0:00:14 | she you not good afternoon |
---|
0:00:17 | i am casey kennington |
---|
0:00:20 | currently boise state university but this is work that i did |
---|
0:00:24 | well i was to build a full university with along with that was long and |
---|
0:00:29 | and i'm gonna give my two cents on |
---|
0:00:31 | a continuation i guess on yesterday's discussion on personal assistants |
---|
0:00:36 | "'cause" we're gonna tell you a little bit about a personal assistant of that we've |
---|
0:00:39 | been working on |
---|
0:00:40 | and if you don't know what a personal assistant is your in the wrong conference |
---|
0:00:46 | you've heard of them you've use them and they're great i mean they their useful |
---|
0:00:51 | not we dialogue people aren't the only ones using and lay people are using |
---|
0:00:55 | quite often quite regularly |
---|
0:00:58 | but |
---|
0:01:00 | when these laypeople use these |
---|
0:01:03 | systems |
---|
0:01:03 | these dialogue systems essentially these personal assistants they do weird things with them and they |
---|
0:01:08 | complain about mary all things |
---|
0:01:11 | and so today want to talk about a few of those things and maybe make |
---|
0:01:15 | a approach addressing a couple of them |
---|
0:01:18 | one thing is that they kind of have a difficulty signalling affordances someone shorter but |
---|
0:01:23 | yesterday and things you can do with your e |
---|
0:01:25 | why doesn't need a book |
---|
0:01:28 | that you need to disney to be signal somehow and it shows be a lot |
---|
0:01:35 | of these sure speech recognition output and sometimes it's great perfect |
---|
0:01:39 | but you know well |
---|
0:01:41 | that speech recognition even if it is perfect does not you know understand |
---|
0:01:46 | that something else that needs to happen here |
---|
0:01:49 | they don't know that understood until it finally does something comes back and the results |
---|
0:01:52 | are |
---|
0:01:53 | maybe what they wanted maybe not |
---|
0:01:55 | another thing is the user has to expressed |
---|
0:01:58 | express their intended one goal |
---|
0:01:59 | that you have to say the whole thing wait for to get back to them |
---|
0:02:02 | and then they can continue wanting |
---|
0:02:05 | sort of like this again with the system |
---|
0:02:08 | looking into that a little bit more if you if you consider a |
---|
0:02:12 | personal assistant on a continuum like there's some one extreme you have these |
---|
0:02:17 | person or systems that i don't even really want to talk to you |
---|
0:02:22 | they |
---|
0:02:23 | want to its apparently easier to predict your life then it is to predict |
---|
0:02:28 | what you're trying to say and so groove allows trying to do this in this |
---|
0:02:31 | is useful |
---|
0:02:34 | on the other side of the continuum you have the full turn |
---|
0:02:38 | personal assistant that is expecting you to |
---|
0:02:40 | given entire intent and then it |
---|
0:02:42 | that was all that's understanding and you do some kind of response maybe there's something |
---|
0:02:46 | in the middle that would be a little bit nicer |
---|
0:02:49 | sub-turn little bit a little bit to the left ear so |
---|
0:02:52 | i say call mom and there's some sort of feedback that it understood be a |
---|
0:02:56 | and i know that understood me a nice to amend it and then i can |
---|
0:02:59 | say on speaker phone and okay good |
---|
0:03:03 | and we can move this may be given a little bit more to the left |
---|
0:03:06 | and say something call a your mom |
---|
0:03:09 | one speaker phone |
---|
0:03:13 | it's |
---|
0:03:14 | exactly that's what i meant to say |
---|
0:03:16 | so there's a little bit production it's not trying to predict your entire life it's |
---|
0:03:19 | allowing it to give at least part of the intent but that's doing some prediction |
---|
0:03:22 | error so we can maybe make our dialogue systems fit some runs continuum that's useful |
---|
0:03:27 | for any particular user |
---|
