January 6, 2025 • 19 mins
From advances in the use of artificial intelligence and new guidance into its research to breakthroughs in biological treatments and approaches to healing to how we measure time, discoveries in many areas were revealed by U.S. National Science Foundation-supported researchers in 2024. As we start a new year, we're looking back at memorable moments from some of the most popular episodes of the last year.
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
(00:03):
This is the Discovery Files podcast
from the US National Science Foundation.
As we start a new year,
we're looking back at some of last year'smost popular episodes and revisiting
some of the stunning researchthat we featured early in 2024.
We looked at several innovationswith medical applications.
One project involved investigatinghow white blood cells interact
(00:26):
with cancer cells.
We caught up with Kolade Adebowale,a post-doctoral fellow at the Harvard
John A.
Paulson School of Engineeringand Applied Sciences,
who was workingwith attaching a small backpack to cells
that could be used to deliver treatmentsto cancer cells.
As you can imagine, we have this particlethat's attached to the surface of a cell.
One can envision a scenariowhere we incorporate
(00:47):
some drug of interestin this microparticles.
This can do actually therapeutic drugreally like a cytokine.
It can be whateverdrug of interest to you.
So that's one reasonone might use this solar backpack.
But even beyond that,it appears as though the process
of attaching this microparticlesto some kind of cell.
So for example, things like natural killercells, which are a type of white blood
(01:11):
cell, or T cells, which are another typeof white blood cell.
Just as physical attachment,it appears that snow is able
to activate these cells, which meansthat they have an enhanced ability.
It almost endows that with some kind of
and super powersto be able to fight cancer specifically.
Amber Doiron, assistant professorin the Department of Electrical
and Biomedical Engineeringat the University of Vermont,
(01:33):
told us about her work developing biofilmresistant wound dressings.
Biofilms are basically a different wayfor bacteria to grow.
So bacteria that we're all usedto envisioning, we kind of think of them
as floating around or swimming,you know, in some sort of, liquid.
But they can also attach to a surface.
And when they attach to a surface,they kind of create this really nice,
(01:55):
happy home for themselves as environmentthat's protective for them.
And that surface can be something like athe surface of a wound.
If you have an injury and it's healing.
The bacteria can attach to itand adhere to that surface
and create a protective environmentfor themselves, which is not necessarily
(02:15):
good for the personwith that healing wound.
They can also attach to surfacesof medical devices, things like catheters,
heart valves, things like thatthat are indwelling or inside the body.
So biofilms present a really large problem
for treatmentwhen the bacteria attach to that surface.
They're just much harder to treat.
(02:36):
They create kind of a slime timeenvironment for themselves to live in.
That makes it hard for antibioticsto get to the bacteria, makes it hard for
cells of our immune systemsto get to the bacteria.
Wound dressings, you know, historicallyobviously have been used for a long time,
maybe just a piece of fabric or,when gauze was developed, you know,
(02:56):
that was, used as a wound dressing
for a very long period of timeand still is more advanced
wound dressings have triedto create environments of that wound.
So if we're talking about a topical wound,say, someone had, burn injury
or a traumatic injury, the skin is broken,and that creates a place where bacteria
and other environmental factors
(03:17):
can come into the body, creatinglots of problems for that patient.
And so a wound dressing createsa protective
barrier between the outside of the bodyand the inside of the body.
And that also is providing, a healthy environment for healing.
But they also havea high incidence of themselves
(03:38):
becoming colonized with bacteria.
So what we've tried to dois create a wound dressing
that essentially is an anti biofilmwound dressing,
where it inhibits the growthof the bacteria both on the wound
dressing and helps the wound itselfavoid the bacteria as well.
So we're trying to help the patient heal
(04:01):
by preventing the wound dressing itself
from becoming attached to bacteria.
Our approach to this relies on eliminating
what essentially is an energy sourcefor the bacteria.
And it turns out that that helpsto inhibit these biofilms.
It actually can both prevent the biofilmsfrom starting
(04:25):
and treat the biofilmsif they're already established.
So by incorporatingsome of those approaches
into our wound dressing,we can cut down on the bacteria,
even being able to start a biofilmon the wound, dressing itself and help
treat the potentially already estab,lished biofilm that's in the wound.
