[Script Info]
Title: 
[Events]
Format: Layer, Start, End, Style, Name, MarginL, MarginR, MarginV, Effect, Text
Dialogue: 0,0:00:00.00,0:00:19.48,Default,,0000,0000,0000,,{\i1}36C3 preroll music{\i0}
Dialogue: 0,0:00:19.48,0:00:24.14,Default,,0000,0000,0000,,Herald Angel: We have Tom and Max here.\NThey have a talk here with a very
Dialogue: 0,0:00:24.14,0:00:28.14,Default,,0000,0000,0000,,complicated title that I don't quite\Nunderstand yet. It's called "Interactively
Dialogue: 0,0:00:28.14,0:00:35.81,Default,,0000,0000,0000,,Discovering Implicational Knowledge in\NWikidata. And they told me the point of
Dialogue: 0,0:00:35.81,0:00:39.19,Default,,0000,0000,0000,,the talk is that I would like to\Nunderstand what it means and I hope I
Dialogue: 0,0:00:39.19,0:00:42.19,Default,,0000,0000,0000,,will. So good luck.\NTom: Thank you very much.
Dialogue: 0,0:00:42.19,0:00:44.31,Default,,0000,0000,0000,,Herald: And have some applause, please.
Dialogue: 0,0:00:44.31,0:00:47.88,Default,,0000,0000,0000,,{\i1}applause{\i0}
Dialogue: 0,0:00:47.88,0:00:54.98,Default,,0000,0000,0000,,T: Thank you very much. Do you hear me?\NDoes it work? Hello? Oh, very good. Thank
Dialogue: 0,0:00:54.98,0:00:58.79,Default,,0000,0000,0000,,you very much and welcome to our talk\Nabout interactively discovering
Dialogue: 0,0:00:58.79,0:01:05.11,Default,,0000,0000,0000,,implicational knowledge in Wikidata. It\Nis more or less a fun project we started
Dialogue: 0,0:01:05.11,0:01:10.89,Default,,0000,0000,0000,,for finding rules that are implicit in\NWikidata – entailed just by the data it
Dialogue: 0,0:01:10.89,0:01:18.85,Default,,0000,0000,0000,,has, that people inserted into the\NWikidata database so far. And we will
Dialogue: 0,0:01:18.85,0:01:23.57,Default,,0000,0000,0000,,start with the explicit knowledge. So the\Nexplicit data in Wikidata, with Max.
Dialogue: 0,0:01:23.57,0:01:28.34,Default,,0000,0000,0000,,Max: So. Right. What what is Wikidata?\NMaybe you have heard about Wikidata, then
Dialogue: 0,0:01:28.34,0:01:33.21,Default,,0000,0000,0000,,that's all fine. Maybe you haven't, then\Nsurely you've heard of Wikipedia. And
Dialogue: 0,0:01:33.21,0:01:36.79,Default,,0000,0000,0000,,Wikipedia is run by the Wikimedia\NFoundation and the Wikimedia Foundation
Dialogue: 0,0:01:36.79,0:01:41.33,Default,,0000,0000,0000,,has several other projects. And one of\Nthose is Wikidata. And Wikidata is
Dialogue: 0,0:01:41.33,0:01:45.49,Default,,0000,0000,0000,,basically a large graph that encodes\Nmachine readable knowledge in the form of
Dialogue: 0,0:01:45.49,0:01:51.73,Default,,0000,0000,0000,,statements. And a statement basically\Nconsists of some entity that is connected
Dialogue: 0,0:01:51.73,0:01:58.20,Default,,0000,0000,0000,,– or some some entities that are connected\Nby some property. And these properties
Dialogue: 0,0:01:58.20,0:02:02.91,Default,,0000,0000,0000,,can then even have annotations on them.\NSo, for example, we have Donna Strickland
Dialogue: 0,0:02:02.91,0:02:09.15,Default,,0000,0000,0000,,here and we encode that she has received a\NNobel prize in physics last year by this
Dialogue: 0,0:02:09.15,0:02:16.29,Default,,0000,0000,0000,,property "awarded" and this has then a\Nqualifier "time: 2018" and also "for:
Dialogue: 0,0:02:16.29,0:02:23.10,Default,,0000,0000,0000,,Chirped Pulse Amplification". And all in\Nall, we have some 890 million statements
Dialogue: 0,0:02:23.10,0:02:31.96,Default,,0000,0000,0000,,on Wikidata that connect 71 million items\Nusing 7000 properties. But there's also a
Dialogue: 0,0:02:31.96,0:02:36.83,Default,,0000,0000,0000,,bit more. So we also know that Donna\NStrickland has "field of work: optics" and
Dialogue: 0,0:02:36.83,0:02:41.42,Default,,0000,0000,0000,,also "field of work: lasers" so we can use\Nthe same property to connect some entity
Dialogue: 0,0:02:41.42,0:02:46.48,Default,,0000,0000,0000,,with different other entities. And we\Ndon't even have to have knowledge that
Dialogue: 0,0:02:46.48,0:02:56.53,Default,,0000,0000,0000,,connects the entities. We can have a date\Nof birth, which is 1959. Nineteen ninety.