0:03:29 | we want to look at this a little bit |
---|
0:03:32 | really quick related work some inspiration joyce tries work on misalignment manners signalling understanding and |
---|
0:03:37 | others work |
---|
0:03:40 | on backchannels stuff on arts and |
---|
0:03:44 | work on goodies which we kind of are gonna do here and then of course |
---|
0:03:48 | lose project |
---|
0:03:50 | we would take inspiration from all of these |
---|
0:03:52 | for some reason they're not none of these people here |
---|
0:03:59 | but we're gonna do something using all this all of these as a sort of |
---|
0:04:03 | inspiration so we're gonna signal ongoing understanding |
---|
0:04:06 | you can agree |
---|
0:04:07 | assuming here of course that people have a way to display agree so this might |
---|
0:04:11 | not work on something like the amazon |
---|
0:04:13 | echo but most people have other phones with them and can use the personal assistant |
---|
0:04:18 | with the display |
---|
0:04:21 | and with it with this really backchannels don't overlap speech so for talking and its |
---|
0:04:25 | updating and showing them its understanding then it's not gonna have any problems importantly works |
---|
0:04:30 | incrementally |
---|
0:04:31 | that is word for word are explained that the moment a little bit more and |
---|
0:04:34 | it works with |
---|
0:04:36 | minimal or no training data |
---|
0:04:41 | the rest the talk is as follows i'm gonna explain our system |
---|
0:04:44 | and the components of it and then |
---|
0:04:47 | see if that system is worth its salt |
---|
0:04:50 | well first the system |
---|
0:04:53 | at first blush looks like any other dialogue system you've ever seen their speech there's |
---|
0:04:57 | nlu errors dialogue management there's some way to convey the it |
---|
0:05:03 | i'm response to the user |
---|
0:05:04 | user with technology in but in this case agree |
---|
0:05:07 | the speech recognition i'm not gonna going too much it's |
---|
0:05:10 | google asr we have it modularised here nicely to give us incremental |
---|
0:05:15 | results so word-byword it's coming back to us and we take the those that incremental |
---|
0:05:22 | output from the asr give it to our nlu |
---|
0:05:25 | and are not use working in lockstep with that so one takes a word |
---|
0:05:30 | and we're gonna use the in the simple incremental update model which we introduced in |
---|
0:05:33 | sect dial and that's in two thousand thirteen |
---|
0:05:36 | and without getting technical you can look at the paper if you like |
---|
0:05:40 | equation thing like that you can if you what you get is you don't word |
---|
0:05:45 | and its going to produce a distribution over slots |
---|
0:05:48 | and that's can be given to the dm the dm the dialogue manager gonna use |
---|
0:05:52 | that somehow |
---|
0:05:54 | with this little provision when someone utters a word |
---|
0:05:57 | asr gives us a word |
---|
0:05:59 | that is the same as more similar to |
---|
0:06:02 | a value that could fill a candidate slot |
---|
0:06:05 | then that's gonna get more credit and this is how we are able to make |
---|
0:06:09 | the system work with little or no training data and then build up from there |
---|
0:06:13 | that's no you're |
---|
0:06:16 | but the dialog managers taking these |
---|
0:06:19 | word for word the not use given this these slot |
---|
0:06:23 | distributions to dialogue management dialogue manager has to do something with that |
---|
0:06:28 | though |
---|
0:06:29 | in fact it's making one of four |
---|
0:06:32 | there are simple decisions one is |
---|
0:06:34 | i get a slot a look at its confidence value and what why do i |
---|
0:06:38 | can wait |
---|
0:06:39 | if it's if the confidence values well just sort of ignore it |
---|
0:06:43 | in particular so particular value isn't enough to make the slot the one that i |
---|
0:06:47 | want |
---|
0:06:49 | or i can select something |
---|
0:06:51 | is above some confidence threshold than the slot as good let's fill it with this |
