so the person methods um this paper okay and assume the presence of so when for and the lab as comedies and uh_huh of a speech two systems where two a this just that the sequence of presence an to assume a present and number some the and so on so should just a but most work two a is a web i'm um uh uh if become a member of members are bones were exposed was in like to as or have has been to do so um uh if you want to become a member can the to and of the members of of a spectrum to some uh and two um we have a to as shows but that's prince and uh the fruit uh well as a perhaps or i'm so it from us and uh is but a man to talk about the process the mission so this session as the presentations of them after each presentation of the few is for questions and and the um more to have come to ask questions to a sparse image okay thank you very much on the hawk and that would like to thank of few for coming have have coming here unfortunately i have one of them right of the this so i apologise i that in that's if i'm of a little bit fast than usual no as you can imagine a a a a speech M is very but are we cover lots of a has lots of interesting at as uh uh in the general of statistical signal processing adaptation and learn in compressive sensing and so forth and and and be a question was you know i should be but a but but are or should be a big some topic what want topic or two topics and but that can be deeper and um for consultations with of a common with the of we decided let's pick a topic "'em" at least expose at the m-th at least to what's going on in that at and many fat interested of course you can go uh i i i and check the reference and a to us and would be not to be you more and more yeah input put into that no i would be talking and or so of a come would be thought of it becomes a possible compliment my essentially that are two uh a general approach is that and women each other in this area yeah and very excited about the same as that going to see he takes adaptation and learning to new levels it adds a i have to this a to this very exciting a year i usually like to to start my by showing this fit you're a piece i have a video here i a look about to i one to see how used to something i not sing anything okay so maybe somebody can help because this was playing a or level see something much showing anything to just so i in that see if we have a thousand about okay okay uh that meant to see here is for a lot of them but it's a i'm down to the left them to the right now you tell me that is not a but tell them what to do that is not but i think these that the words and what direction we should go we have a thousand people in a state will they hit each other or not do you tell be find a lot the people that want to on or but each other how but does do that that's incredible behave i will go two algorithms they you were processing that you like learning that you was how they had job okay a to a two that was i um them at this is a highly sophisticated be here but if you if you look at the dock in system each but is a the processing information just a it to make decisions about how to go on to to go and you have seen patterns of intelligent behaviour like this H nature would probably just go to but the me when it just a look at the just start time to understand but yeah going to see how D this behave at is how difficult actually is to produce that kind of behaviour okay sampling to this is a you to video about the up my video you so and okay now but some the example of intelligent him and that you see that we have a very very smart man and here you have an example of this is a picture a a and joe graphic just look are they a a a a all that have the colour last same most and parallel some such they see an obstacle that not and a nice i that the shape and then they can call to them should now you tell me how to achieve that kind of behaviour is that in the fish that and key a position yourself at this location and you position yourself at that location that's not what's happening and this kind of smile in go be behave that that just to that is the that of highly localized information process i come information processing can do you know that to and that kind of sophisticated behave but this out of N one questions and a very interesting question i and if you look at the literature of course people have we study this or many many using the fun disciplines them by would you complete the science and uh in a colour G and and i think now a seven is an hour in our field uh uh a include this because we had also time to to to understand as best as we can that kind of behaviour and to to produce it not that we want to and and make and my behaviour but i think that is something come from them to back to make a algorithm as the smart a robust morning just a a of and so for okay yeah a just but its of information but also by the a from that the for that i had but also a fly in your formation the have sense for that okay and for the call reasons for that but see it before the ration to this kind of that this i that time to be the first one and then a end and the for to the the bound to flight that a and like this is that is out of highly localised in information processing that is and this kind of i a rational behave and how do they do it that interest be have yeah right so a more to look into it get i you get excited at that this is another example of some which are P about a to use that well if some some press i have a it video to show you here but that like to see so if you were this time okay but that somebody to i mean lots of that you have a this sub time and and it was that i in in a dish where is i believe was some oxygen and like