Dialogue: 0,0:02:56.53,0:03:05.53,Default,,0000,0000,0000,,No. Nineteen fifty nine. Yes. And this is\Nthen just a plain date, not an entity. And
Dialogue: 0,0:03:05.53,0:03:11.51,Default,,0000,0000,0000,,now coming from the explicit knowledge\Nthen, well, we have some more we have
Dialogue: 0,0:03:11.51,0:03:16.21,Default,,0000,0000,0000,,Donna Strickland has received a Nobel\Nprize in physics and also Marie Curie has
Dialogue: 0,0:03:16.21,0:03:21.17,Default,,0000,0000,0000,,received the Nobel prize in physics. And\Nwe also know that Marie Curie has a Nobel
Dialogue: 0,0:03:21.17,0:03:27.78,Default,,0000,0000,0000,,prize ID that starts with "phys" and then\N"1903" and some random numbers that
Dialogue: 0,0:03:27.78,0:03:32.97,Default,,0000,0000,0000,,basically are this ID. Then Marie Curie\Nalso has received a Nobel prize in
Dialogue: 0,0:03:32.97,0:03:38.58,Default,,0000,0000,0000,,chemistry in 1911. So she has another\NNobel ID that starts with "chem" and has
Dialogue: 0,0:03:38.58,0:03:43.59,Default,,0000,0000,0000,,"1911" there. And then there's also\NFrances Arnold, who received the Nobel
Dialogue: 0,0:03:43.59,0:03:48.55,Default,,0000,0000,0000,,prize in chemistry last year. So she has a\NNobel ID that starts with "chem" and has
Dialogue: 0,0:03:48.55,0:03:54.74,Default,,0000,0000,0000,,"2018" there. And now one one could assume\Nthat, well, everybody who was awarded the
Dialogue: 0,0:03:54.74,0:04:00.16,Default,,0000,0000,0000,,Nobel prize should also have a Nobel ID.\NSo everybody who was awarded the Nobel
Dialogue: 0,0:04:00.16,0:04:05.67,Default,,0000,0000,0000,,prize should also have a Nobel prize ID,\Nand we could write that as some
Dialogue: 0,0:04:05.67,0:04:11.79,Default,,0000,0000,0000,,implication here. So "awarded(nobelPrize)"\Nimplies "nobelID". And well, if you
Dialogue: 0,0:04:11.79,0:04:16.35,Default,,0000,0000,0000,,look sharply at this picture, then there's\Nthis arrow here conspicuously missing that
Dialogue: 0,0:04:16.35,0:04:22.55,Default,,0000,0000,0000,,Donald Strickland doesn't have a Nobel\Nprize ID. And indeed, there's 25 people
Dialogue: 0,0:04:22.55,0:04:26.67,Default,,0000,0000,0000,,currently on Wikidata that are missing\NNobel prize IDs, and Donna Strickland is
Dialogue: 0,0:04:26.67,0:04:34.06,Default,,0000,0000,0000,,one of them. So we call these people that\Ndon't satisfy this implication – we call
Dialogue: 0,0:04:34.06,0:04:40.42,Default,,0000,0000,0000,,those counterexamples and well, if you\Nlook at Wikidata on the scale of really
Dialogue: 0,0:04:40.42,0:04:45.35,Default,,0000,0000,0000,,these 890 million statements, then you\Nwon't find any counterexamples because
Dialogue: 0,0:04:45.35,0:04:52.55,Default,,0000,0000,0000,,it's just too big. So we need some way to\Nautomatically do that. And the idea is
Dialogue: 0,0:04:52.55,0:04:58.93,Default,,0000,0000,0000,,that, well, if we had this knowledge that\Nwhile some implications are not satisfied,
Dialogue: 0,0:04:58.93,0:05:03.84,Default,,0000,0000,0000,,then this encodes maybe missing\Ninformation or wrong information, and we
Dialogue: 0,0:05:03.84,0:05:10.87,Default,,0000,0000,0000,,want to represent that in a way that is\Neasy to understand and also succinct. So
Dialogue: 0,0:05:10.87,0:05:16.09,Default,,0000,0000,0000,,it doesn't take long to write it down, it\Nshould have a short representation. So
Dialogue: 0,0:05:16.09,0:05:23.06,Default,,0000,0000,0000,,that rules out anything, including complex\Nsyntax or logical quantifies. So no SPARQL
Dialogue: 0,0:05:23.06,0:05:27.48,Default,,0000,0000,0000,,queries as a description of that implicit\Nknowledge. No description logics, if
Dialogue: 0,0:05:27.48,0:05:33.20,Default,,0000,0000,0000,,you've heard of that. And we also want\Nsomething that we can actually compute on
Dialogue: 0,0:05:33.20,0:05:41.54,Default,,0000,0000,0000,,actual hardware in a reasonable timeframe.\NSo our approach is we use Formal Concept
Dialogue: 0,0:05:41.54,0:05:46.89,Default,,0000,0000,0000,,Analysis, which is a technique that has\Nbeen developed over the past several years
Dialogue: 0,0:05:46.89,0:05:52.07,Default,,0000,0000,0000,,to extract what is called propositional\Nimplications. So just logical formulas of
Dialogue: 0,0:05:52.07,0:05:56.24,Default,,0000,0000,0000,,propositional logic that are an\Nimplication in the form of this
Dialogue: 0,0:05:56.24,0:06:03.02,Default,,0000,0000,0000,,"awarded(nobelPrize)" implies "nobleID".\NSo what exactly is Formal Concept
Dialogue: 0,0:06:03.02,0:06:08.50,Default,,0000,0000,0000,,Analysis? Off to Tom.\NT: Thank you. So what is Formal Concept
Dialogue: 0,0:06:08.50,0:06:14.42,Default,,0000,0000,0000,,Analysis? It was developed in 1980s by a\Nguy called Rudolf Wille and Bernard Ganter
Dialogue: 0,0:06:14.42,0:06:18.54,Default,,0000,0000,0000,,and they were restructuring lattice\Ntheory. Lattice theory is an ambiguous
Dialogue: 0,0:06:18.54,0:06:23.37,Default,,0000,0000,0000,,name in math, it has two meanings: One\Nmeaning is you have a grid and have a
Dialogue: 0,0:06:23.37,0:06:29.05,Default,,0000,0000,0000,,lattice there. The other thing is to speak\Nabout orders – order relations. So I like
Dialogue: 0,0:06:29.05,0:06:34.15,Default,,0000,0000,0000,,steaks, I like pudding and I like steaks\Nmore than pudding. And I like rice more
Dialogue: 0,0:06:34.15,0:06:40.96,Default,,0000,0000,0000,,than steaks. That's an order, right? And\Nlattices are particular orders which can
Dialogue: 0,0:06:40.96,0:06:46.77,Default,,0000,0000,0000,,be used to represent propositional logic.\NSo easy rules like "when it rains, the
Dialogue: 0,0:06:46.77,0:06:52.99,Default,,0000,0000,0000,,street gets wet", right? So and the data\Nrepresentation those guys used back then,
Dialogue: 0,0:06:52.99,0:06:57.08,Default,,0000,0000,0000,,they called it a formal context, which is\Nbasically just a set of objects – they
Dialogue: 0,0:06:57.08,0:07:02.00,Default,,0000,0000,0000,,call them objects, it's just a name –, a\Nset of attributes and some incidence,
Dialogue: 0,0:07:02.00,0:07:07.89,Default,,0000,0000,0000,,which basically means which object does\Nhave which attributes. So, for example, my
Dialogue: 0,0:07:07.89,0:07:13.15,Default,,0000,0000,0000,,laptop has the colour black. So this\Nobject has some property, right? So that's
Dialogue: 0,0:07:13.15,0:07:17.87,Default,,0000,0000,0000,,a small example on the right for such a\Nformal context. So the objects there are
Dialogue: 0,0:07:17.87,0:07:24.38,Default,,0000,0000,0000,,some animals: a platypus – that's the fun\Nanimal from Australia, the mammal which is
Dialogue: 0,0:07:24.38,0:07:30.28,Default,,0000,0000,0000,,also laying eggs and which is also\Nvenomous –, a black widow – the spider –,
Dialogue: 0,0:07:30.28,0:07:35.45,Default,,0000,0000,0000,,the duck and the cat. So we see, the\Nplatypus has all the properties; it has
Dialogue: 0,0:07:35.45,0:07:39.73,Default,,0000,0000,0000,,being venomous, laying eggs and being a\Nmammal; we have the duck, which is not a
Dialogue: 0,0:07:39.73,0:07:44.17,Default,,0000,0000,0000,,mammal, but it lays eggs, and so on and so\Non. And it's very easy to grasp some
Dialogue: 0,0:07:44.17,0:07:49.43,Default,,0000,0000,0000,,implicational knowledge here. An easy rule\Nyou can find is whenever you endeavour a
Dialogue: 0,0:07:49.43,0:07:54.30,Default,,0000,0000,0000,,mammal that is venomous, it has to lay\Neggs. So this is a rule that falls out of
Dialogue: 0,0:07:54.30,0:07:59.64,Default,,0000,0000,0000,,this binary data table. Our main problem\Nthen or at this point is we do not have
Dialogue: 0,0:07:59.64,0:08:03.47,Default,,0000,0000,0000,,such a data table for Wikidata, right? We\Nhave the implicit graph, which is way more
Dialogue: 0,0:08:03.47,0:08:09.03,Default,,0000,0000,0000,,expressive than binary data, and we cannot\Neven store Wikidata as a binary table.
