Speech Transcript - Closing ceremony

0:00:00	good morning but often don't what if they make some friends
0:00:04	so they only overseas candidate and of all this advantage and he what shall
0:00:10	if i pass the mic to zero that's the civil use of course this thing
0:00:14	every month a like to make a couple analysis
0:00:17	regarding the logistic
0:00:20	thus
0:00:22	the what the right form of you have that the both the and all the
0:00:25	when well
0:00:26	this the channel will be remains one sre possible
0:00:31	second
0:00:32	the recordings of all live sessions
0:00:35	will be available on the button that phones in a couple of days
0:00:42	finally
0:00:43	it then they that it's a little advisable for all parties then
0:00:47	next week from what
0:00:49	this check it out
0:00:51	no cost of my tools you know this essay the data that it posses
0:00:58	okay and can you can e
0:01:01	yes okay
0:01:03	and been wanting in be not really but it
0:01:07	the oldest code is you know that rumpled luigi capable concluded that
0:01:13	i among the college of these speak at what he can really box
0:01:20	no i in the peace not
0:01:23	possible at night is from each color used incomplete you dysfunctional
0:01:30	and a big challenge to all what is it constantly
0:01:36	you also the two keynote speaker
0:01:40	you're one i don't of the expectation
0:01:43	and three on the on the original in smoking
0:01:47	and when i four point nine at uni marks
0:01:51	in of course
0:01:53	and he's
0:01:54	or the present
0:01:56	and all the on fifty
0:01:59	you all but you have three into like walks
0:02:04	well i'm using a program you know speech by professor i don't usually
0:02:12	i think you know
0:02:14	he on the counter abusive speaker recognition
0:02:19	i agree you look at the beginning recently you can't it equals okay in
0:02:26	at all
0:02:28	do you recall doesn't
0:02:32	we used in a preparation forty function of modern for you have
0:02:36	i do not logistic being that some problem
0:02:41	problem will be not all along so you know
0:02:46	non-target are used
0:02:48	a difficult you really do you
0:02:50	and you can be uncomfortable
0:02:55	and then you can be just you for only
0:02:59	and you have been doing
0:03:01	in a technique at most of the vocabulary
0:03:07	and unfortunately corners change in
0:03:11	in you of on the box
0:03:13	could be sold
0:03:14	people hoping to all
0:03:17	all email
0:03:19	then we give up and change you could be one nine bucks
0:03:26	i should have found that the local u w
0:03:28	i want to institute of technology
0:03:31	but i can't
0:03:35	okay
0:03:36	it should have been pretty for you
0:03:38	but you can provide you know
0:03:41	could you
0:03:45	nothing in japan
0:03:47	i don't know non of what was
0:03:51	each meeting you each of about clark
0:03:54	and the could not used a screen
0:03:57	and then we stick together unique look
0:04:01	i don't the kind
0:04:03	the and just like you don't caleb recording
0:04:07	and they are open it
0:04:10	use and you used in the original from time
0:04:15	i data you have many are not used in
0:04:18	buddies
0:04:20	maybe you have already a number but in
0:04:23	the last presentation time doing this past three
0:04:28	and that we have applied your presentations i
0:04:31	we you don't
0:04:35	i have not and it
0:04:36	apart
0:04:38	what is the voxel several and i mean will be impressed
0:04:43	i the warm and cosy on the mel-scale community
0:04:49	okay you to use the screen and the united together to make a low
0:04:55	and b
0:04:57	in this corner right
0:04:59	strings knowable
0:05:04	if you want to be able to
0:05:06	one example for the only big next year
0:05:08	here we used a buyer institute
0:05:11	you already there
0:05:14	i'm two hundred and one tuning
0:05:17	and pretty enjoyed recruiting
0:05:19	and you
0:05:22	i guess a lecture of ssm about
0:05:25	no like to ask them like to really gmm fishy
0:05:30	so it's all about the of what
0:05:34	the interface
0:05:41	it can be also can see my site
0:05:47	only ten year
0:05:49	see my site
0:05:53	i guess
0:05:54	so this is seen in ages from and i just tons and on behalf
0:06:02	speaker policy look at local when i think comedy