where the made and then the but this time and there is at least i was size devices you can hardly see them well that's that that's you to left on the item and them on a motion now when you all the of the and that i score my micro quite this is what to see this kind of but you see these time of devices is or and thus the in a the set that action so that's a here a K i like T V "'cause" i don't know i'm just asking questions that a back to you oh this kind of high gram may behave at a notation in a single direction that that's an interesting kind of behaviour so that's happening at me a and we had a eyes level of so this seven we this thing questions that come up when you stop observing this kind of behaviour in nature okay so if we look at these examples case was but to be okay a much of to see is that what each individual H and in a i is not sophisticated but them together that a the interaction action at that highly localised level needs to oh patterns of behaviour or people have of this for years of course look at the question now for us am and just uh a show it on the next the slide but this is the convolution of this examples that to see here examples are actions and the agents i to sophisticated complex behaviour of course on S as i like good as that can do nothing i'd these it to have some some cognition in them okay know okay but when they put them to to have that it leads to that kind of interesting behavior so can and our a little is but i in statistic signal processing an adaptation and learning in cognition in machine learning can i and make that kind of intelligence that lots from that the but but that's in a question that will like for example an important questions that to make a so that i'm i i behave i i i is is i i love for a kind of that action is you to have to generate a data set kind of sophisticated be here the might be interested in a big a level a of G should exist between a and so that they show an information in such a way that that is and be kind of carbonated behave but that you would like to result okay how much should information be quantized should be used at and share information at the very high presentation level is low quantization level enough i'm sure that the fish number see a shop but they can shared information like this P of the shop but the colour of the shot that action but just a segment and that's that's that just a shot i i and everybody gets i about the shark but this is one example of quantization how much information relation position such a that that kind of behaviour yeah appears and the last question is then a of them you know if you think about it at least example that i sure that have a static networks and is at a time maybe and may was five minutes from now the connections that change of the time this is that high networks and a T had and for is learning that inference is mission okay and the other way and i also be a a a a a to five and that may also in france to an environment "'cause" of this call this questions are tied together and i decide address a from the perspective of a a a a two a from the perspective of signal processing in general are there are many many techniques you come up i that this problem like this to the calm is going to but also about some additional techniques which are should have listed yeah so a you have this to the process can of course is at the heart of a lot of this for the estimation adaptation and and the assume game phonetic a and methods uh i i a statistical as that you can use to try to but that kind of behaviour i like i said people problem uh that uh that's have this these behaviours as and they have exploited that and a different ways you even see and nations of naked behaviours in movies mag but i have a computer that science have generated an issues that you see are going to a uh uh i a to produce this kind of behaviour yeah about insist of course and the the different level of processing the different kinds of questions how the whole algorithm that can generate that kind of high level intelligence right from high localised it's set actions at the local level look at this is a high level question we have mine no of course i don't have time have to gossip algorithms that sure i as we have a we have a a a a a large family of i a can can and i guess adaptation i'm of such conditions you have more networks and that that actions that local i with them up to discuss the algorithm this the of everything but i would like to show some simulations examples to example are the things that are able to do that oh and interesting thing this of a lot of them to come up i to my be B here for a in the back to the that behaviour fish a a and have that as so that that is something that are okay for example a to to see in the first example a a fish i don't know okay we to find a different is that want to to gather have to make to have a fine but the for this and in that direction and that's not a shock peers so they have to be a bit of that environment they have to also to that where the shock so for that is the first object of shop as a moving object that they have to track it and this problem so i went to see the simulation this is not a kind of application i would like to sort of consensus because if fish that's scores as to behave in a different manner than a fish that's found away from the shock and the before this is a kind of example that the as we to have something of age i have like to to add that are but in addition to apply to them from the but it's not enough to for a to teach my senses would you but you have to be a a a a a a a and and that's that's the situation but on your on nice estimate so he had to have and and this is just the