Dialogue: 0,0:08:09.03,0:08:13.86,Default,,0000,0000,0000,,Even if you tried to, we have no chance to\Ncompute such rules from that. And for
Dialogue: 0,0:08:13.86,0:08:21.46,Default,,0000,0000,0000,,this, the people from Formal Context\NAnalysis proposed an algorithm to extract
Dialogue: 0,0:08:21.46,0:08:27.16,Default,,0000,0000,0000,,implicit knowledge from an expert. So our\Nexpert here could be Wikidata. It's an
Dialogue: 0,0:08:27.16,0:08:31.24,Default,,0000,0000,0000,,expert, you can ask Wikidata questions,\Nright? Using this SPARQL interface, you
Dialogue: 0,0:08:31.24,0:08:34.74,Default,,0000,0000,0000,,can ask. You can ask "Is there an example\Nfor that? Is there a counterexample for
Dialogue: 0,0:08:34.74,0:08:39.88,Default,,0000,0000,0000,,something else?" So the algorithm is quite\Neasy. The algorithm is the algorithm and
Dialogue: 0,0:08:39.88,0:08:45.38,Default,,0000,0000,0000,,some expert – in our case, Wikidata –, and\Nthe algorithm keeps notes for
Dialogue: 0,0:08:45.38,0:08:49.45,Default,,0000,0000,0000,,counterexamples and keeps notes for valid\Nimplications. So in the beginning, we do
Dialogue: 0,0:08:49.45,0:08:53.57,Default,,0000,0000,0000,,not have any valid implications, so this\Nlist on the right is empty, and in the
Dialogue: 0,0:08:53.57,0:08:56.78,Default,,0000,0000,0000,,beginning we do not have any\Ncounterexamples. So the list on the left,
Dialogue: 0,0:08:56.78,0:09:01.90,Default,,0000,0000,0000,,the formal context to build up is also\Nempty. And all the algorithm does now is,
Dialogue: 0,0:09:01.90,0:09:09.17,Default,,0000,0000,0000,,it asks "is this implication, X follows Y,\NY follows X or X implies Y, is it true?"
Dialogue: 0,0:09:09.17,0:09:14.00,Default,,0000,0000,0000,,So "is it true," for example, "that an\Nanimal that is a mammal and is venomous
Dialogue: 0,0:09:14.00,0:09:18.88,Default,,0000,0000,0000,,lays eggs?" So now the expert, which in\Nour case is Wikidata, can answer it. We
Dialogue: 0,0:09:18.88,0:09:24.86,Default,,0000,0000,0000,,can query that. We showed in our paper we\Ncan query that. So we query it, and if the
Dialogue: 0,0:09:24.86,0:09:28.49,Default,,0000,0000,0000,,Wikidata expert does not find any\Ncounterexamples, it will say, ok, that's
Dialogue: 0,0:09:28.49,0:09:36.20,Default,,0000,0000,0000,,maybe a true, true thing; it's yes. Or if\Nit's not a true implication in Wikidata,
Dialogue: 0,0:09:36.20,0:09:41.78,Default,,0000,0000,0000,,it can say, no, no, no, it's not true, and\Nhere's a counterexample. So this is
Dialogue: 0,0:09:41.78,0:09:48.51,Default,,0000,0000,0000,,something you contradict by example. You\Nsay this rule cannot be true. For example,
Dialogue: 0,0:09:48.51,0:09:52.90,Default,,0000,0000,0000,,when the street is wet, that does not mean\Nit has rained, right? It could be the
Dialogue: 0,0:09:52.90,0:10:01.38,Default,,0000,0000,0000,,cleaning service car or something else. So\Nour idea now was to use Wikidata as an
Dialogue: 0,0:10:01.38,0:10:05.82,Default,,0000,0000,0000,,expert, but also include a human into this\Nloop. So we do not just want to ask
Dialogue: 0,0:10:05.82,0:10:11.71,Default,,0000,0000,0000,,Wikidata, we also want to ask a human\Nexpert as well. So we first ask in our
Dialogue: 0,0:10:11.71,0:10:18.52,Default,,0000,0000,0000,,tool the Wikidata expert for some rule.\NAfter that, we also inquire the human
Dialogue: 0,0:10:18.52,0:10:22.08,Default,,0000,0000,0000,,expert. And he can also say "yeah, that's\Ntrue, I know that," or "No, no. Wikidata
Dialogue: 0,0:10:22.08,0:10:27.20,Default,,0000,0000,0000,,is not aware of this counterexample, I\Nknow one." Or, in the other case "oh,
Dialogue: 0,0:10:27.20,0:10:32.77,Default,,0000,0000,0000,,Wikidata says this is true. I am aware of\Na counterexample." Yeah, and so on and so
Dialogue: 0,0:10:32.77,0:10:37.60,Default,,0000,0000,0000,,on. And you can represent this more or\Nless – this is just some mathematical
Dialogue: 0,0:10:37.60,0:10:41.69,Default,,0000,0000,0000,,picture, it's not very important. But you\Ncan see on the left there's an exploration
Dialogue: 0,0:10:41.69,0:10:46.72,Default,,0000,0000,0000,,going on, just Wikidata with the\Nalgorithm, on the right an exploration, a
Dialogue: 0,0:10:46.72,0:10:51.42,Default,,0000,0000,0000,,human expert versus Wikidata which can\Nanswer all the queries. And we combined
Dialogue: 0,0:10:51.42,0:10:57.72,Default,,0000,0000,0000,,those two into one small tool, still under\Ndevelopment. So, back to Max.
