Geospatial Forum: David DiBiase

Geospatial Forum: David DiBiase


[Money] Let’s go ahead and get started. So it’s
my pleasure to welcome David DiBiase for our Geospatial Forum this week. For those
you that don’t know me, I’m Eric money. I’m an Associate Director here in the
Center for Geospatial Analytics. So David comes to us from ESRI. He is
currently the director of education in ESRI’s global business development group,
which means that he leads an outreach team that promotes and supports GIS used
by educators, students, campus administrators at thousands of
universities across the country and the world. And he teaches introduction to GIS
and professional ethics part-time for Penn State, where he was also a
mastermind behind their pioneering online geospatial programs. And he’s
currently, like I said, with the global business development group but he also
was lead editor for the first editions of the GIST Body of Knowledge. Many of
you have probably come in contact with that that before, as well as the US Department
of Labor’s geospatial technology competency model. So today he’s gonna
be talking about something I think we’ve all heard about but maybe don’t quite
understand, and how it might impact the work that we do as do spatial
professionals. Talking about the Internet of Things. So I will turn it over to David with that. Thank you. [DiBiase] Thank you so much
for showing up. It’s a real privilege to be here with you. I was so impressed with
everything I saw during my visit today, and you all should be very proud to be
part of this outstanding program. And so thanks, Eric, for the introduction and for
mentioning my association with Penn State, which continues. Some of my new
friends who I shared lunch with today heard the story about how my dream
came true in the last few months. My dream was–it’s a pretty humble dream, I
guess–but my dream was that I would get a chance to reconceive and rewrite the introduction in the gateway course to the Penn State programs that I first wrote 20 years ago.
And over the past year I had that chance, and the
thing just debuted, and very excited about. The presentation that I’m
going to read to you today is actually the last presentation in that new course.
So this was like, this is all I got right. I’m laying everything out, everything I
could bring to the table I’m trying to put it into this presentation, and I am
going to read it to you, and there’s there’s a one thing, a truism that
students should keep in mind. You may have heard this before, but it always
helps to hear it again: time passes much more quickly when
you’re talking than when you’re listening, and so in the interests of
making sure that I dispatch this talk in a timely way and don’t get distracted
and don’t forget anything, I am going to read, so I hope you’ll you will bear with
me as I do that. This course that–the new course was the very first course in the
masters and certificate program–started with the premise of a guy named Simon
Cynic. Has anybody heard of Simon Cynic? Simon–you’ve heard of Simon? He
makes the point that whenever you’re trying to convince somebody of something,
anything, start with why. Why does something matter,
before you talk about the what or the how or the ask. Start with why. And so I
did that in this class and tried to make a course about why GIS matters. So each
lesson in this class begins with a story about the roles geospatial
technology, and GIS people, play in fighting disease, in helping assess
progress toward the United Nations sustainable development goals, in
supporting analysis and modeling of everyday tasks, as well as extreme events,
and in enabling developers and app builders to lever the power of geography
to inform the public and policymakers. So it’s uh–should be no mystery to anyone
here that there’s plenty of evidence that GIS matters in today’s world.
For example, this economic development or, I’m sorry, economic impact study by a
group funded by Google estimated that the economic impact of geo services, as
they defined them, is two to three hundred billion dollars per year
worldwide and the operating definition that Google had for geo services didn’t
even include key business sectors of the GIS industry such as state, local, and
national governments. We know that geo services increase efficiency across many
industries and that they make business operations greener and cities smarter,
and that they even sometimes help save lives.
We know GIS matter because maybe a million livelihoods around the world
depend upon this technology, and thousands of those folks show up in
the annual pilgrimage to San Diego every year for the ESRI user conference. You
haven’t done that before you should do it at least once. We know GIS matters
because nearly 10,000 individuals have voluntarily earned certification as GIS
professionals, and thousands of students like you seek education and training
opportunities every year to start or advance careers related to GIS. So
summing it all up, there’s plenty of evidence that GIS matters, but there are
also signs that the world is changing in fundamental ways. Will GIS still matter,
and will GIS professionals continue to have meaningful and rewarding roles to
play, in years to come? First let’s consider whether GIS itself is likely to
persist. This guy is Mark Weise.r Mark was a leading computer scientist at
Xerox PARC labs back at the dawn of personal computing. As you may know, PARC
is where Apple got ideas about graphical user interfaces, like the mouse among other things. In 1995,
Mark famously wrote in Scientific American that the most profound
technologies are those that disappear. They weave themselves into the fabric of
everyday life until they are indistinguishable from it. As an example
Weiser asked us to consider writing. He pointed out that our lives are suffused
with literacy technology, written information in printed or digital forms
and in street signs, billboards, shop signs, and even graffiti. These do not
require active attention, Weiser wrote, but the information to be conveyed is
ready for use at a glance. By contrast, he said, silicon-based information
technology is far from having become part of the environment. How things have
changed since Weiser’s untimely passing in 1999. One big change is the ongoing
evolution of the internet into a ubiquitous network of interconnected
devices and objects–an Internet of Things. The Organisation for Economic
Cooperation and Development and others project that IOT will connect 50 billion
devices by 2020. That is not the clicker I want..this is. What kinds of things?