(04:45):
Dmitry Kireev,
assistantprofessor of biomedical engineering
at the University of MassachusettsAmherst,
joined to discuss developing biosensorsusing graphene.
It is the.
Electronic propertiesthat this combination
of atomic distribution of the materialthat makes this unique sense.
But then it's alsonot only about conductivity itself, it's
(05:06):
about sensitivityto external feel to to the surroundings.
You know, I have this example.
You have somethinghard. Let's say you're cooking, right?
You put your hand on and you can,if you tried to sense something,
you would on this sensewith one side of your hand, right.
Because it's thick. It's not one. Nothing.
You don't sense anythingwith the opposite side of your hand.
(05:26):
But if your hand was one atom thick,you would sense it.
You know much better. Faster.
You know thiswith a greater level of sensitivity.
And that is the case with grapheneand 2D materials.
All of these atoms, the whole surfaceis exposed to the surroundings.
That makes it amazing. Sensor.
(05:47):
We learned about how clocks work and whyprecise measurements of time are important
from Eric Hudson,
a professor in the Department of Physicsand astronomy at UCLA
who has spent the last 15 years workingto accomplish the seemingly impossible
directly manipulating the energy levelof an atomic nucleus using a laser
something that has never been done beforeand could unlock the most accurate
(06:08):
clock ever.
We all have an inherent understandingof what a clock is.
They're very important in our lives,
but we don't really think about them toooften.
What they're doing is pretty profound.
Evidently, there's something we all agreeon called time.
The time is later nowthan when we started this discussion.
But how do we measure that?
How do we agree upon how much timeit's passed and what we do with clocks?
(06:32):
We typically have some sort of processthat repeats itself,
and we assume that that repetitionis sort of constant in time.
So think of a grandfather clockwhere you have a pendulum swinging
back and forthafter one oscillation back and forth.
We call that one second.
You can let two oscillationsbe two seconds, 60 be a minute, and so on.
The wristwatches or our phonesthat we use, they typically have like a
(06:55):
something very similar to this.
It's a piece of quartz that vibratesat about 30,000 times per second.
You could see a lot of detailsin that analogy
that are going to be important
for understanding atomic clocksthat eventually nuclear clocks.
The other thing that's importantis that our clocks be identical.
If you and I build grandfather clocks,and I make my arm
(07:15):
a little bit longer than youor my bearings, maybe not as good as you.
There's a little more friction, right,than my clock will run slower than yours.
And so after some time, we no longer agreeon how much time has passed.
And I'll show up for this interview lateand that is where atomic clocks come in.
As far as we know,
every, let's say hydrogen atom inthe universe is identical.
(07:35):
What that means is the absorbing emitlight at exactly the same frequencies.
Right?
So we know you can shine in a laser,for example, and excite an electron
in an atom to a higher orbital.
And that happens at exactlya certain color of light.
And it's the same here as on the moon,as in the Andromeda Galaxy.
If I make sure my laser can always drivethat transition.
(07:57):
Now, my laser,the electro magnetic field of the laser
is oscillating at a frequencythat's tied to the structure of hydrogen,
which is tied,which is set by the laws of the universe.
And so my laser now becomes my oscillator.
And if I count the oscillations
of the electromagnetic field,I can use that to keep time
(08:18):
that can be very high frequencyand optical atomic clocks.
It can also be identical because allthe atoms are the same of the same type.
So our clocks are guaranteedto tick with the same frequency.
And importantly,because we have techniques now
where we can isolate these atomsfrom their environment,
we don't have this problemof like friction in the pendulum clock.
(08:41):
Can you.
But like so these things are going to beperturbed by their environment as much.
And so we can build these, you know,amazingly accurate clocks that people
at precision timekeeping is actuallyreally important in our daily lives.
We don't normally think about this.
You know, if you're a minutelate to a zoom meeting, it's it's okay.
But your cell phones,your global positioning,
(09:01):
lots of communication technology,some imaging technology,
you really have to havethe most precise clock you can get.
And if you had better clocks,you could do those things better.
Like if we had betteratomic clocks, GPS would be better
that, you know, that is important for,you know, navigation here on Earth.