Dialogue: 0,0:10:57.72,0:11:02.98,Default,,0000,0000,0000,,M: Okay. So far for that to work, we\Nbasically need to have a way of viewing
Dialogue: 0,0:11:02.98,0:11:08.07,Default,,0000,0000,0000,,Wikidata, or at least parts of Wikidata,\Nas a formal context. And this formal
Dialogue: 0,0:11:08.07,0:11:13.61,Default,,0000,0000,0000,,context, well, this was a binary table, so\Nwhat do we do? We just take all the items
Dialogue: 0,0:11:13.61,0:11:18.88,Default,,0000,0000,0000,,in Wikidata as objects and all the\Nproperties as attributes of our context
Dialogue: 0,0:11:18.88,0:11:24.16,Default,,0000,0000,0000,,and then have an incidence relation that\Nsays "well, this entity has this
Dialogue: 0,0:11:24.16,0:11:30.55,Default,,0000,0000,0000,,property," so it is incident there, and\Nthen we end up with a context that has 71
Dialogue: 0,0:11:30.55,0:11:36.43,Default,,0000,0000,0000,,million rows and seven thousand columns.\NSo, well, that might actually be a slight
Dialogue: 0,0:11:36.43,0:11:40.18,Default,,0000,0000,0000,,problem there, because we want to have\Nsomething that we can run on actual
Dialogue: 0,0:11:40.18,0:11:45.81,Default,,0000,0000,0000,,hardware and not on a supercomputer. So\Nlet's maybe not do that and focus on
Dialogue: 0,0:11:45.81,0:11:50.90,Default,,0000,0000,0000,,a smaller set of properties that are\Nactually related to one another through
Dialogue: 0,0:11:50.90,0:11:55.69,Default,,0000,0000,0000,,some kind of common domain, yeah? So it\Ndoesn't make any sense to have a property
Dialogue: 0,0:11:55.69,0:11:59.64,Default,,0000,0000,0000,,that relates to spacecraft and then a\Nproperty that relates to books – that's
Dialogue: 0,0:11:59.64,0:12:05.05,Default,,0000,0000,0000,,probably not a good idea to try to find\Nimplicit knowledge between those two. But
Dialogue: 0,0:12:05.05,0:12:10.26,Default,,0000,0000,0000,,two different properties about spacecraft,\Nthat sounds good, right? And then the
Dialogue: 0,0:12:10.26,0:12:15.00,Default,,0000,0000,0000,,interesting question is just how do we\Ndefine the incidence for our set of
Dialogue: 0,0:12:15.00,0:12:20.15,Default,,0000,0000,0000,,properties? And that actually depends very\Nmuch on which properties we choose,
Dialogue: 0,0:12:20.15,0:12:25.55,Default,,0000,0000,0000,,because it does – for some properties, it\Nmakes sense to account for the direction
Dialogue: 0,0:12:25.55,0:12:32.68,Default,,0000,0000,0000,,of the statement: So there is a property\Ncalled parent? Actually, no, it's child,
Dialogue: 0,0:12:32.68,0:12:38.31,Default,,0000,0000,0000,,and then there's father and mother, and\Nyou don't want to turn those around, as do
Dialogue: 0,0:12:38.31,0:12:43.76,Default,,0000,0000,0000,,you want to have "A is a child of B," that\Nshould be something different than "B
Dialogue: 0,0:12:43.76,0:12:48.93,Default,,0000,0000,0000,,is a child of A." Then there's the\Nqualifiers that might be important for
Dialogue: 0,0:12:48.93,0:12:54.74,Default,,0000,0000,0000,,some properties. So receiving an award for\Nsomething might be something different
Dialogue: 0,0:12:54.74,0:13:00.74,Default,,0000,0000,0000,,than receiving an award for something\Nelse. But while receiving an award in 2018
Dialogue: 0,0:13:00.74,0:13:06.55,Default,,0000,0000,0000,,and receiving one in 2017, that's probably\Nmore or less the same thing, so we don't
Dialogue: 0,0:13:06.55,0:13:11.93,Default,,0000,0000,0000,,necessarily need to differentiate that.\NAnd there's also a thing called subclasses
Dialogue: 0,0:13:11.93,0:13:15.47,Default,,0000,0000,0000,,and they form a hierarchy on Wikidata. And\Nyou might also want to take that into
Dialogue: 0,0:13:15.47,0:13:20.15,Default,,0000,0000,0000,,account because while winning something\Nthat is a Nobel prize, that means also
Dialogue: 0,0:13:20.15,0:13:25.19,Default,,0000,0000,0000,,winning an award itself, and winning the\NNobel Peace prize means winning a peace
Dialogue: 0,0:13:25.19,0:13:32.59,Default,,0000,0000,0000,,prize. So there's also implications going\Non there that you want to respect. So,
Dialogue: 0,0:13:32.59,0:13:38.40,Default,,0000,0000,0000,,to see how we actually do that, let's look\Nat an example. So we have here, well, this
Dialogue: 0,0:13:38.40,0:13:47.03,Default,,0000,0000,0000,,is Donald Strickland. And – I forgot his\Nfirst name – Ashkin, this is one of the
Dialogue: 0,0:13:47.03,0:13:51.72,Default,,0000,0000,0000,,people that won the Nobel prize in physics\Nwith her last year. And also Gérard
Dialogue: 0,0:13:51.72,0:13:57.99,Default,,0000,0000,0000,,Mourou. That is the third one. They all\Ngot the Nobel prize in physics last year.
Dialogue: 0,0:13:57.99,0:14:04.19,Default,,0000,0000,0000,,So we have all these statements here, and\Nthese two have a qualifier that says
Dialogue: 0,0:14:04.19,0:14:10.26,Default,,0000,0000,0000,,"with: Gérard Mourou" here. And I don't\Nthink the qualifier is on this statement
Dialogue: 0,0:14:10.26,0:14:15.16,Default,,0000,0000,0000,,here, actually, but it doesn't actually\Nmatter. So what we've done here is,
Dialogue: 0,0:14:15.16,0:14:21.19,Default,,0000,0000,0000,,put all the entities in the small graph as\Nrows in the table. So we have Strickland
Dialogue: 0,0:14:21.19,0:14:27.85,Default,,0000,0000,0000,,and Mourou and Ashkin, and also Arnold and\NCurie that are not in the picture. But you
Dialogue: 0,0:14:27.85,0:14:33.29,Default,,0000,0000,0000,,can maybe remember that. And then here we\Nhave awarded, and we scaled that by the
Dialogue: 0,0:14:33.29,0:14:37.25,Default,,0000,0000,0000,,instance of the different Nobel prizes\Nthat people have won. So that's the
Dialogue: 0,0:14:37.25,0:14:42.21,Default,,0000,0000,0000,,physics Nobel in the first column, the\Nchemistry Nobel Prize in the second column
Dialogue: 0,0:14:42.21,0:14:48.38,Default,,0000,0000,0000,,and just general Nobel prizes in the third\Ncolumn. There's awarded and that is scaled
Dialogue: 0,0:14:48.38,0:14:55.24,Default,,0000,0000,0000,,by the "with" qualifier, so awarded with\NGérard Mourou. And then there's field of
Dialogue: 0,0:14:55.24,0:15:00.45,Default,,0000,0000,0000,,work, and we have lasers here and\Nradioactivity, so we scale by the actual
Dialogue: 0,0:15:00.45,0:15:06.58,Default,,0000,0000,0000,,field of work that people have. And well\Nthen, if we look at what kind of incidence
Dialogue: 0,0:15:06.58,0:15:11.37,Default,,0000,0000,0000,,we get for Donna Strickland, she has a\NNobel prize in physics and that is also a
Dialogue: 0,0:15:11.37,0:15:17.19,Default,,0000,0000,0000,,Nobel prize, and she has that together\Nwith Mourou. And she has "field of work:
Dialogue: 0,0:15:17.19,0:15:23.22,Default,,0000,0000,0000,,lasers," but not radioactivity. Then,\NMourou himself: he has a Nobel prize in
Dialogue: 0,0:15:23.22,0:15:29.45,Default,,0000,0000,0000,,physics, and that is a Nobel prize, but\Nnone of the others. Ashkin gets the Nobel
Dialogue: 0,0:15:29.45,0:15:33.89,Default,,0000,0000,0000,,prize in physics, and that is still a\NNobel prize, and he gets that with Gérard
Dialogue: 0,0:15:33.89,0:15:40.97,Default,,0000,0000,0000,,Mourou. And also he works on lasers, but\Nnot in radioactivity. So Frances Arnold
Dialogue: 0,0:15:40.97,0:15:47.23,Default,,0000,0000,0000,,has a Nobel prize in chemistry, and that\Nis a Nobel prize. And Marie Curie, she has
Dialogue: 0,0:15:47.23,0:15:50.51,Default,,0000,0000,0000,,a Nobel prize in physics and one in\Nchemistry, and they are both a Nobel
Dialogue: 0,0:15:50.51,0:15:55.32,Default,,0000,0000,0000,,prize. And she also works on\Nradioactivity. But lasers didn't exist
Dialogue: 0,0:15:55.32,0:16:02.49,Default,,0000,0000,0000,,back then, so she doesn't get "field of\Nwork: lasers." And then basically this
Dialogue: 0,0:16:02.49,0:16:10.29,Default,,0000,0000,0000,,table here is a representation of our\Nformal context. So and then we've actually
Dialogue: 0,0:16:10.29,0:16:14.84,Default,,0000,0000,0000,,gone ahead and started building a tool\Nwhere you can interactively do all these
Dialogue: 0,0:16:14.84,0:16:20.32,Default,,0000,0000,0000,,things, and it will take care of building\Nthe context for you. You just put in the
Dialogue: 0,0:16:20.32,0:16:24.54,Default,,0000,0000,0000,,properties, and Tom will show\Nyou how that works.