of an adaptive a cognitive networks are i i-th that to get that in the and to which are with each other and learning from each other and but skip that and this is an example you see that for a is that that i on that i i a that you have before and the fish like to have as before no this is a a and that if network okay it's went to see it from the stop down here that that was the for for a i think that actions and then that but that is present in the environment at two and then go i this is using a a an adaptive algorithm on the kind that should be for nothing or yeah a highly localized that action and is out and to make the kind of used in H okay this is what zap of the next example of to see two networks working dance each other and also X the the kind of be to find in the true that that's the group of shots time time back a of fish in of the so of the fish a so you have a network of shocks score than a with each other so that they the fish and the network of fish score than a with each other that so that they know if it is where the shots i don't not they should do or that's what are able to see next example okay so he it is a function of mm the shots B also also that have the kind of a but this is a adaptation and oh and this is a kind of thing that i like i said you can expect this fish to check check limit of everybody else that i i to say of themselves are K they do that best like an a show then neighbours but not it comes time to decide for your life then you have to take some decisions on your or or or okay so we start to examples that i not to show a about how patient and let as in high and i make a bad environments of every multiple agents so give some going to stop you a from this gives an overview okay i five yeah yeah is okay uh okay so if there questions of you have any questions as a consequence oh and a and and uh i i thank you thank you about sally things okay can i like to uh i ask become "'cause" number of T are also can has i a um about oh signal process and thank you very much it is going to talk about uh if as a a like to to that this that's per session is not a summary of what's happening at time "'cause" mom i'm sure or we find the uh that is that the that uh put to but we were try to to focus here a a a a a as and we can't cover everything so opens the pressure we can uh oh have an appreciation of what the where the we thank you and thank you are so uh uh a a a i only have two slides here a in the first i'd i just want to give a incomplete list of several oh important areas based on submissions and i guess this year and i guess a extrapolation of what might be interesting so this is an incomplete list is always these for like popping to some few leagues and i apologise to any of that so these are six here is we've listed here the first two signal processing a signal dynamic rows and as many of you know that is actually a fairly important here at the moment we but looking at consensus formation information flows games so that and a quite multi disciplinary as well you you think about it not only does it affect signal processing but i guess if you looking at things like social networks you point to model house a siding functions in terms of some of these a dallas sings out with then another area which is also uh of in not most that our T in in take are are uh S P D and society is a dimensional signal processing get sparsity mean uh if you look at this year's i guess is been uh enormous number of people in that area oh from a mathematical point if few many of these results to you with this very uh seminal idea of concentration of measure your room and and compressed sensing is one such example oh many of you would known that there are special sessions uh in i cast on that the transaction a signal processing which is you must papers in that area i so the to special use of journal selected topic the processing again it's multi disciplinary in people and computer science machine learning and also by a medical imaging the uses that another area which is also a significant importance coming from signal processing this financial time-series traditionally a people in mathematical finance do things it continuous time the use things like stochastic was and so on but there is also a because we did you do things a discrete time and and those was are people do financial time to and of course is a smart grid and and most of us know that that's something with a lot of a a potential as distributed sensing decision making a control and that's really what i'll was talking about it in this part of the top my next slide i wanna give a few more ideas behind that so you can think of these as called a systems the uh ability of a system to use feedback to we can figure its behaviour and on top of bit more about that later and and think this is very multi disciplinary or people in economics of done some very serious work in this area or uh and and also with other is and finding the signal processing in the life sciences so we comes a a laid down from oh molecules how you can simulate the behaviour of small molecules and how drop spine to small molecules that really up to a large scale by medical imaging and instrumentation i just found as little uh snippet but from looking at the nature but side nature or column nature one of the highest in that turtles and by uh and as a top here so that actually companies like pfizer or and no part is which are drug companies uh investing billions of dollars in high speed stochastic simulation at the molecular level to see how drugs buying buying to Q and we use mark of T the column so so that's something which i guess people in signal processing to quite a bit of and uh you might be