But also as you go into the solar systemand beyond this sort of deeper
(09:24):
space navigation,this becomes a really big problem.
Technological advances, including machinelearning and artificial intelligence
applications, continueto be popular subjects.
We spoke with Miyoko Chu,
senior director of communicationsat the Cornell Lab of Ornithology,
who was involved in developingthe popular Merlin Bird ID app,
which helps users to identify birdsin the wild from their mobile devices.
(09:47):
The origin of Merlinwent back to about 2009.
Our team here at the Lab of Ornithology
was was wonderinghow can we find some funding
to do more to engage peoplein our websites about birds?
So we sat down and we had heardabout this NSF opportunity,
(10:08):
a grant that could be availablein informal science education.
And we said, what could we dothat could be valuable?
And we noticed thatwhen you look at our All About
Birds website, which is like the onlineencyclopedia of birds,
you looked at what people weretyping into the search box.
It was things like little brown bird
(10:30):
with white line, red bird with a crest,you know.
So there were these descriptorswere clearly people were trying
to identify birds using our websiteas a bird identification tool.
But that's not the most useful toolto help you with birds.
Although if you spent a while,maybe you could figure it out.
So it happenedthat one of the developers on our team
(10:54):
said, I've seen this online gameabout 21 questions where the computer
guesses what you're thinkingbased on clues that you provided.
And that was just at the beginningof AI really taking off.
Like, I hadn't really seen applications ofAI before, but we said, wow,
computer can guess what you're thinking.
(11:15):
And 21 questions.
That sounds neat.
Can't we do that for birds?
So that was the start of the idea,the nugget of the idea.
I have to say, I'm completely astounded
just by how many peoplehave started to use it, but the reaction
has just been delight and a sense of,I don't know, new possibilities.
(11:36):
Because as soon as people realizethat they can use this app
to give them the name of somethingthat they've seen
or heard,it just becomes much more tangible.
Once you have the name, you can learn
all kinds of interesting facts about it,so it really opens a door.
Tanya Berger-Wolf, a professorand computer scientist at The Ohio State
(12:00):
University, joined us to share
how she is using technologies,including artificial intelligence,
to extract information from imagesin the new field of images omics.
So it's the omics of images, images,comics, image comics.
Right?
This is the field of discoveringbiological information
from images, particularlyfocusing on traits and phenotypes.
(12:23):
So traits are the things that we considerimportant about animals.
We can observe and we can connectto function, often through genotype.
So things that started with Mendel'sdiscovery of the size, color, shape
of the peas and the plantand the peapod and the pea flower.
(12:45):
And that's how we started on the pathway
to genetics and then genomics.
But traits can be at the individual level.
So the things like color, shape,
size, weight, height,the pattern of the zebra stripes,
they can be at the population level,the variation of the colors and shapes.
(13:06):
But they can also varyduring individual lifetimes.
Things like behavior or pregnant, sick,
healthy, juvenile, adulthave different proportions.
Things like the fruit of plants.
So not only the animals have traits,certainly plans as well, like peas,
but the width and length of the leafas it grows.
(13:27):
The ratios of all these as the nutritionand the soil changes
and the amount of sunlight changes.
All of these are traits.
And then the product of genes, environmentand the combinations of the two.
Now the traits are
not just importantto observe and understand,
but it is important to connect themto function and the content
(13:49):
and the understanding of why
not only what we're seeing,but why is it changing?
Why do animals have traits?
What is the intra speciestrait variability
within the species, and how is itrelated to the variability in the habitat
or the conditions of the species
along a particular gradient?
(14:12):
And so as humans, we're very biased
by what we look at, what we can see,and what we consider
important in terms of traitsand everything else for that matter.
There's so many thingsthat we're not even capable of seeing,
because our apparatus that we're using tosee is highly unreliable and is limited.
(14:33):
For example, we don't have enough acuityto distinguish red orange variability
in polymorphic moths because that didnot evolve for our benefit.
That variability evolved for birds.
Right?
To fool the birds in the end and to signalsomething to their own species insects.
Our vision is very, very different.
(14:54):
They're also traits that are not evenin the visual spectrum at all.