0:06:06	i want to mentions a few of was
0:06:09	i don't you have already enjoying
0:06:13	it's mixture and then i wonder why zero this paper so
0:06:18	and we have also select it to
0:06:23	this paper award
0:06:25	the moment sitting around the means of this paper would
0:06:29	another nine so just got fees the student a powered
0:06:34	so maybe
0:06:35	explain fast and how we chose those i was
0:06:41	so this fast reset acoustic source state best based on the view at school a
0:06:46	also recommendations from a bus
0:06:50	then six point it was they faced papers presentations
0:06:56	and then provide a school for each candidate
0:07:00	so as to be seen be computed it of at school
0:07:07	selection bananas in "'cause" a resurgent hands and unsummed
0:07:12	and prophesied sienna
0:07:14	fantastic for my that pronunciation and meeting
0:07:19	jean francois clustering
0:07:21	question of thus and the myself
0:07:25	i did ms four
0:07:27	so i valuable time so on
0:07:31	then
0:07:33	this is ugly stops a candidate for bass eight a lot
0:07:38	then it all expect that now we need to estimation it back test set for
0:07:44	into a speaker recognitions it and by
0:07:47	then you "'cause" the middle
0:07:49	big so and a battery
0:07:53	paso a use t matrix sign based on a speaker verification but can't
0:07:58	it invites on the
0:08:00	the unseen by shell late again
0:08:04	gene done ten
0:08:06	so you don't mind that initiation
0:08:09	and that in da and d you know vad in what is for speaker verification
0:08:13	is
0:08:15	it and by she s analogy
0:08:18	past and christian us should i'm gonna f c
0:08:24	and then the bad from phonetic i of at inc what representations and so that
0:08:29	is the and
0:08:32	ninety combinations and shows you named julie and so that use out of the and
0:08:38	thirty cutting each of
0:08:42	keyhole sure so i can i couldn't pronounce quality
0:08:46	and the last one is using monte solution fusion ups feast competition you know in
0:08:52	advance and three s p it and i change and one i mean g
0:08:59	and for me
0:09:01	question
0:09:04	and
0:09:05	and it's time to announce to this fun
0:09:10	chosen by something else
0:09:19	so this paper i lost all do nineteen it all expected and i mean age
0:09:24	estimation it but classical set fourteen "'cause" theocratic mentions it can bite then you "'cause"
0:09:30	male quick same and i do not be
0:09:34	from john hopkins university united states make a
0:09:39	conversations and
0:09:41	i four
0:09:43	so all assess a whole stack was that is
0:09:47	if so and nine want to have a few us from then it is possible
0:09:55	only
0:09:58	"'cause" you usually
0:10:00	yes you great how this is a daniel garcia rimmer
0:10:06	thanks a very much to well the comedian for select in our work i don't
0:10:11	with great grand island are the ugliness
0:10:14	but actually a based on just of there is application to jar
0:10:19	also one or even though this one this is actually
0:10:24	has had a using phones in my career and a testing data
0:10:29	the reason one and here it has a look too quickly so everyone are really
0:10:34	good is these are working
0:10:37	so we want to me
0:10:40	and thanks for the was so i had on a on and are still interacting
0:10:45	with other people in this like
0:10:47	i have to a unique opportunity really "'cause" i cannot given that we can all
0:10:52	the common to use you
0:10:54	so it's been actually
0:10:57	so you very much
0:10:59	eight and two minutes or seconds
0:11:02	and that we also and so maybe have his
0:11:08	so we decided to make that
0:11:10	a thank you yes this is that i i'm anyone to second one then used
0:11:15	to having that's great session project yesterday's
0:11:18	okay process is a greater than the
0:11:21	generated so thank you are very much
0:11:25	thank you much and then and i want to pass a microphones to me
0:11:31	and then joel to announce that this dude and i would i mean and that
0:11:37	you
0:11:39	thank you very much
0:11:41	i'll start by us saying if you variance about
0:11:46	the ward off
0:11:48	so yesterday i was really happy to see so many people
0:11:53	at and the jack godfrey and that are of evaluation benchmarking session
