0:00:14 | but i wanted to presented to a continuum of what nickel are presented them |
---|
0:00:22 | are you |
---|
0:00:22 | about |
---|
0:00:24 | we created a an inquisition game |
---|
0:00:28 | that abstracted completely the natural language out from the game so basically we consider that |
---|
0:00:34 | the negotiation |
---|
0:00:36 | it's |
---|
0:00:38 | it's finding an agreement between two people |
---|
0:00:45 | that's not my |
---|
0:00:48 | it's fa an agreement between two people and |
---|
0:00:52 | and so both there are a number of options possible agreement between them |
---|
0:01:00 | both people have different preferences overall this set of possible agreements |
---|
0:01:06 | and they have to exchange in order to find the best possible way |
---|
0:01:10 | and what we found out |
---|
0:01:13 | so you know very simply for the you know very simplified way of |
---|
0:01:19 | of communicating the different options is that supposed to simple and when we put two |
---|
0:01:26 | humans |
---|
0:01:30 | trying to find an agreement together |
---|
0:01:35 | maybe they didn't the where some of them where more easygoing |
---|
0:01:41 | but in the n the best strategy for the system was quite the same |
---|
0:01:47 | so the continue the |
---|
0:01:51 | the well what we did what we're trying to do now is to of more |
---|
0:01:56 | complex options that are |
---|
0:01:58 | that's not the combination of several features so for instance issue of trying to set |
---|
0:02:04 | an appointment you have to specify the date you have to specify the power |
---|
0:02:10 | or if you want to exchange fruits like in the bargaining task |
---|
0:02:15 | well you have to define how many apple i mean the oranges you want and |
---|
0:02:21 | stuff like that so these are different features and you can have a much complex |
---|
0:02:27 | a set of actions |
---|
0:02:29 | that a i want the apples |
---|
0:02:32 | or a |
---|
0:02:35 | we can meet in the at any time on a on thursday |
---|
0:02:39 | stuff like this and it |
---|
0:02:40 | the number of action then explodes and it would be much more in an interesting |
---|
0:02:44 | to work on this |
---|
0:03:02 | but i actually need |
---|
0:03:04 | i other than that huh show and the unit can try to |
---|
0:03:12 | i mean they can start like to tease to show the example |
---|
0:03:30 | i know it was the presentation and the k |
---|
0:03:40 | so we start the |
---|
0:03:45 | so |
---|
0:03:48 | i think that everybody here agrees that major challenge for the automatic analysis them negotiate |
---|
0:03:54 | negotiation dialogs |
---|
0:03:56 | is that a like modeling be disagreement space which is shaped it different participants in |
---|
0:04:02 | you process of arguing |
---|
0:04:04 | and actually at least it's of the art there like studied in argumentation mining that |
---|
0:04:08 | focuses on due to medic identification of like james and primacy is and all sounded |
---|
0:04:13 | types of relation linking them like that supports |
---|
0:04:17 | agreement or disagreement |
---|
0:04:19 | our have are the current methodologies do not find to what aspects easy to the |
---|
0:04:26 | treatment as scope over |
---|
0:04:29 | and we believe that this is actually important in order to predict registration strategies and |
---|
0:04:35 | also to understand like specific controversy in different contexts |
---|
0:04:42 | so therefore our research question is like how can we model these scope of disagreements |
---|
0:04:47 | you know comment that the context |
---|
0:04:49 | and on these grounds we proposed a to a level ontology an upper lever and |
---|
0:04:55 | the lower level ontology |
---|
0:04:57 | in order to model these agreement space so let's like take as example |
---|
0:05:04 | a discussion around like taken from change maybe you the a subgradient |
---|
0:05:09 | and so as you can see details all of the original post use like diversity |
---|
0:05:15 | is not about race |
---|
0:05:18 | and uncertainty common starts to be at i versus societies the society which have people |
---|
0:05:24 | from different backgrounds and cultures how does raise scamming to play |
---|
0:05:29 | and actually these two sentences are called out or this is how we call nh |
---|
0:05:36 | comments in our like upper level ontology by one random participant that is called like |
---|
0:05:43 | d in the skin |
---|
0:05:45 | that actually challenges like the assertion underlying the rhetorical question we do not a rhetorical |
---|
0:05:52 | question so you're joking right and you can do you know as we descended like |
---|
0:05:56 | talk to some black falls |
---|
0:05:59 | so it's clears the according to our ontology did these first to send its use |
---|
0:06:03 | in like orange in the original forest |
---|
0:06:06 | are a target |
---|
0:06:07 | and the comment is like a whole lot |
---|
0:06:10 | and so it is clear that the relation is that these agreements relation but actually |
---|
0:06:15 | you can i think an e like weekly understand what is challenge is not really |
---|
0:06:22 | like the each row of the statement it is the fact that it's the each |
---|
0:06:26 | roles of the person with like expressing statement |
---|
0:06:29 | so basically what is challenge is not be proposition but speech act so one pretty |
---|
0:06:35 | easy to use conditional like making an argument is that of like having your right |
---|
0:06:39 | to do so |
---|
0:06:41 | and so actually the user is claiming that like the speaker is biased |
---|
0:06:47 | then you original poster goes on and you like you provides an example take these |
---|
0:06:51 | example a white child we immigrated from change are yet this depressed persecution and the |
---|
0:06:56 | black and make an child even next door to each other |
---|
0:06:59 | the african american child was had another way someone childhood is accepted into college for |
---|
0:07:05 | this take a bigger city one channel from chain child that is rejected |
---|