There is chemical, the acoustics,and all the other
kinds of sensory channelsthat we're not great at perceiving.
And so we're missinga whole bunch of the world.
And that
translates into missing understandingof a whole bunch of the world.
(15:16):
So where technology comesin, it makes the invisible visible,
because we're not only can bring in imagesfrom the same spectrum
that we humans are seeing,but from a whole other range of spectra,
and we can make sense of that.
We can extract information from that,and then connected
(15:39):
to a whole other wealth of datathat then really
not only finds new traits,
connects them to different contexts,
and startsthrough these patterns of connections,
understand the function of these traitsand how they evolve, why they evolve,
(16:00):
and what is their functionin the also environmental context
and the ecological contextand understanding the truly.
Maybe the life on earth.
We also explored an inter-agency effortthat resulted in the Global Artificial
Intelligence Research Agenda, a documentcrafted to serve as a starting point
to align a global research visionwith Michael Liittman,
(16:23):
the Division Director of Informationand Intelligence Systems
in the Computer and Information Scienceand Engineering Directorate at the U.S.
National Science Foundation.
So it's worth pointing outthat the idea for the ERA was first
made public in the Executive Orderon Artificial Intelligence,
which the Biden administrationissued in last October.
So we're getting really closeto the one year anniversary of it.
(16:45):
In that document,there was a whole slew of things that were
that were requested of the I guess,demanded of the agencies to actually do.
Some were targeted at particular agencies,some were targeted to all the agencies.
This particular item, it mentionedcreating a global AI research agenda.
So which we've been calling Gaira.
But it doesn't actually say Gaira in thein the document.
(17:06):
We now know that this is actuallythe name of a monster from Godzilla.
So that was you know, for spiteful.
It's not a monster. It'sactually a really great thing.
So the idea of it is that thethe executive order said
that the State Departmentand US aid agencies in
the US would put together a globalAI research agenda.
Basically, what should we all be doingtogether to help move AI forward in a way
(17:29):
that advances artificial intelligencefor the benefit of the entire planet?
The authors of the executive order,
they felt that the State Departmentand U.S.
aid have the expertiseto really connect a US vision to to help
turn it into a global visionto really connect
with our partners in other countries.
They also asked the Department of Energyand the National Science Foundation
(17:49):
to be a part of the developmentof the research agenda because, well, AI
research really happens at these agencieswithin the federal government.
They also ended up bringing in
or we also ended upbringing in the Department of Labor
to help out with the issuesthat have to do with the way that AI is
being developed worldwide involvesemploying a lot of different people
to do a lot of different kinds of tasks,and this is not happening within
(18:11):
any given one country.
This is actually happeningacross the world.
And so a lot of the decisionsthat get made in terms of how AI is
pursued, actually have impact on workerseverywhere.
I think it's really important
to point out that in terms of AI research,I've been an AI researcher.
I don't know when you officially become anAI researcher.
There's no ceremony for that.
But for me,my first research paper was in 1989,
(18:33):
and I went to a conference,
and it was a global conferencein the sense
that there were researchersfrom all over the world.
And so my entire career inAI has been interacting with
not just folks in the USdoing AI research, but folks everywhere
in the world doing AIresearch is a global academic community.
And I think recognizing that, thatthis isn't there, we're not making these
unilateral decisionswhen we're talking about AI research.
(18:55):
We need to be working togetheras as a team.
That's that, to me, is the impetusfor doing something like this.
And so I think this is kindof our offering out into the world saying
this is how we're thinking about this.This is how we'd like to work with you.
Let's get this conversation going.
Special thanks to everyone who took timeout of their schedules to join us
and share stories
about their amazing breakthroughs in 2024,and our thanks to you for listening.
(19:18):
For the discovery Files, I'm Nate Potter.
You can watch video versions
of these conversations on our YouTubechannel by searching @NSFscience.
Please subscribe wherever you get podcastsand if you like
our program, share with a friendand consider leaving a review.
Discover how the U.S.
National Science Foundationis advancing research at NSF.gov.
This is the Discovery Files podcast
from the US National Science Foundation.
As we start a new year,
we're looking back at some of last year'smost popular episodes and revisiting
some of the stunning researchthat we featured early in 2024.