Consider this chart from the OECD’s digital economy outlook. Probably hard
for you to read–you have to look it up online. The chart compares the number and
kinds of connected devices in use in typical households in OECD countries–
that is the global North–in 2012, 2017, and 2022. Now even if you can’t read the
fine print, you can see the list gets a lot longer, right? As time goes on, lots
more connected devices. And I wonder, how does your household compare with the
one described here? Gartner, the influential corporate strategy
advisors, predict 30 trillion dollars of spending on the Internet of Things by
2020–30 trillion dollars. Connected automobiles and other transportation
modes will account for a large share of the spending, as will home
automation, security, and energy management. Gartner stresses that the IOT
is not one thing–it’s the integration of several technologies and that sense and
collect data, that analyze the data, and take action upon the data to accomplish
business goals. Sound familiar? Make no mistake, the IOT is all about
business. Beechum Research is a technology market
research consultancy that specializes in what it calls the connected devices
market, sometimes referred to as M2M or machine to machine sector, also known as
the Internet of Things. Like Gartner, Beecham Sells market research reports
and advice to businesses that seek opportunity in the IOT. Beecham created
this amazing map of the IOT to help its clients identify business opportunities.
The map highlights nine market sectors from left to right: buildings, energy,
consumer and home, health care and life science, industrial, transportation, retail
security and public safety, and IT and networks. So those are the markets
sectors in which that 30 trillion dollars is going to get spent. You might
ask yourself, “Hmm, what doesn’t appear?” Like science? The map also identifies
applications and particular devices within each sector. For example, the
security and public safety sector includes surveillance applications,
connected equipment including weapons, vehicles, ships, and other gear, location
tracking of people and assets, connected public infrastructure such as water
treatment facilities and environmental sensors,
and connected emergency services personnel and equipment. Consider that
every one of those billions of IOT devices has a location, and many are
location-aware. We’ll consider the implication of that in a minute.
If the IOT makes you nervous, it’s probably because of hackers. Many of
those billions of IOT connected devices each have a tiny bit of computing power.
Hackers who can harness millions of tiny computing devices can combine them to
create a massive computing capacity they can use to mount large-scale attacks on
businesses, government agencies, and public infrastructure. Cyber security
threats are also business opportunities. Beechum Research maps these threats, in
this map of the IOT security threat map, and it offers advices to businesses that
aim to sell security solutions. Gartner says it expects many new IOT security
and management vendors to arrive on the scene. Lots of businesses are focusing on
IOT. Did you know there’s even a Geo IOT World Conference? And Geo IOT World
awards for innovative products and services? At the second Geo IOT World
Conference in Brussels earlier this year, four companies were recognized. One of
those is called Sewio. Sewio describes itself as a precise indoor tracking and data
analytics platform for the digitisation of movement in industry 4.0,
retail, and sport. Industry 4.0, by the way, refers to what the World Economic Forum
calls the fourth Industrial Revolution. First came an agricultural revolution
10,000 years ago. Then beginning in the 18th century the
invention of the steam engine and construction of railroads brought the
first Industrial Revolution. A Second Industrial Revolution began in the 19th
century with the advent of mass production. Digital computers
heralded a third Industrial Revolution beginning in the 1960s, and today the
drivers of the fourth Industrial Revolution include a ubiquitous and
mobile internet; smaller, cheaper, and more powerful sensors; and artificial
intelligence and machine learning. So you can think of these diagrams as treasure
maps. Use your imagination. How might geospatial technologies, analytics,
and apps create value from the big data and the big threats produced by the
Internet of Things? As Alec Ross has written in The Industries of Tomorrow, if
you can imagine an innovation in information technology, chances are
somebody somewhere is already working on developing and commercializing it. Tim
Forsman is a former United Nations chief environmental scientist and
national manager for the Digital Earth Initiative under Vice President Al Gore.