Dialogue: 0,0:16:24.54,0:16:29.03,Default,,0000,0000,0000,,T: So here you see some first screenshots\Nof this tool. So please do not comment on
Dialogue: 0,0:16:29.03,0:16:32.52,Default,,0000,0000,0000,,the graphic design. We have no idea about\Nthat, we have to ask someone about that.
Dialogue: 0,0:16:32.52,0:16:36.12,Default,,0000,0000,0000,,We're just into logics, more or less. On\Nthe left, you see the initial state of the
Dialogue: 0,0:16:36.12,0:16:41.12,Default,,0000,0000,0000,,game. On the left you have five boxes:\Nthey're called countries and borders,
Dialogue: 0,0:16:41.12,0:16:47.37,Default,,0000,0000,0000,,credit cards, use of energy, memory and\Ncomputation – I think –, and space
Dialogue: 0,0:16:47.37,0:16:53.18,Default,,0000,0000,0000,,launches, which are just presets we\Ndefined. You can explore, for example, in
Dialogue: 0,0:16:53.18,0:16:57.05,Default,,0000,0000,0000,,the case of the credit card, you can\Nexplore the properties from Wikidata which
Dialogue: 0,0:16:57.05,0:17:02.17,Default,,0000,0000,0000,,are called "card network," "operator," and\N"fee," so you can just choose one of them,
Dialogue: 0,0:17:02.17,0:17:05.53,Default,,0000,0000,0000,,or on the right, "custom properties," you\Ncan just input the properties you're
Dialogue: 0,0:17:05.53,0:17:10.64,Default,,0000,0000,0000,,interested in Wikidata, whatever one of\Nthe seven thousand you like, or some
Dialogue: 0,0:17:10.64,0:17:15.14,Default,,0000,0000,0000,,number of them. On the right, I chose then\Nthe credit card thingy and I now want to
Dialogue: 0,0:17:15.14,0:17:21.86,Default,,0000,0000,0000,,show you what happens if you now explore\Nthese properties, right? The first step in
Dialogue: 0,0:17:21.86,0:17:25.75,Default,,0000,0000,0000,,the game is that the game will ask – I\Nmean, the game, the exploration process –
Dialogue: 0,0:17:25.75,0:17:31.02,Default,,0000,0000,0000,,will ask, is it true that every entity in\NWikidata will have these three properties?
Dialogue: 0,0:17:31.02,0:17:36.36,Default,,0000,0000,0000,,So are they common among all entities in\Nyour data, which is most probably not
Dialogue: 0,0:17:36.36,0:17:41.54,Default,,0000,0000,0000,,true, right? I mean, not everything in\NWikidata has a fee, at least I hope. So,
Dialogue: 0,0:17:41.54,0:17:46.52,Default,,0000,0000,0000,,what I will do now, I would click the\N"reject this implication" button, since
Dialogue: 0,0:17:46.52,0:17:51.48,Default,,0000,0000,0000,,the implication "Nothing implies\Neverything" is not true. In the second
Dialogue: 0,0:17:51.48,0:17:56.36,Default,,0000,0000,0000,,step now, the algorithm tries to find the\Nminimal number of questions to obtain the
Dialogue: 0,0:17:56.36,0:18:01.82,Default,,0000,0000,0000,,domain knowledge, so to obtain all valid\Nrules in this domain. So next question is
Dialogue: 0,0:18:01.82,0:18:06.12,Default,,0000,0000,0000,,"is it true that everything in Wikidata\Nthat has a 'card network' property also
Dialogue: 0,0:18:06.12,0:18:12.56,Default,,0000,0000,0000,,has a 'fee' and an 'operator' property?"\NAnd down here you can see Wikidata says
Dialogue: 0,0:18:12.56,0:18:18.11,Default,,0000,0000,0000,,"ok, there are 26 items which are\Ncounterexamples," so there's 26 items in
Dialogue: 0,0:18:18.11,0:18:22.67,Default,,0000,0000,0000,,Wikidata which have the "card network"\Nproperty but do not have the other two
Dialogue: 0,0:18:22.67,0:18:28.20,Default,,0000,0000,0000,,ones. So, 26 is not a big number, this\Ncould mean "ok, that's an error, so 26
Dialogue: 0,0:18:28.20,0:18:32.86,Default,,0000,0000,0000,,statements are missing." Or maybe that\Nthat's, really, that's the true case.