We looked at several innovationswith medical applications.
One project involved investigatinghow white blood cells interact
(00:26):
with cancer cells.
We caught up with Kolade Adebowale,a post-doctoral fellow at the Harvard
John A.
Paulson School of Engineeringand Applied Sciences,
who was workingwith attaching a small backpack to cells
that could be used to deliver treatmentsto cancer cells.
As you can imagine, we have this particlethat's attached to the surface of a cell.
One can envision a scenariowhere we incorporate
(00:47):
some drug of interestin this microparticles.
This can do actually therapeutic drugreally like a cytokine.
It can be whateverdrug of interest to you.
So that's one reasonone might use this solar backpack.
But even beyond that,it appears as though the process
of attaching this microparticlesto some kind of cell.
So for example, things like natural killercells, which are a type of white blood
(01:11):
cell, or T cells, which are another typeof white blood cell.
Just as physical attachment,it appears that snow is able
to activate these cells, which meansthat they have an enhanced ability.
It almost endows that with some kind of
and super powersto be able to fight cancer specifically.
Amber Doiron, assistant professorin the Department of Electrical
and Biomedical Engineeringat the University of Vermont,
(01:33):
told us about her work developing biofilmresistant wound dressings.
Biofilms are basically a different wayfor bacteria to grow.
So bacteria that we're all usedto envisioning, we kind of think of them
as floating around or swimming,you know, in some sort of, liquid.
But they can also attach to a surface.
And when they attach to a surface,they kind of create this really nice,
(01:55):
happy home for themselves as environmentthat's protective for them.
And that surface can be something like athe surface of a wound.
If you have an injury and it's healing.
The bacteria can attach to itand adhere to that surface
and create a protective environmentfor themselves, which is not necessarily
(02:15):
good for the personwith that healing wound.
They can also attach to surfacesof medical devices, things like catheters,
heart valves, things like thatthat are indwelling or inside the body.
So biofilms present a really large problem
for treatmentwhen the bacteria attach to that surface.
They're just much harder to treat.
(02:36):
They create kind of a slime timeenvironment for themselves to live in.
That makes it hard for antibioticsto get to the bacteria, makes it hard for
cells of our immune systemsto get to the bacteria.
Wound dressings, you know, historicallyobviously have been used for a long time,
maybe just a piece of fabric or,when gauze was developed, you know,
(02:56):
that was, used as a wound dressing
for a very long period of timeand still is more advanced
wound dressings have triedto create environments of that wound.
So if we're talking about a topical wound,say, someone had, burn injury
or a traumatic injury, the skin is broken,and that creates a place where bacteria
and other environmental factors
(03:17):
can come into the body, creatinglots of problems for that patient.
And so a wound dressing createsa protective
barrier between the outside of the bodyand the inside of the body.
And that also is providing, a healthy environment for healing.
But they also havea high incidence of themselves
(03:38):
becoming colonized with bacteria.
So what we've tried to dois create a wound dressing
that essentially is an anti biofilmwound dressing,
where it inhibits the growthof the bacteria both on the wound
dressing and helps the wound itselfavoid the bacteria as well.
So we're trying to help the patient heal
(04:01):
by preventing the wound dressing itself
from becoming attached to bacteria.
Our approach to this relies on eliminating
what essentially is an energy sourcefor the bacteria.
And it turns out that that helpsto inhibit these biofilms.
It actually can both prevent the biofilmsfrom starting
(04:25):
and treat the biofilmsif they're already established.
So by incorporatingsome of those approaches
into our wound dressing,we can cut down on the bacteria,
even being able to start a biofilmon the wound, dressing itself and help
treat the potentially already estab,lished biofilm that's in the wound.
(04:45):
Dmitry Kireev,
assistantprofessor of biomedical engineering
at the University of MassachusettsAmherst,
joined to discuss developing biosensorsusing graphene.
It is the.
Electronic propertiesthat this combination
of atomic distribution of the materialthat makes this unique sense.
But then it's alsonot only about conductivity itself, it's
(05:06):
about sensitivityto external feel to to the surroundings.
You know, I have this example.
You have somethinghard. Let's say you're cooking, right?