0:11:59	the tell me canyon and i curvature
0:12:03	with a panel of six
0:12:06	distinguished colleagues and ryan godfrey representing the
0:12:13	jack godfrey family and unashamed a hack announcing the establishment of this
0:12:20	ward
0:12:21	this
0:12:22	has a significant meaning to all of us
0:12:27	checks passions in life
0:12:30	included science and understanding
0:12:34	not only just trying to achieve the old men performance
0:12:39	but understanding how systems work why they work and the scientific basis for making them
0:12:46	perform better
0:12:48	so that's is passion and of course where is that all begin with student
0:12:55	and jack loved working with students interacting with students
0:13:01	i you know i had the pleasure of well working with jack for a decade
0:13:05	working for him for a decade in working with them for two and a half
0:13:10	decades
0:13:11	and every interaction i had with jack that involve students where there was always something
0:13:17	remarkable
0:13:20	usually we all walked away learning something new in different it might sometimes be about
0:13:27	cultural language
0:13:30	and in the students cases often lead system inspiration which we heard from teachers the
0:13:36	panelists yesterday and just now from a daniel garcia romano and the team at the
0:13:43	johns hopkins h l t c or you where john jack godfrey we used to
0:13:49	work and
0:13:50	with great passion
0:13:52	so and fortunately the recording of the session is available and encased anybody miss that
0:14:01	i encourage you to watch it just see the context for this award and now
0:14:06	let me handed over to tell me canyon
0:14:11	okay so are plans for is a group or more or two
0:14:19	to do this system and a little used in every room is present a real
0:14:23	room from
0:14:24	conversely some of the convolutional will remember from source to
0:14:29	so actually the performance will remember performance was miserable so it doesn't really underscores emotional
0:14:37	armour from everyone used as a front on a proposal on of on one produced
0:14:43	from there is serious i'm not so the military imprisonment this is very differently
0:14:49	personally on a remote we also a good as cousins one thousand two will be
0:14:55	the this is not a scientific principle this remote some of them with remotes
0:15:02	could be able to probably on a list of learning more from there is one
0:15:08	very from here on a remote
0:15:11	so it's really one or several that's over there is a process a three presents
0:15:18	the number of everyone are four hundred and one i'm also a not particularly good
0:15:25	remote simmons so i'm trying to because i think you two remote controls
0:15:31	so we are to go members your be or from using different paper on the
0:15:37	one room b alluded paper
0:15:40	then we are rows one on a similar recall removal and round robin the most
0:15:46	popular "'cause" we're going on one remote but underlings possibly remember system
0:15:51	a and b o one is meaningless from the was
0:15:57	one of these are drawn to the three gram model and language native speaker notices
0:16:02	and using speaker role improvements on
0:16:06	okay sorry we're and no one of the remainder of propose a good proposed remote
0:16:22	this the well known as remember me probably more generally almost improvement is on a
0:16:27	reporter no never remembered hundred and possibly the roses and
0:16:33	so we can probably on if the authors on recently or maybe they can serve
0:16:38	reversals
0:16:41	you
0:16:43	yes
0:16:44	okay
0:16:46	thank you and lee
0:16:49	i nice to operate yes
0:16:53	we are going to be the mean like i think or you another one because
0:16:57	organisers to their nonweighted maker so they don't similar like or not
0:17:05	okay so that it so that
0:17:10	and that simply
0:17:11	and
0:17:13	i very you read on
0:17:18	and congratulations on
0:17:20	i was you or something and joy
0:17:25	and
0:17:26	and then to everybody and thank you also to jack three
0:17:34	yes thank you
0:17:51	and then this is that and of the old fast in many
0:17:55	can you can move onto the next time was decisions
0:18:02	sat they still only
0:18:11	thank you
0:18:16	right
