly the same than the previous one but |
---|
0:10:46 | i perform a tree according to a a or or uh around zero |
---|
0:10:51 | okay |
---|
0:10:52 | yeah again i managed to a of the pleading distribution requirement a and and they able to to from and |
---|
0:11:00 | dating |
---|
0:11:01 | the last |
---|
0:11:02 | partitioning function i tried that |
---|
0:11:05 | is called |
---|
0:11:06 | the no |
---|
0:11:07 | that natural are what a marking where a or or then beating uh region as the same way |
---|
0:11:14 | that is code then does so it's very similar you know to this |
---|
0:11:18 | can of post test scheme all to you a i am but on you know you have this uh secure |
---|
0:11:24 | and bidding in this case |
---|
0:11:27 | okay |
---|
0:11:28 | so i will now uh compels of defer and to uh what a mapping method the a regarding uh a |
---|
0:11:35 | tell of eight okay |
---|
0:11:37 | i want to evaluate knows the last constrains the robustness so i will uh i'm but now assume uh additive |
---|
0:11:44 | white caution noise uh channel |
---|
0:11:47 | and here is the compare is and between the that than that you are what that marking and france pop |
---|
0:11:52 | net you are what that marking needs |
---|
0:11:53 | just |
---|
0:11:54 | to evaluate |
---|
0:11:55 | the benefit of doing |
---|
0:11:57 | inform it coding |
---|
0:11:59 | and as you can see so |
---|
0:12:00 | he's low well well a plot is for |
---|
0:12:04 | that than that you are what that marking so you can see that you can that shape |
---|
0:12:08 | uh small of bit error rate |
---|
0:12:10 | once you can see that in front of coding spatially for lower that uh the value you in a |
---|
0:12:16 | the you and a the what them map and of the ratio of cool |
---|
0:12:20 | so |
---|
0:12:22 | this is so of first benefit |
---|
0:12:24 | no if i can they so as a different |
---|
0:12:27 | uh and they so that than that you to what they marking P that you are what that mapping and |
---|
0:12:32 | P bound that you are what yeah marking |
---|
0:12:35 | again a again but if they are and what are marked to knows the ratio |
---|
0:12:38 | you can see is that for example in this case we have a what they map to content which sure |
---|
0:12:43 | of minus five db |
---|
0:12:45 | in this case so first the P uh and that you are without marking |
---|
0:12:49 | gives the best yeah formants below minus five db and that the are you have to use the P bound |
---|
0:12:56 | that you are what that marking in of the two |
---|
0:12:58 | uh degrees |
---|
0:12:59 | you're a bit error rate |
---|
0:13:02 | if |
---|
0:13:02 | i i degrees again to what that mark to content issue |
---|
0:13:06 | i i uh no sorry yeah |
---|
0:13:07 | um |
---|
0:13:09 | i wanted also to compare now we is and secure but known to be robust what the marketing scheme so |
---|
0:13:14 | you have to mail for |
---|
0:13:16 | to more plots |
---|
0:13:17 | is that |
---|
0:13:18 | got more less picky read for uh in bits picked one here |
---|
0:13:22 | so you can see for example that the improve |
---|
0:13:24 | a like all for of course low well bit error right for the value and and |
---|
0:13:29 | and this get us but that's scheme E which fill also |
---|
0:13:33 | low low well it or right so |
---|
0:13:36 | oh |
---|
0:13:36 | is this is a kind of a known in in the secure what they marking it's very out to be |
---|
0:13:42 | both robust and very secure |
---|
0:13:44 | this the rubber scheme yeah wow |
---|
0:13:47 | a more robust and the secure one |
---|
0:13:50 | and knows of embedding distortion and minus |
---|
0:13:53 | eleven D V yeah you can see is that |
---|
0:13:56 | main is the |
---|
0:13:58 | uh that than that you are what that marking out their forms |
---|
0:14:01 | the was a a a a uh and dating scheme |
---|
0:14:05 | so the bit the it much real well for this one |
---|
0:14:08 | if again i compare our we is this can ask what S T in uh and uh |
---|
0:14:13 | is the input speech spectrum |
---|
0:14:15 | i can see is that |
---|
0:14:17 | the the kind of "'cause" that's in give |
---|
0:14:20 | even beta a bit error rate |
---|
0:14:22 | even if |
---|
0:14:24 | as a at the the value and they to zero you are very you are very close performance as |
---|
0:14:30 | so |
---|
0:14:32 | it is also interesting |
---|
0:14:34 | no i will conclude group michael so |
---|
0:14:37 | um |
---|
0:14:40 | it's its importance so to perform distributions coefficients so to split your this |
---|
0:14:45 | or and you know up to achieve a |
---|
0:14:46 | secure and baiting |
---|
0:14:48 | you have also to find a way to match the two distributions so the one of the all sins one |
---|
0:14:52 | of the what they not go though |
---|
0:14:55 | in order to do so uh a to do this you can use of to much from sports |
---|
0:14:59 | it's very it's walks fine in one dimension mention |
---|
0:15:02 | is the problem is that in a a multi dimensional it's more complicated but you can use all also of |
---|
0:15:08 | optimization trees and mean know up to minimize the distortion |
---|
0:15:12 | uh there is a gain of using in from it in also for secure and baiting |
---|