With operations research specialist Ruth Luscombe of Brisbane,
Australia, Forsman recently proposed a second law of geography for the fourth
Industrial Age. Things that know where they are can act on their
locational knowledge, Forsman and Luscombe assert. Furthermore, they say
spatially enabled things have increased financial and functional utility. This
increased utility, they argue, creates the basis for a spatially enabled economy.
That economy–that is the economy we and our children will inherit. So location is
elemental to the spatially enabled economy and to the IOT. Location
analytics is a defining feature of geographic information systems. So, will GIS as we know it still be a thing in the Internet of
Things? Or will it disappear, like other profound technologies? Well ESRI
certainly believes that GIS is IOT ready and is here to stay.
The ArcGIS Enterprise suite introduced this year includes specialized server
technologies engineered to ingest, analyze, and store millions of sensor
events per second. That’s fast enough, ESRI claims, to monitor all the sensors and
smart meters used by major water, oil, gas, and electric utilities–and to track and
analyze the movement and disposition of large fleets of trucks, ships, and
aircraft. Meanwhile, leading data management and analytic corporations,
like Oracle and SAP, cloud vendors such as Google and Salesforce, and established
industrial technology providers like General Electric are expected to offer
their own IOT platform solutions. Will GIS successfully compete with those? Or
integrate with them? Or will its key capabilities merge into them? Paper maps
and portable navigation devices have already disappeared, in a sense. They are
counted among the ten things killed by the smartphone. By accelerating the trend
toward integration of GIS with mainstream information technology, will
the Internet of Things and the fourth Industrial Revolution kill GIS? On the
other hand, recall Mark Weiser’s observation that profound technologies
disappear. If information systems persist, does that mean they were not a profound
technology in the first place? Is spatial really not special after all, but just
another data type? More important than the fate of GIS technology per se are
the prospects for GIS people. Will the education you’re investing so
much in have lasting value? According to the US Department of Labor, the outlook
for geospatial information scientists and technologists, as it calls them, is
bright. DOL estimates that nearly a quarter of a
million people are employed in this occupation, and although predicted growth
is just two to four percent through 2024, that’s still nearly 38,000 additional
GIS jobs in this one GIS-related occupation, in the US alone. On the other
hand, thought leaders concerned with the impacts of the fourth Industrial
Revolution worry that many of today’s occupations may not be sustainable. In a
widely cited research article, economist Karl Benedict Frey and machine learning
researcher Michael Osborn estimate that 47 percent of US workers are at risk of
technological unemployment. Of the 702 occupations Frey and Osbourne analyzed,
one of the most susceptible was surveying and mapping technicians. Frey
and Osbourne calculated a 96% probability that workers in that
occupation will be displaced by automation in the coming decade or two.
Although the Bureau of Labor Statistics, shown here, predicts only an 8 percent
decline, it does attribute the decline to advances in technology. Now, although Frey
and Osborne’s research has its critics, their prognosis is generally consistent
with a body of research by economists, tech leaders, and forward-looking
historians who anticipate fundamental disruption of traditional employment by
increasingly capable machines. What does this mean for GIS work? Innovation expert
Alec Ross observes that through history our most valuable commodities have gone
from salt and sugar to chemicals and fuels to data and services.
Not just the internet of things but international finance, social media, and
other human activities generate an unprecedented an ever-increasing volume,
velocity, and variety of data. Human analysts and their employers, Ross and
others foresee, will rely increasingly on machine learning and artificial
intelligence to cope with the data deluge. In 2014, Jong Jin and colleagues
published an illuminating paper about how the IOT enables planners and
engineers to design smart cities. Illustrated by a case study involving
noise mapping in Melbourne, Australia, Jin and team discuss the data collection,
data processing and management, and data interpretation aspects of an IOT-enabled
urban information system. GIS plays a role in their framework, specifically for
the integration and visualization of geo-referenced data. Considering the
massive data throughputs generated by the IOT, Jin and colleagues
observe that to make sense of the information and convert it into
knowledge, state-of-the-art computational intelligence techniques such as genetic
algorithms, evolutionary algorithms, and neural networks are necessary. Machine
learning, they conclude, will help achieve automated decision-making and provide
useful policy. Think about that for a minute.
Automated decision-making. Consider this scenario about self-driving cars,
published by the philosopher Eric Schwitzgebel in the LA Times. You and
your daughter are riding in a driverless car along the Pacific Coast Highway.
The autonomous vehicle rounds a corner and detects a crosswalk full of children.
It brakes, but your lane is unexpectedly full of sand
from a recent rock slide. It can’t get traction. Your car does some calculations.