0:07:10 | so in this case there is an adder user id that like calls out this |
---|
0:07:14 | time like the challenge is you know corrine somebody rejection event |
---|
0:07:19 | and it is like well it's hard to say which key it would have a |
---|
0:07:23 | better shot at getting into the same competitive school and then you can tune you |
---|
0:07:27 | like expressing like a these agreement again but towards like the last sentence of the |
---|
0:07:33 | original what was so that prosody should be about the result of experience in background |
---|
0:07:37 | not skin |
---|
0:07:39 | so in this case what is challenge is not the events but actually it's really |
---|
0:07:44 | like the you truth of deeper position |
---|
0:07:47 | another type of these agreement is they one expressed by be all actually like in |
---|
0:07:52 | is |
---|
0:07:54 | commenting on d meaning of the verb novelty word by diversity |
---|
0:07:59 | which is also part of the last a statement which actually in like different linguistic |
---|
0:08:05 | theories is considered like an entity soft first order and d |
---|
0:08:10 | so there are also other types of relation you can also have like agreement relations |
---|
0:08:15 | for example like the original poster answers to user c and this is like no |
---|
0:08:20 | i agree |
---|
0:08:22 | that one what it was like trying to say or a to relations type that |
---|
0:08:26 | are hard to classify |
---|
0:08:29 | for example like when the what we jump was to a of words i delta |
---|
0:08:35 | so you e actually that they like |
---|
0:08:37 | you have or stating |
---|
0:08:40 | so |
---|
0:08:42 | yes i just want to finish we award punch lines our pension is that these |
---|
0:08:48 | ontologies readable to leverage outer existing semantic and pragmatic layers of annotation |
---|
0:08:54 | and then reach the information the information that they provide and for example because relation |
---|
0:08:59 | between like these agreement relations and you types of targets that they select |
---|
0:09:05 | promises to improve the detection of like about those which are like a type of |
---|
0:09:09 | disagreement |
---|
0:09:12 | and so then we then we like these are like a more points of discussion |
---|
0:09:16 | that basically muir those that have already been addressed by dependencies |
---|
0:09:23 | thank you |
---|
0:09:46 | so i'm happy to announce it just happens like two weeks ago via released out |
---|
0:09:52 | the corpus and you corpus met the log multi-issue bargaining corpus in ldc catalog so |
---|
0:09:58 | it's |
---|
0:10:00 | i presented for all for our group in here in several on and that colleagues |
---|
0:10:05 | in grounding and cool build the cognitive model for the corpus for the corpus collection |
---|
0:10:10 | and for the future system and i will be present the system to model |
---|
0:10:15 | so what the corpus is about as a scenario is multi-issue bargaining so it's not |
---|
0:10:20 | just negotiation to buy banana or oranges is i issues based preferences involve therefore issuers |
---|
0:10:29 | it's integrated negotiation as win been situation |
---|
0:10:34 | it's featuring actually negotiation the value in my it's a complex negotiation strategic negotiation the |
---|
0:10:41 | domain it was real scenario took |
---|
0:10:46 | in a at the necessity of buttons past the anti smoking legislation also this each |
---|
0:10:53 | year of new york could force the debate on the on this one and it |
---|
0:10:58 | was not very efficient so they me to come immediate many by just need to |
---|
0:11:02 | come together and everything negotiate |
---|
0:11:04 | why is not walking a half adjustments so basically the corpus is collected to be |
---|
0:11:10 | is that a negotiation train there so dct council needs to train a to be |
---|
0:11:18 | trained to negotiate beginnings different body sit giddings business represent the thief |
---|
0:11:24 | against police officers against house insurance et cetera et cetera so the preferences brag even |
---|
0:11:31 | for them raise up references in the in a sense that the right couple of |
---|
0:11:39 | scenarios designs and because |
---|
0:11:43 | it was not real politicians involve by the out you again parliamentarians so that where |
---|
0:11:49 | got a preferences and they need to defend their positions no time constraints and they |
---|
0:11:55 | were instructed to a weight negotiate a negative agreement i will explain later tomorrow so |
---|
0:12:01 | the basically |
---|
0:12:03 | cannot be all they we're at college not to accept |
---|
0:12:09 | this preferred |
---|
0:12:11 | this preferred options |
---|
0:12:13 | so the we will release for parts of this corpus now we release human dialogues |
---|
0:12:20 | for to have a was eight subjects |
---|
0:12:24 | two thousand turns ten thousand tokens bill the release they next part more which is |
---|
0:12:31 | more argument that eve we had they need to defend their position we will release |
---|
0:12:36 | debate corpus on the same topic and we will release the evaluation corpus which is |
---|
0:12:41 | larger human |
---|
0:12:43 | machine dialogue or human machine dialogues |
---|
0:12:48 | corpus asr recordings and transcriptions and you can use these to retrain the speech recognition |
---|
0:12:54 | is obviously too small but you can use it for adaptations of the speaker diarisation |
---|
0:12:59 | is done manually correctly everything a high quality |
---|
0:13:06 | into format in many format's also dear and also transcriptions in t i is include |
---|
0:13:13 | eats |
---|
0:13:15 | down we have dialogue annotations or semantic and pragmatic annotations |
---|
0:13:21 | about nine cells and to vent that this out annotated this types six dives up |
---|
0:13:26 | before the dialogue acts discussed structuring acts rhetorical relations according to the newest eyes are |
---|
0:13:33 | stand that |
---|
0:13:36 | that's it and this is you your l where you can download the |
---|
0:13:40 | corpus if you have membership |
---|