0:15:17 | and uh and thought and that is so the best partitioning depends of the and dating |
---|
0:15:23 | used option the but use yeah and also of the noise |
---|
0:15:26 | so i |
---|
0:15:27 | we don't have a a a a a magic old up to perform both secure and and and dating |
---|
0:15:34 | the best pick either a fell to link |
---|
0:15:37 | what we can do using secure adaptation of the get us but that's you |
---|
0:15:41 | so they have been applied by around the last week and is using at at at that position of a |
---|
0:15:46 | uh this yeah that's but that's scheme |
---|
0:15:49 | and they are or so a more fundamental problems as like i would to compute exactly theoretically is a secure |
---|
0:15:55 | get by is so |
---|
0:15:56 | how much information they can convey a while keeping being uh security |
---|
0:16:02 | and also |
---|
0:16:04 | another problem is that |
---|
0:16:06 | if you want to what secure embedding betty a secure uh and baiting uh |
---|
0:16:10 | with is low bit error rate you have also to use secure uh iraq weighting |
---|
0:16:15 | yeah well only can code |
---|
0:16:17 | and this is uh not known right now also we have to find a way to |
---|
0:16:23 | improve the quality of cell is but also guarantee security it's not easy because if you use uh |
---|
0:16:29 | they're all calling |
---|
0:16:30 | encode you gonna add dependence between the |
---|
0:16:33 | symbols that you're gone and that |
---|
0:16:35 | and this dependence can be a security uh equal |
---|
0:16:39 | thank you for your attention |
---|
0:16:45 | and the time for a couple |
---|
0:16:52 | that to a kind of nice the |
---|
0:16:54 | didn't on the first like to set you word and work on a C is just a done the quite |
---|
0:16:58 | sure what type of security does your |
---|
0:17:01 | net the dress |
---|
0:17:03 | and and is you cure in the first place because you kind of |
---|
0:17:06 | i use you could in first two slice |
---|
0:17:08 | you never argue why he's |
---|
0:17:10 | Q also okay |
---|
0:17:11 | so uh |
---|
0:17:13 | here are the signal you is that so that very will have only what a market the content and we |
---|
0:17:19 | try to |
---|
0:17:21 | it's steam so you will have a bunch a |
---|
0:17:23 | what a market contents |
---|
0:17:25 | all more more to be uh what L mark only a set of then will be what they matt |
---|
0:17:30 | and we assume that all this can this as a cushion and and i a E okay |
---|
0:17:35 | so the goal of that the we will be two |
---|
0:17:38 | for example is to the set |
---|
0:17:39 | oh |
---|
0:17:40 | seven |
---|
0:17:42 | that uh |
---|
0:17:44 | convey |
---|
0:17:45 | so what am a |
---|
0:17:47 | so if you want to do this |
---|
0:17:50 | i i has to use some kind of uh uh security at that it can be for example looking at |
---|
0:17:56 | the distribution |
---|
0:17:57 | fine cluster |
---|
0:17:58 | yeah from independent can it uh and component and that is this would you have |
---|
0:18:02 | very very different way to perform from uh what uh what am looking at that in the security uh a |
---|
0:18:08 | lot at the security level |
---|
0:18:10 | here |
---|
0:18:11 | we uh i assume that |
---|
0:18:14 | is there is not possible at that |
---|
0:18:16 | since |
---|
0:18:17 | all the coefficient have the same distributions so what it is that the sre |
---|
0:18:21 | won't be a bore |
---|
0:18:22 | to know which coefficient |
---|
0:18:24 | i what alma |
---|
0:18:27 | and still not quite a what what |
---|
0:18:29 | what is the goal of the attacker |
---|
0:18:31 | i'm still not getting a |
---|
0:18:33 | so what is that the vet analysis that you trying to |
---|
0:18:35 | event of time |
---|
0:18:36 | so it's to if you at this you know you a few out of the N coefficient and only you |
---|
0:18:42 | set of and coefficient |
---|
0:18:44 | conveys or information |
---|
0:18:46 | but |
---|
0:18:46 | is the N coefficients the location of this and clean fish and uh come from us |
---|
0:18:51 | secret key |
---|
0:18:52 | so the of the idea so you will be two |
---|
0:18:55 | uh try to estimate |
---|
0:18:58 | which coefficient fusion |
---|
0:18:59 | conveys or information |
---|
0:19:01 | okay |
---|
0:19:02 | so |
---|
0:19:03 | and then then the next |
---|
0:19:05 | why that's the case why is this in system because it didn't see any secure you know |
---|
0:19:09 | i only solar boston's in |
---|
0:19:12 | yeah the security is grounded because you assume |
---|
0:19:15 | because uh with the em baiting |
---|
0:19:18 | you have exactly the same distribution before and after of baiting |
---|
0:19:21 | it's means that |
---|
0:19:22 | if i want if i plays a but yeah i and and the i am that that's every |
---|
0:19:26 | i won't be a ball |
---|
0:19:28 | to locate |
---|
0:19:29 | which coefficient carries a what that matt so if i want to |
---|
0:19:33 | to try if i tells a say J you have to add the noise on all the coefficient of my |
---|
0:19:39 | team |
---|
0:19:40 | i cannot say act |
---|
0:19:41 | which one carries a a age |
---|
0:19:45 | as a a just every |
---|
0:19:50 | a left so offline |
---|
0:19:54 | yeah |
---|
0:19:55 | right |
---|
0:19:56 | so thank you know match again |
---|