If it continues braking, there’s a 90% chance that it will kill at least three
children. Should it save them by steering you and your daughter off the cliff? Now,
can you imagine the ethical algorithms that would be needed for an autonomous
urban planning system? Richard and Daniel Susskind, authors of The Future of the
Professions, foresee that in the long run increasingly capable machines will
transform the work of professionals, leaving most to be replaced by less
expert people and high-performing systems. Their hope is that practical
expertise will become more openly available, freeing many users from
obstacles currently imposed by gatekeepers like physicians, lawyers,
accountants, and, well, surveying and mapping technicians. Predictions like the
Susskinds’ about a coming robopocalypse have given rise to what Wired magazine
called the great tech panic. Columnist James Surowiecki argues that the evidence
disagrees that automation will take away our jobs.
Neither the increased productivity that should accompany automation nor growing
unemployment are evident. Sara Ricci points out that US corporate
investment in robotics in 2016 was just eleven point three billion dollars.
That’s about one-sixth of what Americans spend every year on their pets. And he
cites economist James Besson who found that of the 271 occupations listed in
the 1950 US census, only one had been rendered obsolete by automation: elevator
operators. So if your GIS work feels like this, you probably should worry. Sir Ricci rightly points out that the
outsourcing of work to machines is not new. From the cotton gin to the washing
machine to the car, jobs have been destroyed but others have been created.
Over and over, he reminds us, we’ve been terrible at envisioning the new jobs
that people would end up doing. The Susskinds recognize this and don’t predict
future occupations that may replace the traditional professions. However, they do
suggest 12 future roles that education should help people prepare for–they’re
listed there. One that I highlighted– the highlight doesn’t come out very well–
I highlighted this one. Several of these new roles, and one in particular–data
scientists, are related to the knowledge and skills you are acquiring
in these programs. Here’s a witty diagram of the knowledge and skills that data
scientists possess. For Steven Kolasa who drew the diagram, data scientists
combined competencies in statistical analysis, spatial analysis in other words,
programming, coding, app development, business–which in general means
understanding what your organization is trying to accomplish and what you can do
to help– doesn’t mean business in the narrow
sense–and communication abilities like those of the folks who gave such
articulate demos to me this afternoon. The Susskinds describe data scientists
as “masters of the tools and techniques required to capture and analyze large
bodies of information with the intent of identifying correlations, trends, and
causal insights.” Does that sound familiar? This diagram represents the US
Department of Labor’s view of the foundational, academic, workplace, and
industry-specific competencies that are characteristic of what it calls the
geospatial technology industry. In particular, notice the three industry
sectors specified in tier five: positioning in data acquisition, analysis
and modeling, and software and app development. Do these seem relevant to
the future role of spatial data scientist? I believe that if students are
able to pursue an education that balances knowledge and skills in data
acquisition and wrangling, spatial analysis and modeling, and coding and app
building, they’ll be pretty well prepared for current and future roles. And of all
the competencies needed to navigate an uncertain future, the most valuable may
be the ongoing voluntary and self-motivated pursuit of knowledge,
independently or part of a team. Lifelong learning is a cornerstone of the
geospatial technology competency model, and programs like yours that support lifelong learning will help GIS people continue to matter in the
Internet of Things. Thank you. Yeah, that came in at 30 minutes, right? Good. Not too short? I
could read the last few pages over again. Yes? [Mitasova] There were lots of things that can be discussed. We have seen lots of discussions… [inaudible] I was wondering, you mentioned with
surveying there is a decline in the jobs, but–because it’s highly automated–
but what we see is that the more automated it is, the more data we are collecting, and it
means generating the jobs for people who need to actually look at these data. So the more data we have, the more people you need to actually analyze the data and use the data for something. And I
would like to ask you, like, what’s your vision of how do we–where do you see we
are going with this massive amount of data that we are getting and we are not–
we are using maybe 10% of it, or maybe 20% of it–where do you see the future with–how are we going to use
it. What should we do, let’s say in the education, in the research, to make the
data more usable and create probably new jobs and help with decision-making? [DiBiase] So first, with regard to your observation
about the surveying jobs, that that artifact has to do with that particular
occupational definition of surveying and mapping technicians. So that’s not
surveyors; that’s the people who help surveyors, and those are very routine
roles that people believe can be automated. The GIS occupation
that I cited from the Department of Labor didn’t exist when Benedict Frey
did this study; that’s a relatively new occupation, so we don’t have a measure on that. In terms of what do we do about about all these data, well, from
what I read and the people I talk to you, the fact is that machines are becoming
smarter than people. And machines know how to learn now, right? The computer
scientists can teach a computer to learn a complex thing without us teaching them
how to do it. We don’t have to understand something for a computer to understand
it, and they understand it. Computers understand things totally differently
than humans do. What’s worrisome about that is that computers are more
intelligent than people now, but they’re not conscious, and that separation of
consciousness from intelligence is a worrisome thing for the the fate of our
planet, I believe, and that’s what I was trying to allude to in talking about
what would the ethics be for automated decision-making in an urban planning, in
an urban planning system. That might sound like an absurd scenario, but it is
not at all absurd. I mean, that is totally plausible,
and I can–I predict we’ll see cities some places in the world doing that. I
can imagine Singapore doing that, for example. It depends on the culture and all.