0:18:23	okay the dataset one is you would like to design a certificate of recognition to
0:18:29	a loaded or speakers
0:18:32	so i with this and this one not in any of that
0:18:44	so clear that this kinda stuff t and they ist japan
0:18:49	to those speakers of the topics
0:18:52	the what's their opinion on the shin speech interface the that is and t and
0:18:57	this and was taking
0:19:07	the value of terrible with the results of that
0:19:11	judges filters
0:19:12	and the topics neural speech recognition
0:19:25	miss the set at time either
0:19:28	where is that japan to toss features on the topic was speech recognition
0:19:41	of the civil
0:19:42	which is the most informative japan to their speakers and the topics
0:19:47	we will statistical parametric speech synthesis
0:19:59	the that massively journal of this for your conference to their speakers on the topics
0:20:05	and these fools the in automatic speaker recognition
0:20:16	last i don't is
0:20:18	note that you're and within well known as the of technology to their speakers on
0:20:22	the topics
0:20:23	as speaker recognition why when and how to do
0:20:36	okay now once again by and idea
0:20:39	to give us a summary
0:20:42	for the obviously didn't the option
0:20:46	like this
0:21:00	i everybody conventional slice
0:21:03	okay so capable over some of the highlights all these work your
0:21:09	so overlappers okay and organisers or the one amazing you know given the sequence of
0:21:17	all the whole weight and you got have to do that whilst at the beginning
0:21:22	named and you have proposed a new data to get a chance completely the impression
0:21:27	and make it remotely so really what one and about my heart and everybody here
0:21:33	we did you have been successful
0:21:37	with emotion
0:21:39	so this is this motion vectors these tutorials that was not are only in the
0:21:45	final step earlier estimate role give us some really nice learning and speech and bindings
0:21:52	have given by flavour of that
0:21:55	let's i believe that really less than one and of course of calibration you know
0:21:59	and everybody to death so that's what i don't remember that
0:22:05	we also have the menu or something was privacy issues of training
0:22:12	we should but this what we see mentioned it so i'm sorry of the five
0:22:18	are also one but i would probably do what don and women
0:22:24	you know
0:22:25	consists in a remote just fine you are not at all
0:22:30	so far as we have about eight to seven based within decision obviously duration itself
0:22:37	is the main over and plastic a language based on an easy life are able
0:22:43	we have corpus regularization
0:22:46	most moving in other measures in we have a special session was a lot lately
0:22:53	voice conversion set is this
0:22:56	you know there are nice this and evaluation benchmark in you know only a nist
0:23:01	is used in video audio and video so work about it you know there are
0:23:06	also one about speech applications you know the overall you know isr annotations
0:23:13	and also have a nice class yesterday memorial of jack three on the one friend
0:23:19	and the more in this
0:23:22	so maybe you can imagine working
0:23:26	so first of all you know for this little sister this is a very quickly
0:23:30	so the precision at all possible future the best are war by danielle and an
0:23:37	n-gram so congratulation
0:23:39	so this paper is about how to learn you know mismatch duration by nor and
0:23:44	its car i don't like at the end of it was there you know
0:23:49	inside and it was actually complementary colours also results are so we go back to
0:23:55	what one and one video
0:24:00	we also have proposed a extractor on a which you know
0:24:05	which
0:24:08	one of the best free best and or more so on their solution and you
0:24:13	know that this paper because and are already some uncertainty in your statistically conditions not
0:24:19	like just one year some as in the around how much you're gmm estimation
0:24:25	this for source on that are not a clustering of them or something and variance
0:24:30	you know we also have some supervised training well model the baffle you know and
0:24:37	they show that the future that extracts from this networks or you know about integer