Dialogue: 0,0:18:32.86,0:18:36.89,Default,,0000,0000,0000,,That's also ok. But you can now choose\Nwhat you think is right. You can say, "oh,
Dialogue: 0,0:18:36.89,0:18:40.47,Default,,0000,0000,0000,,I would say it should be true" or you can\Nsay "no, I think that's ok, one of these
Dialogue: 0,0:18:40.47,0:18:46.38,Default,,0000,0000,0000,,counterexamples seems valid. Let's reject\Nit." I in this case, rejected it. The next
Dialogue: 0,0:18:46.38,0:18:51.02,Default,,0000,0000,0000,,question it asks: "is it true that\Neverything that has an operator has also a
Dialogue: 0,0:18:51.02,0:18:56.29,Default,,0000,0000,0000,,fee and a card network?" Yeah, this is\Npossibly not true. There's also more than
Dialogue: 0,0:18:56.29,0:19:03.11,Default,,0000,0000,0000,,1000 counterexamples, one being, I think a\Ntelecommunication operator in Hungary or
Dialogue: 0,0:19:03.11,0:19:10.34,Default,,0000,0000,0000,,something. And so we can reject this as\Nwell. Next question, everything that has
Dialogue: 0,0:19:10.34,0:19:15.36,Default,,0000,0000,0000,,an operator and a card network – so card\Nnetwork means Visa, MasterCard, whatever,
Dialogue: 0,0:19:15.36,0:19:21.69,Default,,0000,0000,0000,,all this stuff – is it true that they have\Nto have a fee?" Wikidata says "no," it has
Dialogue: 0,0:19:21.69,0:19:27.57,Default,,0000,0000,0000,,23 items that contradict it. But one of\Nthe items, for example, is the American
Dialogue: 0,0:19:27.57,0:19:32.09,Default,,0000,0000,0000,,Express Gold Card. I suppose the American\NExpress Gold Card has some fee. So this
Dialogue: 0,0:19:32.09,0:19:36.14,Default,,0000,0000,0000,,indicates, "oh, there is some missing data\Nin Wikidata," there is something that
Dialogue: 0,0:19:36.14,0:19:40.68,Default,,0000,0000,0000,,Wikidata does not know but should know to\Nreason correctly in Wikidata with your
Dialogue: 0,0:19:40.68,0:19:46.52,Default,,0000,0000,0000,,SPARQL queries. So we can now say, "yeah,\Nthat's, uh, that's not a reject, that's an
Dialogue: 0,0:19:46.52,0:19:51.47,Default,,0000,0000,0000,,accept," because we think it should be\Ntrue. But Wikidata thinks otherwise. And
Dialogue: 0,0:19:51.47,0:19:55.80,Default,,0000,0000,0000,,you go on, we go on. This is then the last\Nquestion: "Is it true that everything that
Dialogue: 0,0:19:55.80,0:20:00.95,Default,,0000,0000,0000,,has a fee and a card work should have an\Noperator," and you see, "oh, no counter
Dialogue: 0,0:20:00.95,0:20:05.93,Default,,0000,0000,0000,,examples." This means Wikidata says "this\Nis true," because it says there is no
Dialogue: 0,0:20:05.93,0:20:09.58,Default,,0000,0000,0000,,counterexample. If you're asking Wikidata\Nit says this is a valid implication in the
Dialogue: 0,0:20:09.58,0:20:15.40,Default,,0000,0000,0000,,data set so far, which could also be\Nindicating that something is missing, I'm
Dialogue: 0,0:20:15.40,0:20:20.31,Default,,0000,0000,0000,,not aware if this is possible or not, but\Nok, for me it sounds reasonable. Everyone
Dialogue: 0,0:20:20.31,0:20:23.80,Default,,0000,0000,0000,,has a fee and a card network should also\Nhave an operator, which meens a bank or
Dialogue: 0,0:20:23.80,0:20:29.22,Default,,0000,0000,0000,,something like that. So I accept this\Nimplication. And then, yeah, you have won
Dialogue: 0,0:20:29.22,0:20:34.41,Default,,0000,0000,0000,,the exploration game, which essentially\Nmeans you've won some knowledge. Thank
Dialogue: 0,0:20:34.41,0:20:40.30,Default,,0000,0000,0000,,you. And the knowledge is that you know\Nwhich implications in Wikidata are true or
Dialogue: 0,0:20:40.30,0:20:44.34,Default,,0000,0000,0000,,should be true from your point of view.\NAnd yeah, this is more or less the state
Dialogue: 0,0:20:44.34,0:20:50.70,Default,,0000,0000,0000,,of the game so far as we programmed it in\NOctober. And the next state will be to
Dialogue: 0,0:20:50.70,0:20:54.97,Default,,0000,0000,0000,,show you some – "How much does your\Nopinion of the world differ from the
Dialogue: 0,0:20:54.97,0:20:59.95,Default,,0000,0000,0000,,opinion that is now reflected in the\Ndata?" So is what you think about the data
Dialogue: 0,0:20:59.95,0:21:05.43,Default,,0000,0000,0000,,true, close to true to what is true in\NWikidata. Or maybe Wikidata has wrong
Dialogue: 0,0:21:05.43,0:21:10.68,Default,,0000,0000,0000,,information. You can find it with that.\NBut Max will tell me more about that.
Dialogue: 0,0:21:10.68,0:21:18.22,Default,,0000,0000,0000,,M: Ok. So let me just quickly come\Nback to what we have actually done. So we
Dialogue: 0,0:21:18.22,0:21:23.67,Default,,0000,0000,0000,,offer a procedure that allows you to\Nexplore properties in Wikidata and the
Dialogue: 0,0:21:23.67,0:21:30.72,Default,,0000,0000,0000,,implicational knowledge that holds between\Nthese properties. And the key idea's here
Dialogue: 0,0:21:30.72,0:21:34.66,Default,,0000,0000,0000,,that when you look at these implications\Nthat you get, while there might be some
Dialogue: 0,0:21:34.66,0:21:39.28,Default,,0000,0000,0000,,that you don't actually want because they\Nshouldn't be true, and there might also be
Dialogue: 0,0:21:39.28,0:21:46.22,Default,,0000,0000,0000,,ones that you don't get, but you expect to\Nget because they should hold. And these
Dialogue: 0,0:21:46.22,0:21:51.84,Default,,0000,0000,0000,,unwanted and/or missing implications, they\Npoint to missing statements and items in
Dialogue: 0,0:21:51.84,0:21:56.13,Default,,0000,0000,0000,,Wikidata. So they show you where the\Nopportunities to improve the knowledge in
Dialogue: 0,0:21:56.13,0:22:00.10,Default,,0000,0000,0000,,Wikidata are, and, well, sometimes you\Nalso get to learn something about the
Dialogue: 0,0:22:00.10,0:22:04.08,Default,,0000,0000,0000,,world, and in most cases, it's that the\Nworld is more complicated than you thought
Dialogue: 0,0:22:04.08,0:22:10.26,Default,,0000,0000,0000,,it was – and that's just how life is. But\Nin general, implications can guide you in
Dialogue: 0,0:22:10.26,0:22:17.22,Default,,0000,0000,0000,,your way of improving Wikidata and the\Nstate of knowledge therein. So what's
Dialogue: 0,0:22:17.22,0:22:22.38,Default,,0000,0000,0000,,next? Well, so what we currently don't\Noffer in the exploration game and what we
Dialogue: 0,0:22:22.38,0:22:27.71,Default,,0000,0000,0000,,definitely will focus next on is having\Nconfigurable counterexamples and also
Dialogue: 0,0:22:27.71,0:22:32.03,Default,,0000,0000,0000,,filterable counterexamples – right now you\Njust get a list of a random number of
Dialogue: 0,0:22:32.03,0:22:36.88,Default,,0000,0000,0000,,counterexamples. And you might want to\Nsearch through this list for something you
Dialogue: 0,0:22:36.88,0:22:42.52,Default,,0000,0000,0000,,recognise and you might also want to\Nexplicitly say, well, this one should be a
Dialogue: 0,0:22:42.52,0:22:48.60,Default,,0000,0000,0000,,counterexample, and that's definitely\Ncoming next. Then, well, domain specific
Dialogue: 0,0:22:48.60,0:22:53.75,Default,,0000,0000,0000,,scaling of properties, there's still much\Nwork to be done. Currently, we only have
Dialogue: 0,0:22:53.75,0:23:00.50,Default,,0000,0000,0000,,some very basic support for that. So you\Ncan have properties, but you can't do the
Dialogue: 0,0:23:00.50,0:23:03.78,Default,,0000,0000,0000,,fancy things where you say, "well,\Neverything that is an award should be
Dialogue: 0,0:23:03.78,0:23:10.84,Default,,0000,0000,0000,,considered as one instance of this\Nproperty." That's also coming and then
Dialogue: 0,0:23:10.84,0:23:15.55,Default,,0000,0000,0000,,what Tom mentioned alread: compare your\Nknowledge that you have explored through
Dialogue: 0,0:23:15.55,0:23:21.61,Default,,0000,0000,0000,,this process against the knowledge that is\Ncurrently on Wikidata as a form of seeing
Dialogue: 0,0:23:21.61,0:23:26.54,Default,,0000,0000,0000,,"where do you stand? What is missing in\NWikidata? How can you improve Wikidata?"