But I think we’re going to have to–my guess is, and I’m an observer on this I’m
not an expert; I’m an observer trying to make sense of this–my sense is it’s inevitable that expert people are going to be working side by side with
computers that are smarter than they are, and human beings have to be
knowledgeable enough, agile enough being able to continue to make
themselves valuable. [Mitasova] [inaudible] Computer might not be able to make the right decision because it’s based on fast data rather than the whole system. [DiBiase] Yeah, I’m not sure that’s quite right, but then again I
don’t claim to be an expert. I claim to be a very interested
observer who’s reading a lot. I think what you’re saying about past basis on past
data that that refers to simulation but not the machine learning. Machine
learning–the computer teaches itself. If a solution is defined as optimal,
and then the computer teaches itself how to reach that solution without specific
guidance from people. [Mitasova] But it teaches itself based on some data that you give it. [DiBiase] Well, an outcome that we give it, yeah. [Mitasova] So if we give them only a limited–or if the computer collects only a limited set of data, then their machine learning will be– [DiBiase] But what is limited? I mean, how did IBM’s Watson beat the two best jeopardy
players ever? And the way way Watson beat
the two best Jeopardy players ever is that it had access to all the data in
the world. All of it. The way it studied, the way it crammed for Jeopardy
was to study the entire internet. Everything. And it could do that, in like
in a few months time it could learn all of that stuff and then be able to
recall it in order to beat human experts in Jeopardy. [Mitasova] But what if that data is not available and it cannot search it? [DiBiase] Well, then humans are in as much trouble
as computers are. [Mitasova] You had there some of the occupations that will be valuable are, for example, design or research. Those are really the occupations where people need to be creative. So they are really going beyond what is known. So, I wonder, whether the creativity that hopefully we have can save us, and hopefully make us smarter than computers? [DiBiase] Thats tricky, because in terms of creativity is not, is no longer
a good metric. The computers now can compose music that is–that cannot be
discriminated from music of the quality of Bach and Mozart. That’s just true.
So we really can’t say that it’s not creative jobs that
humans are going to be good at, because computers can learn to do
too. [Mitasova] I will let others… [DiBiase] Hi, what’s your name? [Nick] Nick. [DiBiase] Hi, Nick. [Nick] I’m interested to see the direction of technology and get your opinion where you see ESRI fitting into this progression. I think ESRI has largely moved toward the cloud, has moved towards big data analytics, but has largely created infrastructure for a human in the middle, right, to run a lot of this stuff. Does ESRI foresee that trend continuing? [DiBiase]The hottest hiring
area, personnel hiring area, in ESRI right now is machine learning. Now, the
the concern–um, so I’ve spoke just in the past week, I’ve spoken with the new hot
new machine learning folks, and their concern is very much contextualized
within data science. Their concern is that GIS generally–not just ArcGIS but GIS generally–is seen as having limited utility in data
science, and that utility is limited typically to visualizing end results. And
what ESRI is trying to do is to make the case that GIS is relevant throughout the
process. So a technology like Geo Event Server that’s monitoring millions of
sensor events in real time, that’s part of ESRI’s answer to IOT, and
machine learning is a big part of it too. At this point I think, I don’t know that
you can say that–I’m not aware that ESRI has a strategy for machine learning.