---|
0:06:54 | value |
---|
0:06:55 | or to others here is we're close to that threshold |
---|
0:06:58 | but not quite there so let's make a clarification request and somehow display that agree |
---|
0:07:05 | and then of course they have to be able to confirm that request |
---|
0:07:08 | i want to point out here that it is here between sort of the nlu |
---|
0:07:12 | on the dialogue manager |
---|
0:07:14 | where this and pointing is done we're not doing and pointing with speech recognition that's |
---|
0:07:18 | just always on |
---|
0:07:20 | and it's here that where |
---|
0:07:22 | so they can stop and pause and think and what do something it'll wait for |
---|
0:07:25 | them to finish so they can do things in instalments so it sort of semantic |
---|
0:07:28 | driven and pointing |
---|
0:07:30 | and we can use of and i'll |
---|
0:07:32 | for this it's sort of rulebased at the moment but we have the provisions are |
---|
0:07:36 | there now for |
---|
0:07:38 | reinforcement learning and learning on-line to improve the system as people interact with it |
---|
0:07:43 | now we do we |
---|
0:07:45 | the dialogue manager decides which was to be filled and it says gui here's what |
---|
0:07:50 | the decision i've made please convey this information to the user |
---|
0:07:53 | and the golay you'll notice right off the bat we aren't |
---|
0:07:58 | obviously aren't you are designers |
---|
0:08:01 | but here's the here is that you turn the system on and |
---|
0:08:04 | this comes up it's in java script so |
---|
0:08:07 | and it just looks like a right branching tree and really that's all it is |
---|
0:08:10 | but right here you can already see what the importance as r o we can |
---|
0:08:13 | do these five things are nice |
---|
0:08:14 | i don't have to guess i'd have to play with it in figure out what |
---|
0:08:17 | it knows and what it doesn't know |
---|
0:08:19 | and so i look at this thing is a well you know i am kinda |
---|
0:08:22 | hungry and it will go then into the food domain and sort of open up |
---|
0:08:26 | the treatments a lot |
---|
0:08:28 | if you if you're hungry then i |
---|
0:08:30 | you know one where you want you know what you want and where |
---|
0:08:34 | you're gonna unit |
---|
0:08:35 | and i can say you know i'm among we first and thai food and at |
---|
0:08:38 | that point in |
---|
0:08:40 | go to the top here and |
---|
0:08:43 | shoulders note and read a question mark for this clarification state did you say tie |
---|
0:08:47 | in to the and this to me as |
---|
0:08:50 | into it in that it |
---|
0:08:52 | is trying to understand me and i have to do is say yes or i |
---|
0:08:55 | mean time and that would fit |
---|
0:08:56 | basically feel that slot which |
---|
0:09:00 | conveyed visually means that it just collapses that are the tree and shows like this |
---|
0:09:03 | so the here's a here's a frame that is filled |
---|
0:09:06 | and it shown visually like this |
---|
0:09:11 | that's our system |
---|
0:09:13 | recall right |
---|
0:09:16 | now well we did some experiments to see if that's system it was everything we |
---|
0:09:21 | hoped it would be and where to put some people in front of it |
---|
0:09:25 | though |
---|
0:09:27 | we want to test a couple of things about this system so we're gonna break |
---|
0:09:31 | it up in the basically for different |
---|
0:09:34 | different settings |
---|
0:09:36 | we want to test |
---|
0:09:38 | we want to see if our incremental system is better than or more useful i |
---|
0:09:42 | suppose than the traditional one |
---|
0:09:46 | so we're gonna let them play with that of first and give them a trial |
---|
0:09:49 | phase here's our system here some tasks to do them and get used to the |
---|
0:09:52 | interface and then we're gonna |
---|
0:09:55 | sort of move start on the very right side of the continue one where they're |
---|
0:09:58 | doing this |
---|
0:09:59 | traditional |
---|
0:10:02 | current turn taking full fully intend mentioning |