0:24:43	linear programming and of this is more speaker and then used a list you see
0:24:49	what else that's is you know whether a speaker recognition and there's over the of
0:24:55	you know
0:24:57	that's thing different embedding is willing did not improve and robustness of duration you know
0:25:01	that's important so even for the analysis you know can that and you know people
0:25:07	are really trying to so it is the prior to the problem of duration mismatch
0:25:11	and in the bn nn framework
0:25:14	class was the least we now that is able to for the nn so is
0:25:18	still there so far were explored in this
0:25:22	this is one of the feature for this new error vad learning can then be
0:25:26	seen speaker recognition language
0:25:30	score fusion and success housing in extraposed a cinematic space not in the feature really
0:25:36	i don't cts mismatch between training and about how you that
0:25:41	i don't know in addition to that can sell you know the you know classifier
0:25:45	itself does also the topic
0:25:51	would have that it diarization you know the suspicion diarization and one of the domain
0:25:56	mismatch a nice feature going back to channel
0:26:00	and can see that is still problem is how to and whitening the there is
0:26:04	this is how far
0:26:06	and so even when you use the state-of-the-art speaker embedding you know and a speaker
0:26:12	model and you learning is the problem is you don't want it on a mishmash
0:26:16	so
0:26:17	so be released to look into that improving clustering you know you know and we
0:26:22	have little or paper that you know that are sensitive to parameters in validation by
0:26:28	one or whatever was in dover algorithm which perform effective wadding across various additional business
0:26:34	i think was andreas is d
0:26:38	are in this is so
0:26:41	that was to i was information to this basic iteration and working with limited vision
0:26:48	be were there are more
0:26:52	by the you know minimum cost someone their own name
0:26:59	so you know there was information alarm area under different wainwright unit somewhere it happens
0:27:06	people say you know it's to me and mine english you know so that's a
0:27:10	special one so i think is permissible then with a single curve routine that there
0:27:13	is you can also how to separate considering the same tree
0:27:18	that was a good thing is the always this is kind of you know information
0:27:23	a compromise or you know that it is the first one there is a lot
0:27:27	of you know and as a state-of-the-art on nist through to that selecting challenge you
0:27:35	know frequency masking locking also "'cause" i lost motivation for this task also seem to
0:27:41	be helpful
0:27:43	refers to learn on them automatic speech recognition of the matter whether and how diverse
0:27:49	on automatic speaker recognition using feedback control was conversation and finland water
0:27:59	so there are very the ml based you know that you know i'm here is
0:28:03	giving rise to go and one of the meeting you missed and that's the beauty
0:28:07	imponderable we're gonna was done by and you know used and then i and watch
0:28:11	speakers of than going to move you know
0:28:15	so a very nice thought a system that the voice twenty two which was especially
0:28:22	association there was a feature to do speaker recognition in far field you know sixteen
0:28:30	and to focus lost
0:28:33	you can everyone
0:28:34	but you know there
0:28:37	i was literally microphone channel you know i'm so that was also a this is
0:28:43	one of the speaker really in the future so we have to learn how to
0:28:47	do it is as we find that are involved in unison everything but if you
0:28:54	don't of the progress there
0:28:56	we made use progress and now this is a new area where we need one
0:29:01	and how humans or a is the thing
0:29:06	voice processes that is you know and within a very of papers about using spectral
0:29:11	what the variance along rather well against an all us i guess for singing voice
0:29:18	conversion a nice interesting to see he's on your own voice test you know
0:29:25	and i will assume how can sing beautifully you know