I’m just I am certain–I report to the guy who is responsible for this–I’m
certain that ESRI understands we need to be relevant to AI and machine
learning in order for GIS to remain a brand. And the incentive there is
that there are thousands and thousands of people whose livelihood depends on a
thing called GIS, and if that thing is no longer a thing, what happens to all those
livelihoods? Hey, Ross. [Meentemeyer] I really like the slide with the Venn Diagram. I had never seen that before. And I think that’s the way we’re training a lot of our students. [DiBiase] Yeah, this is
very clever. It’s kind of a tongue-in-cheek thing. It’s a response to
somebody, a blogger by the name of Drew Conway, and Drew Conway has
posts on what data science is about. And this guy doesn’t agree with Drew Conway,
so he says well Drew Conway’s data scientist is there, and so look at here’s
the intersection of business and programming, that’s the IT guy, right? And just statistics as the data nerd. Just communication is hot air. This is
really fun, right? It’s really fun, yeah. Well there’s the Comp Sci prof.
But look, it’s–so it’s the intersection of those four is what this
guy calls the perfect data scientist. Yeah, and he calls it in terms like
statistics and programming and business, but you should clearly be able to see
how that’s relevant to what you do, what you do here, right? Communication
obviously. Business I think just means knowing what your organization is trying
to accomplish and how you can contribute to that end. Right, so it doesn’t
have to be in business, although data scientists now are busy with what?
They’re busy wrangling finance data to try to make sense of the billions,
sometimes trillions, of transactions per second that Wall Street generates, and
trying to make sense out of that stuff is–and machine learning is already
at work at that, right? In fact, we have– we have to regulate how much computers
can do on their own, because they’re, they get so feisty if they if you
just turn them loose and let them make decisions autonomously. [Meentemeyer] But some other domains, the business circle could be the why. Why does this matter. [DiBiase] It’s the why, yeah. So if you
were working for a non-profit, or an NGO, in Africa, that organization has
organizational goals, or if you’re working for a government agency in
Raleigh, the business of that organization is to serve the public, or
whatever it might happen to be. So yeah, I think you can generalize these into the
kinds of things that you’re doing here. What I–one of the things I like
about this diagram is that the hard skills are balanced by what some call
the softer skills, right? And, but they’re balanced perfectly–it’s symmetrical,
right? These are just as important, these two, as these, and in fact that
geospatial technology competency model makes the same argument: that
communication abilities, the abilities to work in teams are every bit as important,
if not more, than the hard skills. So if you can pull off all of that,
you’re adding a lot of value to society. I think, I think my friend is up
next. [Audience member] So, where do you define visualization. Somewhere in communication or programming? [DiBiase] Well, visualization, you said? Yeah, so I
think of visualization as analysis personally, but when I–so the course
that I talked about at the beginning then I just started off–it’s, that course
has–so that course has a set of outcomes that it’s supposed to achieve. I
think there four of them. I should know this, right? But one of them
is explicitly the ability to tell stories with maps. So not just
communication–not just being able to read a paper and write a paper, but to be
able to tell a story and convince somebody that something matters.
That’s an explicit outcome of this. And I think that whether–it doesn’t have
to be at the level of an individual course–but I think any program, any
academic program, should have a very clear sense of why this matters and how
it’s going to make, how it’s going to make people matter, which in the end
I think I tried to pivot this so that you saw it really the technology isn’t the
point. It’s how do people stay valuable. How do
people stay part of the game. [Audience member] So, I think of visualization as part of communication. [DiBiase] Fair enough. So in my early days–in my early days I
was trained as a cartographer at University of Wisconsin long long ago.
And back then cartography was conceived as a communication, a
communication medium. And so the communication and Weavers communication
model was what was the framework for much cartographic research, beginning
with the guy who established my program, Arthur Robinson and others. And the
1988 publication of that scientific visualization study–this study that
came out of NSF–that really opened a lot of people’s eyes to the possibility that
cartography doesn’t need to be just about communication at the end, it can be
about exploration and analysis at the beginning. And that
insight led to a rebirth in cartography I think, at least from the
perspective I was working in at the time. So I completely agree that it’s both, it’s made full circle. Hi, what’s your name? [Vukomanovic] Jelena Vukomanovic. [DiBiase] Jelena,
pleased to meet you. [Vukomanovic] I found what you were saying about the ethical decision-making so interesting. And not just for driverless cars but for data… [DiBiase] Anything autonomous, right? Anything
autonomous. [Vukomanovic] It’s just such an interesting question. I can see as a society we could possibly come up with rules. [DiBiase] There are people in Google working on this right now,
today.They’re developing this stuff with with driverless cars. [Audience member] What are the rules? [DiBiase] Imagine the work that you all do. I mean,
driving is one thing, right? There’s a pretty obvious good outcome and bad
outcome in driving. But imagine something as complex as what you all
study with natural systems and human systems, and with cities and
environments, and who the heck is going to come up with ethical rules to govern
an autonomous system? And I just don’t buy the argument that “Well, it’s
impossible, it’s too hard, you can’t do it.” They’re going to do it. I mean, not
everywhere. Not everything that can be automated will be automated, but lots of
things will be. And there there’s some big business interests at stake to see
to it that the latent value in automation is realized. [Meentemeyer] Right, but with sustainability science the role of optimization is falling out of vogue. We’re thinking now of tradeoffs. And stakeholders have different values. [Vukomanovic] But having to make those things explicit, what is valuable… [DiBiase] What could go wrong? [Vukomanovic] What gets traded off at the expense of something else. There’s an interesting democrative decision. It’s a process.