You put your hand on and you can,if you tried to sense something,
you would on this sensewith one side of your hand, right.
Because it's thick. It's not one. Nothing.
You don't sense anythingwith the opposite side of your hand.
(05:26):
But if your hand was one atom thick,you would sense it.
You know much better. Faster.
You know thiswith a greater level of sensitivity.
And that is the case with grapheneand 2D materials.
All of these atoms, the whole surfaceis exposed to the surroundings.
That makes it amazing. Sensor.
(05:47):
We learned about how clocks work and whyprecise measurements of time are important
from Eric Hudson,
a professor in the Department of Physicsand astronomy at UCLA
who has spent the last 15 years workingto accomplish the seemingly impossible
directly manipulating the energy levelof an atomic nucleus using a laser
something that has never been done beforeand could unlock the most accurate
(06:08):
clock ever.
We all have an inherent understandingof what a clock is.
They're very important in our lives,
but we don't really think about them toooften.
What they're doing is pretty profound.
Evidently, there's something we all agreeon called time.
The time is later nowthan when we started this discussion.
But how do we measure that?
How do we agree upon how much timeit's passed and what we do with clocks?
(06:32):
We typically have some sort of processthat repeats itself,
and we assume that that repetitionis sort of constant in time.
So think of a grandfather clockwhere you have a pendulum swinging
back and forthafter one oscillation back and forth.
We call that one second.
You can let two oscillationsbe two seconds, 60 be a minute, and so on.
The wristwatches or our phonesthat we use, they typically have like a
(06:55):
something very similar to this.
It's a piece of quartz that vibratesat about 30,000 times per second.
You could see a lot of detailsin that analogy
that are going to be important
for understanding atomic clocksthat eventually nuclear clocks.
The other thing that's importantis that our clocks be identical.
If you and I build grandfather clocks,and I make my arm
(07:15):
a little bit longer than youor my bearings, maybe not as good as you.
There's a little more friction, right,than my clock will run slower than yours.
And so after some time, we no longer agreeon how much time has passed.
And I'll show up for this interview lateand that is where atomic clocks come in.
As far as we know,
every, let's say hydrogen atom inthe universe is identical.
(07:35):
What that means is the absorbing emitlight at exactly the same frequencies.
Right?
So we know you can shine in a laser,for example, and excite an electron
in an atom to a higher orbital.
And that happens at exactlya certain color of light.
And it's the same here as on the moon,as in the Andromeda Galaxy.
If I make sure my laser can always drivethat transition.
(07:57):
Now, my laser,the electro magnetic field of the laser
is oscillating at a frequencythat's tied to the structure of hydrogen,
which is tied,which is set by the laws of the universe.
And so my laser now becomes my oscillator.
And if I count the oscillations
of the electromagnetic field,I can use that to keep time
(08:18):
that can be very high frequencyand optical atomic clocks.
It can also be identical because allthe atoms are the same of the same type.
So our clocks are guaranteedto tick with the same frequency.
And importantly,because we have techniques now
where we can isolate these atomsfrom their environment,
we don't have this problemof like friction in the pendulum clock.
(08:41):
Can you.
But like so these things are going to beperturbed by their environment as much.
And so we can build these, you know,amazingly accurate clocks that people
at precision timekeeping is actuallyreally important in our daily lives.
We don't normally think about this.
You know, if you're a minutelate to a zoom meeting, it's it's okay.
But your cell phones,your global positioning,
(09:01):
lots of communication technology,some imaging technology,
you really have to havethe most precise clock you can get.
And if you had better clocks,you could do those things better.
Like if we had betteratomic clocks, GPS would be better
that, you know, that is important for,you know, navigation here on Earth.
But also as you go into the solar systemand beyond this sort of deeper
(09:24):
space navigation,this becomes a really big problem.
Technological advances, including machinelearning and artificial intelligence
applications, continueto be popular subjects.
We spoke with Miyoko Chu,
senior director of communicationsat the Cornell Lab of Ornithology,
who was involved in developingthe popular Merlin Bird ID app,
which helps users to identify birdsin the wild from their mobile devices.
(09:47):
The origin of Merlinwent back to about 2009.