0:29:30	so please "'cause" this paper is
0:29:34	yes so this is a
0:29:38	evaluation benchmarking its colour
0:29:43	features the nist speaker recognition evaluation and here s the you know it shows that
0:29:50	the nist two thousand and eighteen a lie detector machine was a very good success
0:29:55	with other features you audio and visual a visual and visualise at or images
0:30:02	and we well as to combine two and here is that we show under different
0:30:07	paper presented their you know combine still that how long as those though and also
0:30:14	all have already spoofing you can measure you know make the design is just and
0:30:20	right column
0:30:21	that was really good session and but personally i was really one global people don't
0:30:27	you know it's features along case you know you know what is mainly
0:30:33	then post it is given versus those different model more robust models are explore you
0:30:39	know and you know basic task was to have a lot of the goal in
0:30:43	mismatched condition like language you know like that so this is due to these features
0:30:50	only shift for a year
0:30:54	another you know statistical thing on the matter you know and here is to think
0:31:01	i like you know and careful you know carefully if you know
0:31:09	if you have the notation so we have to be careful how the segmentation because
0:31:13	one can have or physical access to not have access
0:31:17	so to have one a carefully designed for a human a result and carefully designed
0:31:23	to assess to require in order to buy something recognition so that those design are
0:31:30	actually one that's longer than there is the one going
0:31:37	so you nist speaker recognition you know calibration in the in the back end issues
0:31:42	with an additional across various condition where reducing or side-information and then there's gender dependent
0:31:49	condition integration stage where side information is learned jointly what the rest of the model
0:31:54	all z
0:31:56	an operating base where into that you are we expect or a speaker locations us
0:32:02	to do re there was just considerably also interesting
0:32:07	you see maybe too optimistic irrigation work well over multiple application scenarios you know and
0:32:15	over different or operating point it is optimized parameter of a square mile manhattan distance
0:32:21	metric to maximize the parallel versus area under the raw
0:32:25	quicker or interesting rest of a false positive rate does not ignore the biggest or
0:32:32	weighted sort of this paper from all
0:32:35	okay
0:32:37	and use it for a set of training from of to learn speaker and i
0:32:42	think that could disconnect speaker but it's invariant to environment you know something really you
0:32:47	know how diverse a training is coming over complete having lots of this new image
0:32:54	processing
0:32:56	well one
0:32:57	in this a new problem we should maybe more
0:33:02	and if you is just as in the audience doesn't last long as are you
0:33:06	can be efficiently get should work on that
0:33:11	so other you know so this is specification on no and over a lot of
0:33:17	combined model a baseline forensic speaker comparison assessing shot communications speaker detection the while you're
0:33:26	all alone
0:33:27	combining speaker or and that in and one of the information no this also nist
0:33:32	all home and no not try to combine you know in the same embedding it
0:33:37	is sometimes useful
0:33:40	you know that this is another speaker recognition
0:33:47	no decision and we have personal see that some interesting that's gonna period well i'm
0:33:53	not wrong from the chime challenge you know are defined as this is you know
0:33:58	about the speaker maybe a four pairs and how many conditions of the speaker and
0:34:04	speaker
0:34:06	and so we have speech data with expression and how i-vectors speaker recognition not telephone
0:34:12	number one or back while they're number one and in the expansion
0:34:18	by a sufficient has to
0:34:19	and still the lid system in a single there on the nn framework using a
0:34:25	question that will also good as you know the far-field task maybe shares
0:34:32	analysis is of the for unit is just a that's also missing them in the