---|
0:10:11 | personal assistant |
---|
0:10:12 | so and points |
---|
0:10:14 | as usual |
---|
0:10:15 | kind of like the traditional personal system |
---|
0:10:17 | so we then we |
---|
0:10:19 | then move the continuum move on the continuum a little bit to the left and |
---|
0:10:23 | nouns incremental now we're doing some terms |
---|
0:10:25 | and you can |
---|
0:10:29 | do things in instalments |
---|
0:10:31 | and then we have phase three for removing that |
---|
0:10:34 | a little bit more to the left on a continuum answering |
---|
0:10:37 | now it's going to adapt to you a little bit and try to predicts and |
---|
0:10:41 | fill some these slots for you |
---|
0:10:44 | or expanded a little bit phase one acted like a standard personal assistant silence and |
---|
0:10:48 | pointing before they can we would even show and the asr was shown like it |
---|
0:10:52 | is in your standard personal system |
---|
0:10:55 | based to is incremental phase so they did phase one for four minutes |
---|
0:10:59 | and then they began face-to-face to did not display asr is just the query and |
---|
0:11:03 | it just was always there are showing always updating |
---|
0:11:07 | and the endpointing as i mentioned was done semantically |
---|
0:11:11 | s two and determine there was a question and we just asked them you know |
---|
0:11:14 | what you think about |
---|
0:11:16 | these different systems so there was a ten questions and we ask some you know |
---|
0:11:20 | that they prefer the first system the second system either or both |
---|
0:11:24 | and case three started this was the adaptability adaptive phase |
---|
0:11:29 | which is basically the same as face to with adaptation and the wayward is that's |
---|
0:11:33 | very simple way |
---|
0:11:35 | if base |
---|
0:11:36 | if they did it task |
---|
0:11:38 | basically build a slot or frame |
---|
0:11:41 | and they |
---|
0:11:42 | did that same thing again it will remember it and start to |
---|
0:11:45 | ask them just immediately ask a clarification so instead of saying i want this i |
---|
0:11:50 | want the thai food they would say i'm hungry and then it would say then |
---|
0:11:53 | it just have to say yes and it was shown slots for them |
---|
0:11:56 | and then after three times we just filling all the frame entirely for |
---|
0:12:00 | and also an example of that much for video card movement |
---|
0:12:03 | and then after face three we had another questionnaire that compared phases two three |
---|
0:12:08 | so here's that video |
---|
0:12:10 | so this is in german i'm doing this |
---|
0:12:13 | so if you speak your mind you apologise from my accent and so anyway so |
---|
0:12:18 | i'm saying something like this i'm hungry us i want to eat something around here |
---|
0:12:22 | maybe thai food |
---|
0:12:23 | and it does a clarifications are to say exactly |
---|
0:12:26 | and then i repeat this several times to show you the adaptability of this |
---|
0:12:30 | this isn't something you would do you're not gonna take your personal assistant read be |
---|
0:12:34 | yourself five times |
---|
0:12:36 | it's gonna give us a lot |
---|
0:12:38 | but just to show the functionality of this |
---|
0:12:45 | stress |
---|
0:12:49 | are |
---|
0:12:52 | i |
---|
0:12:54 | so |
---|
0:12:56 | it's filter not just one more kiss |
---|
0:13:00 | we are hungry and now it's also |
---|
0:13:03 | i feel like |
---|
0:13:05 | and i don't see that same thing i am hungry |
---|
0:13:09 | so |
---|
0:13:16 | and then the last time i said calmly |
---|
0:13:19 | if someone else |
---|
0:13:25 | i'm a pretty |
---|
0:13:27 | pretty easy going to predict yes but this is common |
---|
0:13:30 | it will use their people want to use these personal assistance data the same thing |
---|
0:13:33 | over and over again |
---|
0:13:35 | my brother here's an act my brother everyday twice a day all opens up as |
---|
0:13:40 | i phone subspace yuri |
---|
0:13:42 | google voice you traffic |
---|