[DiBiase] Let’s see. These, these guys here that came up with this list of
future roles–that Richard and Daniel Susskind–this is a really good read. Well,
okay I’ll be honest: I listened to it as a book on tape. It was a good listen, but
then I went and got the hard copy afterward so I could write on it. But
they actually come up with that in the end. Okay, what should be automated and
what must not ever be automated? And they said–their stake was that any
decision that resulted in loss or potential loss of human life should
never be automated. That is, a human being should always intercede. And they talked
about exceptions to that, like for example, how about drone raids in the
hills of Pakistan. People get killed without human intervention there. But for
example, they believe that not only will physicians be displaced by nurse
practitioners and highly capable systems, they believe that should happen
not just that it will. They believe it should. But they also go on to say that
no life-and-death decision about a human being should ever be left to a machine.
So they do draw a line, it’s just a pretty far line. Hi, what’s your name? [McCord] Hi, I’m
Marian McCord. Isn’t it already–aren’t machines already making decisions that affect whether people live or die? [DiBiase] More than we probably know. [McCord] Yeah, just even think of 911, and the way that calls come in and you can think about how they’re dispatching, and how that’s going to become more automated. It’s not a direct… not, you know, I’m a surgeon working in an operating room where the person’s life is literally in my hands…literally, but figuratively. You know, any decision that’s in routing or traffic affects whether one person lives or one person dies. So I think that line is very fuzzy, and we’re all over the place with it already as we become more automated with transportation. [DiBiase] I think you’re so right, and there’s no
turning back. [McCord] Right. To some extent, we surrender a little bit more every day, and we should, but where’s the line? [DiBiase] Hi. Good to see you again. [Audience member] Good seeing you. From an educational standpoint, do you have recommendations or examples of how you would introduce someone coming in as an undergraduate– how do we, after 4 or 6 years, help them be ready for the machine learning and data sciences positions that are out there? [DiBiase] Right. Well, I mean just the starting
point: I’ll talk about the class that I’m in, and this is a class for probably the
same ones you attract in your master’s and certificate program–very young
adults looking to advance their careers, some with background, some not so much.
And so the main thing I try to do in this course was to actually balance the coursework
so that it was one-third positioning in data acquisition, one-third analysis and
modeling, and one-third coding and app development. To echo the the
technology competency model. And very few students had any experience with coding
or app building at all. They were scared of it, and so we just started easy. Like,
here’s this much code. Here’s this much code, and what do you think it does? What
do you think happens if you were to run this much code? And start really, really,
really, really simple and and just get people aware of some of the terms
and then I think you have to keep building up on that. I think you guys are–
from what I can tell, you’re already doing a pretty darn good job at that. But this is not the status quo. GiS still–look at a GIS textbook
and see how much coding is in there. There isn’t any.
How much IT is in there? Not very much. Lots of analysis, a fair bit of data, but
the third part of that isn’t present. And if we ignore that, I think GIS as a
field as it is in peril, because you we are–this field is, the workplace is,
converging on data science and we’re expected to be able to wrangle data,
analyze data, and create apps that that share it. I was really thrilled to
come here and see you guys are doing that. I’m trying to make that–you know, at
Penn State there, they’ve got half a dozen great programming classes of all
kinds, but they’re all advanced specialties. Students don’t hear anything.
They go through the whole certificate program, never think about coding. Never
even comes up, and that was the way I felt I had to change that, try to change
the culture that this is part of this work, right? And if you don’t want
to know something–I talked to, interviewed a bunch people,
particularly people who employ–if you’ve ever heard of a company called Blue
Raster–that’s the kind of place that any GIS student would want to go work at.
They’re really, really super innovative, wonderful people. And the guy that
started that company up said to me, “Any student who’s not doing Python is headed
for a dead-end job now.” That was an overstatement for dramatic effect, but I
think it was fundamentally true that, you know, coding’s just got to be part of
what we do as educators and as students. [McCord] So, you talked about professions and changes and what we’re doing here at NC State and other programs to prepare students. One of the things that I’m most
interested in–it really fascinates me–is what is the future for higher education?