Dialogue: 0,0:23:26.54,0:23:32.60,Default,,0000,0000,0000,,And well, if you have any more suggestions\Nfor features, then just tell us. There's a
Dialogue: 0,0:23:32.60,0:23:39.53,Default,,0000,0000,0000,,Github link on the implication game page.\NAnd here's the link to the tool again. So,
Dialogue: 0,0:23:39.53,0:23:46.14,Default,,0000,0000,0000,,yeah, just let us know. Open an issue and\Nhave fun. And if you have any questions,
Dialogue: 0,0:23:46.14,0:23:50.23,Default,,0000,0000,0000,,then I guess now would be the time to ask.\NT: Thank you.
Dialogue: 0,0:23:50.23,0:23:52.73,Default,,0000,0000,0000,,Herald: Thank you very much, Tom and Max.
Dialogue: 0,0:23:52.73,0:23:55.02,Default,,0000,0000,0000,,{\i1}applause{\i0}
Dialogue: 0,0:23:55.02,0:24:01.51,Default,,0000,0000,0000,,Herald: So we will switch microphones now\Nbecause then I can hand this microphone to
Dialogue: 0,0:24:01.51,0:24:07.25,Default,,0000,0000,0000,,you if any of you have a question for our\Ntwo speakers. Are there any questions or
Dialogue: 0,0:24:07.25,0:24:14.37,Default,,0000,0000,0000,,suggestions? Yes.\NQuestion: Hi. Thanks for the nice talk. I
Dialogue: 0,0:24:14.37,0:24:18.72,Default,,0000,0000,0000,,wanted to ask what's the first question,\Nwhat's the most interesting implication
Dialogue: 0,0:24:18.72,0:24:25.02,Default,,0000,0000,0000,,that you've found?\NM: Yeah. That would have made for a
Dialogue: 0,0:24:25.02,0:24:31.85,Default,,0000,0000,0000,,good back up slide. The most interesting\Nimplication so far –
Dialogue: 0,0:24:31.85,0:24:36.01,Default,,0000,0000,0000,,T: The most basic thing you would expect\Neverything that is launched in space by
Dialogue: 0,0:24:36.01,0:24:41.92,Default,,0000,0000,0000,,humans – no, everything that landed from\Nspace, that has a landing date, also has a
Dialogue: 0,0:24:41.92,0:24:46.45,Default,,0000,0000,0000,,start date. So nothing landed on earth,\Nwhich was not started here.
Dialogue: 0,0:24:46.45,0:24:55.20,Default,,0000,0000,0000,,M: Yes.\NQ: Right now, the game only helps you find
Dialogue: 0,0:24:55.20,0:25:00.71,Default,,0000,0000,0000,,out implications. Are you also planning to\Nhave that I can also add data like for
Dialogue: 0,0:25:00.71,0:25:04.31,Default,,0000,0000,0000,,example, let's say I have twenty five\NNobel laureates who don't have a Nobel
Dialogue: 0,0:25:04.31,0:25:08.22,Default,,0000,0000,0000,,laureate ID. Is there plans where you\Ncould give me a simple interface for me to
Dialogue: 0,0:25:08.22,0:25:12.76,Default,,0000,0000,0000,,Google and add that ID because it would\Nmake the process of adding new entities to
Dialogue: 0,0:25:12.76,0:25:17.40,Default,,0000,0000,0000,,Wikidata itself more simple.\NM: Yes. And that's partly hidden
Dialogue: 0,0:25:17.40,0:25:23.05,Default,,0000,0000,0000,,behind this "configurable and filterable\Ncounterexamples" thing. We will probably
Dialogue: 0,0:25:23.05,0:25:28.38,Default,,0000,0000,0000,,not have an explicit interface for adding\Nstuff, but most likely interface with some
Dialogue: 0,0:25:28.38,0:25:32.27,Default,,0000,0000,0000,,other tool built around Wikidata, so\Nprobably something that will give you
Dialogue: 0,0:25:32.27,0:25:37.10,Default,,0000,0000,0000,,QuickStatements or something like that.\NBut yes, adding data is definitely on the
Dialogue: 0,0:25:37.10,0:25:41.71,Default,,0000,0000,0000,,roadmap.\NHerald: Any more questions? Yes.
Dialogue: 0,0:25:41.71,0:25:48.86,Default,,0000,0000,0000,,Q: Wouldn't it be nice to do this in other\Nlanguages, too?
Dialogue: 0,0:25:48.86,0:25:52.60,Default,,0000,0000,0000,,T: Actually it's language independent, so\Nwe use Wikidata and then as far as we
Dialogue: 0,0:25:52.60,0:25:58.11,Default,,0000,0000,0000,,know, Wikidata has no language itself. You\Nknow, it has just items and properties, so
Dialogue: 0,0:25:58.11,0:26:02.64,Default,,0000,0000,0000,,Qs and Ps, and whatever language you use,\Nit should be translated in the language of
Dialogue: 0,0:26:02.64,0:26:06.18,Default,,0000,0000,0000,,the properties, if there is a label for\Nthat property or for that item that you
Dialogue: 0,0:26:06.18,0:26:12.42,Default,,0000,0000,0000,,have. So if Wikidata is aware of your\Nlanguage, we are.
Dialogue: 0,0:26:12.42,0:26:15.02,Default,,0000,0000,0000,,Herald: Oh, yes. More!\NM: Of course, the tool still needs to be
Dialogue: 0,0:26:15.02,0:26:18.36,Default,,0000,0000,0000,,translated, but –\NT: The tool itself, it should be.
Dialogue: 0,0:26:18.36,0:26:21.85,Default,,0000,0000,0000,,Q: Hi, thanks for the talk. I have a\Nquestion. Right now you only can find
Dialogue: 0,0:26:21.85,0:26:25.99,Default,,0000,0000,0000,,missing data with this, right? Or surplus\Ndata. Would you think you'd be able to
Dialogue: 0,0:26:25.99,0:26:31.56,Default,,0000,0000,0000,,find wrong information with a similar\Napproach.
Dialogue: 0,0:26:31.56,0:26:37.00,Default,,0000,0000,0000,,T: Actually, we do. I mean, if we Wikidata\Nhas a counterexample to something we would
Dialogue: 0,0:26:37.00,0:26:42.83,Default,,0000,0000,0000,,expect to be true, this could point to\Nwrong data, right? If the counterexample
Dialogue: 0,0:26:42.83,0:26:47.45,Default,,0000,0000,0000,,is a wrong counterexample. If there is a\Nmissing property or missing property to an
Dialogue: 0,0:26:47.45,0:26:58.16,Default,,0000,0000,0000,,item.\NQ: Ok, I get to ask a second question. So
Dialogue: 0,0:26:58.16,0:27:06.00,Default,,0000,0000,0000,,the horizontal axis in the incidence\Nmatrix. You said it has 7000, it spans
Dialogue: 0,0:27:06.00,0:27:10.30,Default,,0000,0000,0000,,7000 columns, right?\NM: Yes, because there's 7000 properties in
Dialogue: 0,0:27:10.30,0:27:13.85,Default,,0000,0000,0000,,Wikidata.\NQ: But it's actually way more columns,
Dialogue: 0,0:27:13.85,0:27:17.85,Default,,0000,0000,0000,,right? Because you multiply the properties\Ntimes the arguments, right?