Our team here at the Lab of Ornithology
was was wonderinghow can we find some funding
to do more to engage peoplein our websites about birds?
So we sat down and we had heardabout this NSF opportunity,
(10:08):
a grant that could be availablein informal science education.
And we said, what could we dothat could be valuable?
And we noticed thatwhen you look at our All About
Birds website, which is like the onlineencyclopedia of birds,
you looked at what people weretyping into the search box.
It was things like little brown bird
(10:30):
with white line, red bird with a crest,you know.
So there were these descriptorswere clearly people were trying
to identify birds using our websiteas a bird identification tool.
But that's not the most useful toolto help you with birds.
Although if you spent a while,maybe you could figure it out.
So it happenedthat one of the developers on our team
(10:54):
said, I've seen this online gameabout 21 questions where the computer
guesses what you're thinkingbased on clues that you provided.
And that was just at the beginningof AI really taking off.
Like, I hadn't really seen applications ofAI before, but we said, wow,
computer can guess what you're thinking.
(11:15):
And 21 questions.
That sounds neat.
Can't we do that for birds?
So that was the start of the idea,the nugget of the idea.
I have to say, I'm completely astounded
just by how many peoplehave started to use it, but the reaction
has just been delight and a sense of,I don't know, new possibilities.
(11:36):
Because as soon as people realizethat they can use this app
to give them the name of somethingthat they've seen
or heard,it just becomes much more tangible.
Once you have the name, you can learn
all kinds of interesting facts about it,so it really opens a door.
Tanya Berger-Wolf, a professorand computer scientist at The Ohio State
(12:00):
University, joined us to share
how she is using technologies,including artificial intelligence,
to extract information from imagesin the new field of images omics.
So it's the omics of images, images,comics, image comics.
Right?
This is the field of discoveringbiological information
from images, particularlyfocusing on traits and phenotypes.
(12:23):
So traits are the things that we considerimportant about animals.
We can observe and we can connectto function, often through genotype.
So things that started with Mendel'sdiscovery of the size, color, shape
of the peas and the plantand the peapod and the pea flower.
(12:45):
And that's how we started on the pathway
to genetics and then genomics.
But traits can be at the individual level.
So the things like color, shape,
size, weight, height,the pattern of the zebra stripes,
they can be at the population level,the variation of the colors and shapes.
(13:06):
But they can also varyduring individual lifetimes.
Things like behavior or pregnant, sick,
healthy, juvenile, adulthave different proportions.
Things like the fruit of plants.
So not only the animals have traits,certainly plans as well, like peas,
but the width and length of the leafas it grows.
(13:27):
The ratios of all these as the nutritionand the soil changes
and the amount of sunlight changes.
All of these are traits.
And then the product of genes, environmentand the combinations of the two.
Now the traits are
not just importantto observe and understand,
but it is important to connect themto function and the content
(13:49):
and the understanding of why
not only what we're seeing,but why is it changing?
Why do animals have traits?
What is the intra speciestrait variability
within the species, and how is itrelated to the variability in the habitat
or the conditions of the species
along a particular gradient?
(14:12):
And so as humans, we're very biased
by what we look at, what we can see,and what we consider
important in terms of traitsand everything else for that matter.
There's so many thingsthat we're not even capable of seeing,
because our apparatus that we're using tosee is highly unreliable and is limited.
(14:33):
For example, we don't have enough acuityto distinguish red orange variability
in polymorphic moths because that didnot evolve for our benefit.
That variability evolved for birds.
Right?
To fool the birds in the end and to signalsomething to their own species insects.
Our vision is very, very different.
(14:54):
They're also traits that are not evenin the visual spectrum at all.
There is chemical, the acoustics,and all the other
kinds of sensory channelsthat we're not great at perceiving.
And so we're missinga whole bunch of the world.
And that
translates into missing understandingof a whole bunch of the world.
(15:16):
So where technology comesin, it makes the invisible visible,
because we're not only can bring in imagesfrom the same spectrum
that we humans are seeing,but from a whole other range of spectra,
and we can make sense of that.