interested in the things that but that map lab is also developing a so i all the G two which allows no no uh people to and i chemical part ways and also estimate parameters and sensitivity of these part we so so it's be a very active area as the normal significance to signal processing that so that's a but incomplete list of some areas of importance i and i one a in my last slide simply focus on the area of distributed sensing which which i guess uh what or something which are ads on to what all these said in his still so that's just kind of a a look at the big picture your once again so it's standard signal processing we do things open loop we have a sensor measuring a a signal and noise and then we want a process that noisy measurement to estimate the the like signal no you can think of a philosophical that bounce in that were use some form of feedback where the estimate a if signal this fed back to the sensor so that the sensor can reconfigure its peak a pure in time one such example could be a distributed system where the the picture of the right hand side but you have multiple agents sensors or agents and each should are one individual control unit i each guy wants to we figure its behaviour to do something so i mean one can think of a that experiment where perhaps you sensor so is more drink some sort of target too many sensors to and all the the wasting battery life so they don't all all to at all to a few sensors to not if too few measurements and then if you do what day of data fusion and you have a very high variance so all the the sense decide without a thing to all the sensor should turn on a roll keeping in mind that it's not completely capitalistic was company capitalistic it would say to have with everybody else i what sleep but not to at because that present about battery life so a company socialistic goes that it would simply be to all the time but out to die and uh wouldn't be very useful so without talking to other people by using rationality the ability to predict what other senses of there to do how does the decide when it which is on a is it's which off now what economists do uh and really a talking in this slide from economics point an economist would do to add a lie such a problem is they would introduce the assumption that each agent is rational a we we say to devour rationality that of actually a very similar in some sense so by shall only needs the following that if i get a measurement i i can put big from that right are the agents a gonna do because all the had estimation so i are use by measurement to predict what other people are gonna do and what i a based on that and i don't people know that i'm gonna react to a measure it in the way i'm gonna do so that you know what i'm doing but you know that i i know that the you know what i'm doing and so on ad infinitum that really results is something the called equilibrium a types of put the breed a because if you're nash equilibrium correlated equilibrium so on but the main idea being that here as a system on the right hand side where eventually if things were done correctly individual agents like fig or shots or whatever can we can figure the because you would minimal communication to achieve a common goal so one can play was your at all three are really which people and micro economics of studied in great detail of the last ten fifteen years uh the first question can be how can agents autonomously man the behavior and that's really the i guess the question i poles now you can think of it off sure of that as if each image and that something simple how can you ensure that it a little your is complex or ration and again a uh i i just what a point out to one it'll paper we didn't by so your heart use who's a really famous economist i to look at it can a metric or two thousand and five yeah the fifty page people which gives you an amazing perspective of how individual a in a can we do something very simple like grand a simple adaptive filter the performance is remarkable but eventually every agent at a user consensus not just good as an estimation but can as an action that is for all is eventually this same decision policy C so they have some remarkable to that in micro economics how to come up with some and the to the user is really game theory as an analysis but that all also so into reading the learning out with them which further refine how agents can act a second question one could pose the standard power in signal processing typically typical one is looking at agents and sensors each sensor has measurements and it up estimates based of those measurements now you can think of a further the refinement of that where agents not only get measurements but look at actions other each so you look and see what of the people done how could i learn from what other people have done if you think about that and i really is some sense quantization an observation because the agent gets and observation the agent processes that maybe no rational way by bayes rule and then picks an action if you simply look at the action it's in some sense a plant ties version of the observation you but some information so how can be agents learned from other agents that if you do this do things actually work out well there is this really classical uh for example which has been studied in great detail in economics at i just one quickly mention that results in something called a rational for and so just to give you the example of that in nineteen ninety five to manage but gurus C and mum wrote a book the market for reviews so what they did was we secretly for fifty thousand copies of the old ball and i got the best seller a new york times people so this is a best or it must be good