Dialogue: 0,0:27:17.85,0:27:21.36,Default,,0000,0000,0000,,M: Yes. So if you do any scaling then of\Ncourse that might give you multiple
Dialogue: 0,0:27:21.36,0:27:23.38,Default,,0000,0000,0000,,entries.\NQ: So that's what you mean with scaling,
Dialogue: 0,0:27:23.38,0:27:27.77,Default,,0000,0000,0000,,basically?\NM: Yes. But already seven thousand is way
Dialogue: 0,0:27:27.77,0:27:35.58,Default,,0000,0000,0000,,too big to actually compute that.\NQ: How many would it be if you multiply
Dialogue: 0,0:27:35.58,0:27:48.06,Default,,0000,0000,0000,,all the arguments?\NM: I have no idea, probably a few million.
Dialogue: 0,0:27:48.06,0:27:55.31,Default,,0000,0000,0000,,Q: Have you thought about a recursive\Nmethod, as counterexamples may be wrong by
Dialogue: 0,0:27:55.31,0:28:00.35,Default,,0000,0000,0000,,other counterexamples, like in an\Nargumentative graph or something like
Dialogue: 0,0:28:00.35,0:28:06.71,Default,,0000,0000,0000,,this?\NT: Actually, I don't get it. How can a
Dialogue: 0,0:28:06.71,0:28:14.04,Default,,0000,0000,0000,,counterexample be wrong through another\Ncounterxample?
Dialogue: 0,0:28:14.04,0:28:24.45,Default,,0000,0000,0000,,Q: Maybe some example says that cats can\Nhave golden hair and then another example
Dialogue: 0,0:28:24.45,0:28:31.26,Default,,0000,0000,0000,,might say that this is not a cat.\NT: Ah, so the property to be a cat or
Dialogue: 0,0:28:31.26,0:28:38.00,Default,,0000,0000,0000,,something cat-ish is missing then. Okay.\NNo, we have not considered so far deeper
Dialogue: 0,0:28:38.00,0:28:44.57,Default,,0000,0000,0000,,reasoning. This horn-propositional logic,\Nyou know, it has no contradictions,
Dialogue: 0,0:28:44.57,0:28:47.74,Default,,0000,0000,0000,,because all you can do is you can\Ncontradict by counterexamples, but there
Dialogue: 0,0:28:47.74,0:28:52.74,Default,,0000,0000,0000,,can never be a rule that is not true, so\Nfar. Just in your or my opinion, maybe,
Dialogue: 0,0:28:52.74,0:28:56.37,Default,,0000,0000,0000,,but not in the logic. So what we have to\Nthink about is that we have bigger
Dialogue: 0,0:28:56.37,0:29:01.78,Default,,0000,0000,0000,,reasoning, right? So.\NQ: Sorry, quick question. Because you're
Dialogue: 0,0:29:01.78,0:29:04.93,Default,,0000,0000,0000,,not considering all the 7000 odd\Nproperties for each of the entities,
Dialogue: 0,0:29:04.93,0:29:07.57,Default,,0000,0000,0000,,right? What's your current process of\Nfiltering? What are the relevant
Dialogue: 0,0:29:07.57,0:29:14.82,Default,,0000,0000,0000,,properties? I'm sorry, I didn't get that.\NM: Well, we basically handpick those. So
Dialogue: 0,0:29:14.82,0:29:19.94,Default,,0000,0000,0000,,you have this input field? Yeah, we can go\Nahead and select our properties. We also
Dialogue: 0,0:29:19.94,0:29:26.87,Default,,0000,0000,0000,,have some predefined sets. Okay. And\Nthere's also some classes for groups of
Dialogue: 0,0:29:26.87,0:29:30.78,Default,,0000,0000,0000,,properties that are related that you could\Nuse if you want bigger sets,
Dialogue: 0,0:29:30.78,0:29:35.96,Default,,0000,0000,0000,,T: for example, space or family or what\Nwas the other?
Dialogue: 0,0:29:35.96,0:29:43.41,Default,,0000,0000,0000,,M: Awards is one.\NT: It depends on the size of the class.
Dialogue: 0,0:29:43.41,0:29:47.39,Default,,0000,0000,0000,,For example, for space, it's not that\Nmuch, I think it's 10 or 15 properties. It
Dialogue: 0,0:29:47.39,0:29:51.52,Default,,0000,0000,0000,,will take you some hours, but you can do\Nbecause they are 15 or something like
Dialogue: 0,0:29:51.52,0:29:58.15,Default,,0000,0000,0000,,that. I think for family, it's way too\Nmuch, it's like 40 of 50 properties. So a
Dialogue: 0,0:29:58.15,0:30:04.54,Default,,0000,0000,0000,,lot of questions.\NHerald: I don't see any more hands. Maybe
Dialogue: 0,0:30:04.54,0:30:09.76,Default,,0000,0000,0000,,someone who has not asked the question yet\Nhas another one we could take that,
Dialogue: 0,0:30:09.76,0:30:14.27,Default,,0000,0000,0000,,otherwise we would be perfectly on time.\NAnd maybe you can tell us where you will
Dialogue: 0,0:30:14.27,0:30:18.86,Default,,0000,0000,0000,,be for deeper discussions where people can\Nfind you.
Dialogue: 0,0:30:18.86,0:30:22.40,Default,,0000,0000,0000,,T: Probably at the couches.\NHerald: The couches, behind our stage.
Dialogue: 0,0:30:22.40,0:30:26.72,Default,,0000,0000,0000,,M: Or just running around somewhere. So\Nthere's also our DECT numbers on the
Dialogue: 0,0:30:26.72,0:30:35.96,Default,,0000,0000,0000,,slides; it's 6284 for Tom and 6279 for me.\NSo just call and ask where we're hanging
Dialogue: 0,0:30:35.96,0:30:38.47,Default,,0000,0000,0000,,around.\NH: Well then, thank you again. Have a
Dialogue: 0,0:30:38.47,0:30:40.21,Default,,0000,0000,0000,,round of applause.\N{\i1}applause{\i0}
Dialogue: 0,0:30:40.21,0:30:42.65,Default,,0000,0000,0000,,T: Thank you.\NM: Well, thanks for having us.
Dialogue: 0,0:30:42.65,0:30:45.31,Default,,0000,0000,0000,,{\i1}Applause{\i0}
Dialogue: 0,0:30:45.31,0:30:49.74,Default,,0000,0000,0000,,{\i1}postroll music{\i0}
Dialogue: 0,0:30:49.74,0:31:12.00,Default,,0000,0000,0000,,subtitles created by c3subtitles.de\Nin the year 2020. Join, and help us!