So we work on strategic plans here at the university–looking five, ten years out. And we look at things like increasing graduation,
decreasing the time it takes to graduate students. And the numbers infinitely marching up of Ph.D. students, etc. And what concerns me is could be in five years, the university ceases to exist in the way that we know it and who’s really preparing for it? So I wonder in your disappearance
of the role of experts, how that impacts institutes of higher education? Is the goal of the faculty member as we know it going to disappear as well? [DiBiase] Yeah, it’s highly problematic. My friend,
David Howard here, we know each other from long ago. He works over at DELTA,
right. So we were together at Penn State many moons ago, and
we had a conversation today about a book that I read that envisions a future
for education that was the first really plausible revolutionary vision. It’s a
book called Junana, which is easy to get hold of. Junana. Yeah,
traditional educational institutions are pretty much left behind unless they find
supporting roles. Now I’m not predicting that. I’m just
saying that’s what I read in the book, Junana. I, for one, work for a university
and hope that it will continue to be a viable role for me, but I do
know this: I didn’t talk about pedagogy for this class, but I took the step
of stopping teaching and it’s a true constructivist environment in
which students are responsible for their own learning. I tell them what their
objectives are and I don’t provide any content, other than talks like this.
They have got to go out and research it themselves, share what they learn through
discussion, and then demonstrate what they what they do. And they don’t like it,
but it sure is a better class than it was. [Audience member] They don’t like it? [DiBiase] They don’t like it. They want to be taught. But I think it’s a
fabulous question. So I told David this morning that
fifteen years ago at Penn State I was fortunate enough to be on a little
panel that somebody, some education administrator put together, that said,
“Okay, I want you guys to go away and envision what education will be like
20 years from now.” What a fantastically fun and useless exercise, right? Because
nobody’s going to pay attention to that, right? But we got to do it, and so
what we imagined was that in 20 years time, that students would come to
university and their orientation would include them meeting their personal
digital assistant, that their personal digital assistant would take care of
tasks like helping them navigate the library, helping them find and evaluate
we sources, all of the scut work that students have to do this personal
digital assistant would do it for them. And that would now be called
Cortana. If you take the time to train
Cortana. In other words, that’s happened, and not that I was visionary–
just that’s what happened and it is happening. And I think, though, if you
read the book Junana, it’s a book about what happens when we have fully
realized digital assistants who have machine learning capabilities, and who
know you better than you know yourself. All of which can be hacked, which makes
this a really fun. All of it can be hacked. [Meentemeyer] Another question, David. So, in the list of future roles, outlined by the Susskinds–should we be concerned that scientists are not there? Or do they fall under this new category of scientist, the data scientist? [DiBiase] Yeah, scientists don’t appear there.
Acientific applications don’t appear in that IOT map either. [Meentemeyer] Do you think that’s realistic? [DiBiase] Well, I think that
science isn’t an industry. Science consumes resources; it doesn’t create
them in terms of creating direct revenue, sales, and economic impact. [Meentemeyer] But you might say that the whole Triangle region– [DiBiase] That’s a very good example. Point taken. Point taken. But that is missing from the
IOT map that I showed you. In other words, that company, Beechum Research, doesn’t
see economic opportunities at scale. Right. Do you have a question? [Audience member] What is an empathizer? Caregiver. Caregiver. So right now, right
now, the chief consumer application in robotics is happening in Japan. Japan
culturally has more of an affinity towards home robots, and the hottest and
latest thing I’ve read–the most activity going on in commercial robots
are little robots that can stay with elderly people and cheer them up and
talk to them. Because Japan has a problem with aging population, and the
whole mechanism before of that the wife stays home and takes care of Grandma–
that’s all breaking down, and so they are working on replacing that with consumer
robots that are cuddly and friendly and can empathize. That’s like, if
somebody’s sad, they can tell–the computer can, the robot can tell, and
they can do, “Oh, I’m sorry,” and “What can I do to make it better?” But that also
extends to–it also extends to, I was almost going to say nurse but the nurse
is almost too qualified–somebody who can just be beside somebody who is ill
and empathize with them, or people who are in counseling programs. And it’s
not that–it’s not that computers can’t learn to do that; it’s that it’s not
cost-effective to teach them to do that. How much do nurses’ aides get paid? It’s
not worth automating them.

Danny Hutson

Leave a Reply

Your email address will not be published. Required fields are marked *