0:13:44 | every day |
---|
0:13:45 | is it just like that and it gets the response he once in people do |
---|
0:13:49 | this and it could probably just pop up and shown the traffic |
---|
0:13:54 | where am here |
---|
0:13:55 | so we got fourteen participants to come and sit down with our system so we |
---|
0:14:00 | set them data at a table there is a |
---|
0:14:02 | a screen that show the task that they were to do not spend a moment |
---|
0:14:05 | and then there is a chat with it was a turn on its side it |
---|
0:14:08 | shows the gui and the gooey was this was as i showed you and it's |
---|
0:14:13 | it's javascript so it was in a in a web browser basically a motel what |
---|
0:14:16 | and then as a keyboard push a button to let them know that they couldn't |
---|
0:14:19 | one |
---|
0:14:21 | but to signal about that the task was complete rather so the tasks were like |
---|
0:14:26 | this there are five possible tasks call reminder |
---|
0:14:29 | find a restaurant leave a message or find a route between two cities |
---|
0:14:34 | and that asks questions icons and the task items were randomly chosen randomly chosen task |
---|
0:14:39 | randomly chose the slot so we want them to convey to the system and then |
---|
0:14:43 | there is a fifty percent chance later that the task would be repeated |
---|
0:14:48 | here's an example |
---|
0:14:49 | they were said they'd be sitting down playing with this the system and then something |
---|
0:14:53 | like this would pop up on the screen and that thousand or call |
---|
0:14:56 | peter |
---|
0:14:57 | and the system with then |
---|
0:15:00 | due to its magic then show |
---|
0:15:03 | google really show it's gooey and once they |
---|
0:15:06 | recognise that understood then they would push a button and a new task pop up |
---|
0:15:12 | and they were charged with doing so many of these task as possible |
---|
0:15:15 | because the we wanted to do this |
---|
0:15:19 | and not just let him play with it because the tasks |
---|
0:15:22 | help us |
---|
0:15:25 | collect some objective measures as well if we tell them we want them to do |
---|
0:15:28 | is many tasks as possible in the four minutes of to have to interact with |
---|
0:15:32 | each setting of the system then we can learn a little bit more about how |
---|
0:15:35 | productive they work |
---|
0:15:37 | so here's the other tasks they would see stuff like this |
---|
0:15:39 | so we have the twenty most common german names you know how to most published |
---|
0:15:43 | cities in germany billfold it turns out as among them |
---|
0:15:48 | and you know everything else part of the so there's quite a few possibilities that |
---|
0:15:52 | could be said here |
---|
0:15:54 | but again |
---|
0:15:55 | we didn't train this at all we just sort of type these and got a |
---|
0:15:58 | list of stuff and threw it into to the system important that was the end |
---|
0:16:01 | of it and then worked |
---|
0:16:05 | but here some results from the questionnaire as we get we can we can conclude |
---|
0:16:09 | the following based on sums some significance courses that they generally like the gucci |
---|
0:16:15 | they counterintuitive to use an easy and understandable |
---|
0:16:18 | and that was our main focus now something goal |
---|
0:16:22 | the grill optimistic to be taken care of locally and they did this a lot |
---|
0:16:26 | if a mistake if the if of slot was filled with the wrong thing they |
---|
0:16:29 | would immediately try to fix it |
---|
0:16:31 | it didn't always just push a button move on to the next task or |
---|
0:16:34 | there is a keyword they could say that could we start from the beginning they |
---|
0:16:37 | generally trying to fix it right there and it was able to do it for |
---|
0:16:40 | the most the time |
---|
0:16:42 | and they didn't generally notice that the between face to face three the incremental and |
---|
0:16:47 | adaptive phase they didn't really know there's |
---|