We can extract information from that,and then connected
(15:39):
to a whole other wealth of datathat then really
not only finds new traits,
connects them to different contexts,
and startsthrough these patterns of connections,
understand the function of these traitsand how they evolve, why they evolve,
(16:00):
and what is their functionin the also environmental context
and the ecological contextand understanding the truly.
Maybe the life on earth.
We also explored an inter-agency effortthat resulted in the Global Artificial
Intelligence Research Agenda, a documentcrafted to serve as a starting point
to align a global research visionwith Michael Liittman,
(16:23):
the Division Director of Informationand Intelligence Systems
in the Computer and Information Scienceand Engineering Directorate at the U.S.
National Science Foundation.
So it's worth pointing outthat the idea for the ERA was first
made public in the Executive Orderon Artificial Intelligence,
which the Biden administrationissued in last October.
So we're getting really closeto the one year anniversary of it.
(16:45):
In that document,there was a whole slew of things that were
that were requested of the I guess,demanded of the agencies to actually do.
Some were targeted at particular agencies,some were targeted to all the agencies.
This particular item, it mentionedcreating a global AI research agenda.
So which we've been calling Gaira.
But it doesn't actually say Gaira in thein the document.
(17:06):
We now know that this is actuallythe name of a monster from Godzilla.
So that was you know, for spiteful.
It's not a monster. It'sactually a really great thing.
So the idea of it is that thethe executive order said
that the State Departmentand US aid agencies in
the US would put together a globalAI research agenda.
Basically, what should we all be doingtogether to help move AI forward in a way
(17:29):
that advances artificial intelligencefor the benefit of the entire planet?
The authors of the executive order,
they felt that the State Departmentand U.S.
aid have the expertiseto really connect a US vision to to help
turn it into a global visionto really connect
with our partners in other countries.
They also asked the Department of Energyand the National Science Foundation
(17:49):
to be a part of the developmentof the research agenda because, well, AI
research really happens at these agencieswithin the federal government.
They also ended up bringing in
or we also ended upbringing in the Department of Labor
to help out with the issuesthat have to do with the way that AI is
being developed worldwide involvesemploying a lot of different people
to do a lot of different kinds of tasks,and this is not happening within
(18:11):
any given one country.
This is actually happeningacross the world.
And so a lot of the decisionsthat get made in terms of how AI is
pursued, actually have impact on workerseverywhere.
I think it's really important
to point out that in terms of AI research,I've been an AI researcher.
I don't know when you officially become anAI researcher.
There's no ceremony for that.
But for me,my first research paper was in 1989,
(18:33):
and I went to a conference,
and it was a global conferencein the sense
that there were researchersfrom all over the world.
And so my entire career inAI has been interacting with
not just folks in the USdoing AI research, but folks everywhere
in the world doing AIresearch is a global academic community.
And I think recognizing that, thatthis isn't there, we're not making these
unilateral decisionswhen we're talking about AI research.
(18:55):
We need to be working togetheras as a team.
That's that, to me, is the impetusfor doing something like this.
And so I think this is kindof our offering out into the world saying
this is how we're thinking about this.This is how we'd like to work with you.
Let's get this conversation going.
Special thanks to everyone who took timeout of their schedules to join us
and share stories
about their amazing breakthroughs in 2024,and our thanks to you for listening.
(19:18):
For the discovery Files, I'm Nate Potter.
You can watch video versions
of these conversations on our YouTubechannel by searching @NSFscience.
Please subscribe wherever you get podcastsand if you like
our program, share with a friendand consider leaving a review.
Discover how the U.S.
National Science Foundationis advancing research at NSF.gov.
Popular Podcasts
Dateline NBC
Current and classic episodes, featuring compelling true-crime mysteries, powerful documentaries and in-depth investigations. Follow now to get the latest episodes of Dateline NBC completely free, or subscribe to Dateline Premium for ad-free listening and exclusive bonus content: DatelinePremium.com
24/7 News: The Latest
The latest news in 4 minutes updated every hour, every day.
Therapy Gecko
An unlicensed lizard psychologist travels the universe talking to strangers about absolutely nothing. TO CALL THE GECKO: follow me on https://www.twitch.tv/lyleforever to get a notification for when I am taking calls. I am usually live Mondays, Wednesdays, and Fridays but lately a lot of other times too. I am a gecko.