T everybody started by it so it rational people i know because close following the actions of other regions see if of the observation as opposed to the action the observation would be who bought the book and then you what was the all as buying the book they want to port so that do you act purely only based all all the actions of other and you can show in many cases eventually or agents and up sampling something called the rational for the old eventually make everybody else and they throw away the actual observations so that means a very interesting work done in this area in economics and if you look at for example the book by we published by cambridge university press and rational words to to you a very interesting perspective on how you can model complicated system and how they before now this also that obviously applications in sensors where you might put them but the she's agent and which to look at the at a particular way i it makes every the each would eventually act in that way so you can think of a applications in in several areas but one question which one it again can pose again it's and by people that micro economics as how to local decisions a it will bill decision making so uh in in a little picture or the better of the slide have uh a on the left hand side a bunch of red and green dots that it's actually just a a a a a very straight for example of a sequential change detection problem so that as you me have a bunch of agents each agent gets an observation of a particular thing uh i and we can't really the observation to the next guy a can relate is you do they detect a change so it's change or it's not change to maybe green means it hasn't changed but means it's changed thing so the first i it's maybe it's change a second that is not change and so on each each make making a local decision if you could or take these local decisions how how a global decision maker decide to say that there was a change in stop that's a standard quick as time sequential detection problem turns out that the more you raise your global decision on local actions take scroll catastrophic leave wrong so you can show that the optimal policy in this case as a double threshold and to look at it's hi country to it so that's of the right hand side of the ball so on the X axis as a probability posterior probabilities a patient detection problem of change and the S S uh and the vertical axis is the decision each agent takes so you can see that i as the belief as a posterior distribution that the changes happened goes no you declare a change amazingly but it gets even larger you signal does not change the reason is you have a discontinue reading a bayesian update because you basing it based on actions of the previous agents and that of course but the around you say there is a change so these multiple threshold policies are cut very contrary to because why would you she and decision if you get more sure about what you measure and it's it's something again which is big study uh in great detail uh by people in economics who really try to more complex phenomena in terms of local and global decision and how decisions affect each of so those are three things are just a i'd mention in terms of applications of some of the methodologies all these side pose a unifying theme behind all of this is how does not behavior affect global will behaviour a if you like the by just you can view to say if i have the structure which is little cool how could predict the function which is global that's or by all the just do they have the structural of different small molecules a wanna see if these modules function in the cell how to predict what the actually do given the structure and it's really really that's all i wanted say thank you X and you much read from so can can see to it that to this you two has become a no research to i and the this is a uh one came out of these two do so a the comments or uh questions from the is i'm so this session is very short but uh with the a lot some questions is then i as one when with had an example of to the case of the model of my malicious uh a up lot of i yeah so uh so i i i could do another example was i do quite a bit of mathematical finance and many if you would do to do that all forty percent of all treating is done by computers oh where street runs now add second time scale and and wants all buy and sell it a millisecond time see little quick enough also a lot of the buying and selling is done by computer programs no many people know that so if you want to design a computer program which deliberately for example started buying your role a the the the value for you would go well if if you are a computer program was running from a reputable agency like woman sex a hope will be go but set as you but anyway you would make but we because of a reputation that the your is really going up are these are the computer programs a custom trained to for a will once as doing they was start by and then you would suddenly cell it's called a speculative can see attack a big liens and the for the other that's can react that's a you're all of the market then you made a lot of money so this of things are widely use an economics the call speculative currency C tax and so if you look at for example how paper she talks about those in great detail to about but you shouldn't and not be should versions of such games for four these are all of dynamical systems so not not none of this is static we talking but dynamical systems that how you can control K be another question we four set session goes only till twelve fifteen so round question or comment if this is not the case are like to thank you again a a to have a with the session and the C or not