0:16:48 | something adapting but for those who did not which was about half of them they |
---|
0:16:52 | notice that was face three nineveh did get wrong and there's a listing of all |
---|
0:16:56 | the questions and there's more in the in the results section of the paper on |
---|
0:16:58 | this because of the |
---|
0:17:00 | this is what some things we want to highlight from that |
---|
0:17:04 | so |
---|
0:17:05 | the objective results we are these tell in interesting story so we just cut we |
---|
0:17:10 | just kinda that the number of tasks of their able to do in the different |
---|
0:17:14 | settings |
---|
0:17:15 | and once they get increments one adaptive variable to do quite a few more tasks |
---|
0:17:19 | at least they thought the tasks were complete |
---|
0:17:22 | and here the next the next rows frame accuracies so when all the slots in |
---|
0:17:26 | the framework the same as the one that we wanted them to convey in the |
---|
0:17:30 | task that we showed |
---|
0:17:32 | and the adaptive wanna |
---|
0:17:33 | does quite well because basis it's part of the time the slots are already field |
---|
0:17:38 | for them |
---|
0:17:39 | so it score one for google now |
---|
0:17:41 | i guess trying to predict your life is actually maybe easier than learning how to |
---|
0:17:45 | understand language |
---|
0:17:48 | the other to tell an interesting the more interesting story we get f-score which is |
---|
0:17:52 | basically maybe the entire frame wasn't correct but the this gives a and idea of |
---|
0:17:58 | the correctness of the slots of the frame maybe wanted to the slots were correct |
---|
0:18:01 | one wasn't |
---|
0:18:03 | and |
---|
0:18:04 | in this case incremental lower and then look at the time the time is about |
---|
0:18:08 | the same across all and this tells us that the degree was |
---|
0:18:12 | intuitive enough that in the in the printed |
---|
0:18:15 | phase where they are just playing with it in the trial phase |
---|
0:18:19 | they learn enough about an experience enough that they are just getting used to it |
---|
0:18:22 | over time |
---|
0:18:26 | and |
---|
0:18:28 | what both these rules tell kind of that story |
---|
0:18:31 | so it helps to be a little bit more productive especially in the adaptive the |
---|
0:18:34 | adaptive |
---|
0:18:36 | ending |
---|
0:18:37 | so they're kinda nice results not the most stellar thing this thing is and you |
---|
0:18:42 | know going to be in everyone's phone next month |
---|
0:18:46 | but |
---|
0:18:47 | like i said we didn't use any training data and it was fairly robust |
---|
0:18:55 | some discussion here |
---|
0:18:57 | our incremental personal assistant or ip a different i suppose allow users to make mistakes |
---|
0:19:02 | easier and sooner allow the users to interpret the state of the system's understanding |
---|
0:19:08 | and under the adaptive settings it allows users to be more productive you get more |
---|
0:19:12 | tasks done in this kind of the setting where we're driving them to do tasks |
---|
0:19:16 | like this |
---|
0:19:17 | and endpointed based on semantics not based on site |
---|
0:19:20 | i have a nice thing |
---|
0:19:23 | future work |
---|
0:19:27 | i mandarin is the obvious thing we have a system no training data let's interact |
---|
0:19:31 | with it and it should start to learn and do things better |
---|
0:19:34 | and the mechanisms of their siam the nlu model we have the dialogue manager we |
---|
0:19:39 | have all have provisions for this we just need some kind of a supervision signal |
---|
0:19:42 | which we have if the frames filament get sent on their happy with that |
---|
0:19:46 | we can give feedback now to say those utterances led to this then that should |
---|
0:19:50 | that should help the nlu and hope that the dialogue manager work better |
---|
0:19:53 | same for additive |
---|
0:19:55 | and better use user modelling and adaptability |
---|
0:19:58 | like to be improved |
---|
0:20:00 | also web based authoring loose does this a lot of systems other that do this |