0:34:37	second one is
0:34:40	that's so i would like to minimize to this as a null and j and
0:34:45	have you of all this together and you know
0:34:49	so i didn't ask whistle
0:34:51	thirty s and enjoy the
0:34:54	well this animal of this you all animals used in the next one easy for
0:34:59	me
0:35:03	i say that it for the precise and to the for a summary which compress
0:35:09	but i comparison rate was able to the ten minutes
0:35:13	thank you
0:35:14	okay so now i think well as it were quite nicely
0:35:21	of this a ninety then doing so now we have on the to have
0:35:27	what is the one function wireless e
0:35:30	we will bring us to use or what
0:35:33	the see what will happen as in of this a and b and it'll
0:35:38	so of this along with
0:35:42	it is too much
0:35:47	one and two well known you to harness t and we were or whether a
0:35:53	onto other to present something about but unlike the only thing to
0:35:59	and you don't into and it would be having made in china
0:36:03	and this will be job of annihilation fitting problem to one university shot anyway you
0:36:11	using an annuity a simple heuristic
0:36:14	and this is i of our strongly indian problem for what was this
0:36:22	and so this is a slight problem was that are correctly
0:36:30	a remote i looking at different screen
0:36:35	no
0:36:36	that this sharing because i v
0:36:42	the screen for
0:36:45	we have all universities can you see
0:36:48	no we have is that the first six
0:36:53	is doing the first page
0:36:55	i think i think we mentioned a different screen
0:36:58	the presentation more and
0:37:04	so how the writer
0:37:07	yes now we have a state k
0:37:11	okay i don't know used for screenwriter
0:37:15	so this is that okay
0:37:18	okay cool
0:37:20	so this is of course this like target and this is the over a very
0:37:27	indian
0:37:29	and here sure will have a union paging the
0:37:36	time will be at
0:37:39	was so small
0:37:42	is the u
0:37:45	is it but we haven't drawing a went you clustering to and people is you
0:37:50	mean round tuesday into writing
0:37:53	and the two in this study that no in between but you okay
0:37:58	and in most of time as jenny the it will where will disappear at a
0:38:05	higher so
0:38:08	paging the utterance of reasoning where ut are that a lot of well writings that
0:38:15	you can really try to a or a low wall that we know the problem
0:38:21	we probably will range so that a volunteer is to have transportation
0:38:27	and well basically no to that but probably not decided yet o one is just
0:38:34	a it says that you one investing and study will be willing to the us
0:38:40	quite spacious or a percentage but it is that what would
0:38:44	and another one is
0:38:47	it's a all morning is called my it is there a but besides discrete well
0:38:55	but it sort of you don't well problem of the c but with you can
0:38:59	read a asked to be there but we will use that with where go
0:39:04	and proposed seconds one of the hair cells of trying to
0:39:12	try to make some other social activities and this is really a this is quite
0:39:17	that a patient style of china the cost the we are there are is quite
0:39:26	and also
0:39:30	this is the two of us the will be included about something interesting places are
0:39:39	but you see here
0:39:43	it's o a work okay ageing of this you to twenty two that much
0:39:55	thank you thank you for so long data for what to us
0:40:00	for now
0:40:02	so
0:40:03	of all this that and
0:40:05	this
0:40:07	this feedback allow writing papers
0:40:10	and okay we're research thus by davis
0:40:14	okay i'll
0:40:16	okay
0:40:17	okay so no i would like to
0:40:22	that the my two so my only
0:40:26	okay the we the
0:40:30	the sound a physician those tools on sign
0:40:34	the noise okay a wider this is only around this is because there was no
0:40:39	problem
0:40:41	and a local almost university
0:40:46	like to thank you for what was accomplished
0:40:50	it's also
0:40:53	can be casting committees and a score and looking
0:40:59	a secular match and all
0:41:02	okay i have been working a file together with
0:41:06	scoring can work with a single that conditional that
