44 LXF249May 2019 http://www.linuxformat.com
IntervIew Omoju Miller
theyreally?ThefirstAIclassItookwas
asortofcognitivescienceseminar;it
wasadifferentkindofthingbecause
itgaveasortofhistoricprecedenceof
whythiswholethingstarted.Itwent
allthewaybackintoanexplorationof
themetaphysical,andhowhumanity
questionedthenatureofreality,whywe
arehereandifthereisagod,andhow
thoselinesofinquiryledtothefoundingof
mathematics,tothefoundingofphysics.
Mathematicswereawaytoseparate
whatweknowfromwhatwedon’tknow
andmakeinferences.Wecontinueddown
thatpathallthewaydowntologic,allthe
waytovonNeumannmachines,tothe
inventionofcomputersandallthatkind
ofstuff.It’sjustanaturalprogression,
logicmachinestoAI.It’sthesamequest
offiguringoutthenatureofhumans,andseewhatotherpeoplearedoing.More
importantly,forme,istogetaviewoutside
ofjustAI.Whatarepeoplethinkingabout
ineducation?Whatarepeoplethinking
aboutinpublichealth?Thatkindofthing.
Andthenseeingwhatkindofinitiativesthe
WorldEconomicForumisupto.
lXF:Whatmadeyouspecialiseinartificial
intelligenceandmachinelearning?
om: I was already fascinated with
Computer Science, but I became even
more fascinated when I discovered AI, and
it was specifically for one reason... In high
school, you know, you get your math book.
If you’re learning Pythagoras’ theorem
there will be a picture of Pythagoras and
a little blurb about his life. I was always
fascinated about all these people in
these books – what kind of people were
if we can figure out the nature of humans
that would be like a scientific way to
question the existential. So I thought this
was all fascinating, and that was it, I just
wanted to do that. It’s like metaphysical
exploration but from a scientific stance.
I didn’t really think of it as machine learning
orAI.lXF:Wasitmoreofacontinuation
ofsomebiggerjourney?
om: Right. Basically I thought it was really
interesting, but it turns out it’s also really
usefultoo.lXF: I’m quite interested in computer
science education, so let’s continue
with that theme. I enjoyed your TEDx
talk about dispelling “The myth of the
innate ability in tech” [see http://bit.ly/
lxf249tedx 2 ], particularly as it applies
to coding. What is wrong with this idea,
that when someone shows some inkling
of being good at maths, or whatever, that
wethenthink,ohthey’regifted,they’re
definitelygoingtobeagoodcoder?
om: So this myth, I don’t think we quite put
that term to it, but what we’re saying is it’s
essentially biological determinism. We’re
saying some people are born with a special
genome that gives them a special ability to
do computationally rigourous work, while
others can’t. It’s the same idea behind
people that say “I’m a math person” or “I’m
not a math person”.
There’s no such thing as predisposition
to mathematics or computation or
to reading. It’s about how we haven’t
done enough research in education to
actually understand how people acquire
computation. They acquire knowledge in
two modes: an emotional path and a logical
path. These two paths have to weave
together for the knowledge to seep into a
human being. Human beings are embodied
creatures and they are deeply emotional;
the concepts that they have in their minds
are emotionally mediated as well.
Even in the language we use, when
you think of falling in love, you think of
love as warm and the absence of love
as cold. These are deeply emotionally
triggered concepts and it’s related first to
the warmth felt by the baby inside their
mother’s womb. Later, when the baby
is laid on the mother’s chest, warmth is
transferred from mother to child. If the
baby is put down, then the heat dissipates
and the baby doesn’t like that.
And that is one of our first associations
with something negative, that loss of heat,
that cold. We humans try and cheat that
by swaddling, we’re trying to trap some
of that heat. This is how we learn and it’s
deeplywiredintoourbrain.WhenitcomesbecomIng non-toXIc
lXF: Open source has become quite
popular, and unfortunately that has
brought with it some pretty toxic
behaviour in some communities. What
doyouthinkwecandotofurther
addressthisproblem?
om:Ithinkwehavetohaveverystrong
codesofconductandverystrong
communityandsafetyteams.Itmight
seemstrange,becauseopensource
isopensource–there’snotsomeone
sittingthereworkingfulltime[inevery
project].However,therearesomeopen
sourceprojectsthathavegrantsand
actuallydohavefull-timeemployees.
Whenaprojectgetstothatstage,I
thinktheyneedtoprioritisehavinga
communityandsafetyteam:agroup
ofpeoplewhosewholejobistocurate
thecultureandsetthetoneforwhatis
acceptableandwhatisnotacceptable.
Andthensomethingradical,which
Iwouldliketoseemoreof:instead
ofbanningpeoplefromprojects,
whatifwedidthingslikerestorativejustice?Likereformandworkwiththe
communitytobringthemback.That
seemshealthierthankickingpeople
out,whichdoesn’thelpanybody.
SomethingelseI’minterestedin
deeply:canwedosomeresearchto
seethegenesisofcommunitiesand
whataretheleadingindicatorsofa
communitybecomingtoxic?Thenwe
canwenipthosethingsinthebud.
Anotherthingisthatit’squite
difficulttofigureoutwhatistoxic
andwhatisn’t,becauseit’soften
community-dependent.Sometimes
certainbehaviourisn’ttoxicinone
communitybecauseit’sthenorm,
anditdoesn’tmarginaliseorostracise
anybody.Butwhensomeonefromthat
communityjoinsanotherandcarries
thatnormover,itmightbetoxicinthis
newcommunity.
Howcanwechangethis?Bydoing
researchinthisarea,hard-core
researchanddatamining,andthen
actingonthatresearch.2) http://www.youtube.com/watch?v=BFWVHSeakkg