---|
0:20:04 | right now it's not too bad you can after adjacent file and it'll important there's |
---|
0:20:08 | tools for that and is actually fairly quick and easy but where they softly might |
---|
0:20:11 | be nice and then of course we need to scale up to more |
---|
0:20:15 | larger domains degrees the bottleneck here and it's sort of a two edged sword you |
---|
0:20:18 | wanna show your stuff but also be able to handle lots and lots of general |
---|
0:20:23 | things so |
---|
0:20:25 | that is it thank you |
---|
0:20:33 | note that focus on |
---|
0:20:37 | if the |
---|
0:20:51 | right |
---|
0:20:58 | right like a like i said we're not ui |
---|
0:21:02 | experts bring us to if you're right it's gives call i guess on but what |
---|
0:21:05 | we have right now is sort of a max after their seven or eight knows |
---|
0:21:09 | that is just sort of dot the thing you have to do there is there |
---|
0:21:13 | and what gets shown what are the top seven that you will show and if |
---|
0:21:17 | those are if there's something that's not english on their then you doing something wrong |
---|
0:21:21 | so there's more user modelling that happens in that regard what get shown on the |
---|
0:21:24 | gui |
---|
0:21:26 | better no you would help with that |
---|
0:21:29 | better user model and help with the |
---|
0:21:31 | good question |
---|
0:21:33 | research future stuff |
---|
0:21:35 | i q |
---|
0:21:47 | right that i'm not that the future work i mean the way we don't the |
---|
0:21:51 | provisions are there are also in this you can you can click on of the |
---|
0:21:54 | clicking doesn't do anything about the idea is kind of like the stuff on larson |
---|
0:21:57 | to his gui as you can talk about the gui itself and navigate to go |
---|
0:22:01 | insane know why don't want any of those go down a little bit we start |
---|
0:22:04 | right there are some exactly |
---|
0:22:06 | exactly so you can flip through it put stuff and you can add something if |
---|
0:22:10 | it's not there that would be nice to and i guess but right and system |
---|
0:22:12 | in as becomes intent that you can use in the future the gui should be |
---|
0:22:15 | able to help with that |
---|
0:22:18 | okay |
---|
0:22:38 | right so it |
---|
0:22:42 | right so the common question comment was on the semantic endpointing bit of it i |
---|
0:22:49 | something to look at i don't have |
---|
0:22:52 | don't have an answer |
---|
0:22:54 | definitely something considering |
---|
0:23:05 | right |
---|
0:23:06 | agree |
---|
0:23:19 | no not |
---|
0:23:20 | i want to be really clear on that they're in the trial phase maybe they've |
---|
0:23:24 | done all the adapting they're done adapting but the system is so rudimentary and simple |
---|
0:23:30 | and the gui is that it doesn't it doesn't do much you know there's only |
---|
0:23:34 | a couple of things that it that it does they learn about a very quickly |
---|
0:23:38 | that's why that time to really change |
---|
0:23:40 | you know the average time per |
---|
0:23:43 | for task |
---|
0:23:45 | so they weren't just |
---|
0:23:46 | getting used to it over time because they are already used to before they even |
---|
0:23:49 | started the first phase that's kind of the taken thing i got from the objective |
---|
0:23:53 | scores |
---|
0:23:55 | that's something we were concerned with that's why we designed it this way |
---|
0:23:59 | that was i need i knew someone asks a question i'm glad somebody did exactly |
---|
0:24:03 | we because of the way we wanted to do the comparisons we wanted to do |
---|
0:24:08 | this objective comparisons and we wanted to do some objective scores and this was a |
---|
0:24:12 | debate we had what we ended up doing it this way with the hope that |
---|
0:24:14 | if we designed the right way |
---|
0:24:16 | you don't get used to write beginning we will have as facts and the numbers |
---|
0:24:20 | can show that |
---|
0:24:22 | i'm glad you ask that |
---|