0:41:09	my main task one memorable asks for u s a e
0:41:14	i s and the
0:41:17	it's policy so let me talk about these
0:41:20	so hostile a me mission our sponsors
0:41:25	that's a really is my face that we can we could have a these scores
0:41:30	into one is a very
0:41:32	lc highness final set condition
0:41:36	small a very we have we could the whole far was almost corresponds to a
0:41:41	free the free registration and the
0:41:45	and some possible a very special fee all also as can be funny to a
0:41:52	saul's
0:41:54	because of a classic sponsors an easy one c i think
0:41:59	google can extend okay sequences
0:42:03	so we would like to express some a signal x two source also
0:42:12	somewhere ones
0:42:13	and then we also show the say thanks to our local team members
0:42:23	okay we have to have a big change is due to cope with nineteen
0:42:31	our movements this local team members
0:42:35	well we basically and
0:42:38	a no matter what is a or assignment and the have also be how the
0:42:44	crystal
0:42:46	so mm a rematch okay everyone a few minutes
0:42:55	and okay thank you very much to all the participants
0:43:01	and the
0:43:02	an ask possible command was the a squad steering committees
0:43:06	a whole we can see okay home base to us to use a
0:43:12	thank you
0:43:14	okay
0:43:15	how about you by all the local can be given about the penalties and
0:43:23	so if there was a good idea and i'm not sure how who k u
0:43:29	then we began phone i
0:43:32	okay this will use the reset button
0:43:37	i think about this to the logan
0:43:41	okay this is
0:43:45	i
0:43:46	a somewhat so i can say that it's a
0:43:51	okay as it got this
0:44:13	corpus on all so
0:44:17	some people have left
0:44:20	when there you have all this
0:44:25	i
0:44:27	where n is
0:44:28	so costs on and basically so the
0:44:35	this six phone videos if you
0:44:51	hello services and
0:44:55	no
0:44:56	also presents
0:45:00	in a
0:45:03	well
0:45:12	right
0:45:16	for
0:45:18	i think the l off
0:45:22	cool
0:45:32	usefully
0:45:36	it's l
0:45:39	but i guess online for the
0:45:43	well can i think you
0:45:47	now we will starting from we has at all
0:45:51	the at enough that we get a b
0:45:54	and
0:45:55	with that of the solution this is physical conference i would say that is then
0:46:03	matthew nice
0:46:05	well it's transformed into average of all
0:46:09	so
0:46:12	to have a
0:46:14	all of you
0:46:16	it'll
0:46:17	still with us
0:46:19	before you
0:46:21	it was i in the ones
0:46:24	well bands
0:46:26	and allow things
0:46:28	so there is a much
0:46:36	can some useful once
0:46:39	okay a model
0:46:42	a daily that of things the about this that and all those as well
0:46:48	and this leads us
0:46:49	the doors because there is because
0:46:52	says that
0:46:54	voice
0:46:56	wife and all those we have a bus the may have anything
0:47:01	possible
0:47:02	it just so much
0:47:04	i'll a seal
0:47:05	okay the reference the open
0:47:08	but i was i was able to yesterday the
0:47:13	painting
0:47:14	in obviously the detected
0:47:17	i guess unless
0:47:18	okay
0:47:24	what we
0:47:40	so i mean and thirty minutes maybe i should say that before you say
0:47:45	i think i this is phase okay
0:47:50	we
0:47:52	he's a half period was given this could be made
0:47:58	i receive a lot emails and everything i is a is this to this is
0:48:03	all value so that a
0:48:08	well it's nice it's nice to
0:48:11	the right you know
0:48:14	it newspaper selection
0:48:19	so psyched of image thinking
0:48:21	s
0:48:22	exactly z
0:48:25	okay bye
0:48:26	the i
0:48:33	is this is done using a than any other ideas
0:48:40	and my family of the search string
0:48:47	they still
0:48:49	a u s
0:48:51	thank you and
0:48:53	thanks for the wonderful work i couldn't ask for a better team to work with
0:49:00	and
0:49:01	all these you know last minute changes you will handle them with such extreme politeness
0:49:08	in great i'm so grateful thank you
0:49:12	thank you
0:49:26	yes the right number is and us
0:49:35	so it is
0:49:54	so you fool
0:49:57	cool
0:50:16	okay sufficiently close
0:50:18	yes that's the thing
0:50:20	okay different amendments eight hundred and of is that
0:50:26	okay

Closing ceremony

Live sessions