# AI in Materials Science: Discovery, Data Gaps, and Active Learning

**Podcast:** Latent Space: The AI Engineer Podcast
**Published:** 2026-03-24

## Transcript

There's a school of thought that why should I bother to learn chemistry or physics or whatever when ChatGPT no as PhD level understanding of that anyway?
ChatGPT is super good at Wikipedia level chemistry knowledge.
I'm really interested in molecular design.
Like how do you find a new ligand that can go into a transition metal complex?
And what that means is it's some combination of atoms and it's gonna bind to the metal and it's gonna change its properties.
The thing I constantly do every time an LLM is updated is I just ask it, please design me a ligand that has 22 atoms.
I can never get an answer that has 22 atoms.
Hi, we're really excited to have um Heather Kulik here.
She's a professor of chemical engineering at MIT.
Uh Heather has done some like amazing work in material science and computational chemistry.
Um, but we're particularly excited to have her today because she has for almost her entire career been working on the intersection of using data-driven methods, AI, and using applying them to improve materials and under understanding of materials.
And uh, she has a lot of like really interesting opinions about what works and how do you approach these problems to get the most out of them?
So um, yeah, we're really excited to um have you here and um yeah, maybe to get started.
Can you just tell us about like one of the coolest things you've done in your opinion for a kind of an AI engineering audience?
Yeah, so uh my my group, we work a lot in accelerated discovery of new materials.
When I first started out, we were just really using AI to make predictions we'd normally make with computational models, just make them faster.
But the question I would often get when we were doing that was, okay, but what's what's surprising?
What's what's sort of something from AI that like I wouldn't have already known if I were a really smart chemist or a really smart material scientist?
And you know, you make all these computational predictions.
Has anyone actually made in the lab something that you predicted?
Recently I was I was able to do a really nice demonstration where the answer to both of those questions, you know, was very clear from the work.
So we were able to screen with artificial intelligence a set of uh thousands, tens of thousands of materials where uh each individual experiment, if it were done in the lab, would have taken months to years.
And through AI, we uncovered this sort of unexpected chemical phenomenon that led to a emergent property in in what's known as a polymer network, so plastics, um, that would make the polymer about four times tougher.
And when we showed uh the design that AI had come up with to the experimentalists, they were really surprised.
They would have never come on this on their own.
Um, and then we were able to convince them to make it in the lab.
And in fact, it was it was this tougher material.
And where this has applications is if we can make plastics um tougher, then we, you know, can get more use out of them, and it'll ultimately address some of the problems we have with overall durability and use of plastics.
Um I think that's that's an example of some of the promise of AI and materials discovery.
Cool.
So can you can you uh dig into a little bit?
Um what was the surprising chemical discovery there?
So it's sort of hard for me to think about how to how to explain it without getting too deep into the chemistry.
But basically, these are molecules that have to break apart, and when they break apart, they make the overall structure that they're in tougher.
So a little part of the material breaks and that helps to dissipate the force.
Normally, the way you would think about making it easier to break apart these small molecular components might be to create a hinge so they can kind of peel open instead of sliding apart.
But what we discovered was that there was a fully quantum mechanical phenomenon.
There was really no way for us to predict this, you know, based on anything else, where the electrons just move around in a different way so that at this moment where the molecule is going to break apart, it's a lot more stabilized.
Um, these types of concepts, they're sort of similar to what's kind of known about how catalysts and enzymes work, but it had never before been shown in in uh these polymer materials.
So this is sort of like the the the fuse in the bay bridge that sort of like allows the the bridge to keep his structural integrity during an earthquake by having a controlled break, is that kind of yeah, yeah.
So we weren't the first ones to discover that phenomenon on its own.
The general phenomenon that putting little places that could break to make the network stronger.
That was published in Science Magazine a couple years ago, but the specific way we came up with to design the material to do this, that was that was our new contribution.
How did you you mentioned that you know you were started off in accelerating kind of existing methods using uh comp you know sort of the enhanced computation?
What caused you to take that leap to more machine learning-based methods?
So, you know, I I was drawn to data-driven discovery pretty early on, sort of before I even knew the phrase machine learning.
And I guess I was just really excited by what you could learn from patterns and data.
Back then we were trying to call it cheminformatics and just sort of trying to think about um, you know, in what ways could you unearth trends in data?
Because I I started my career actually working kind of one molecule at a time or one material at a time.
And I was just impatient.
I wanted to be able to sort of understand not just one molecule at a time and write one paper about it, which is something people would have been happy to do back when I was starting my career in the mid-2000s, um, but to actually kind of unearth broader trends in in how you understand how material is gonna behave.
Somewhere around 2015, 2016, I realized it was a bad idea to call things cheminformatics, and it was a good idea to start calling things machine learning.
Um and I had a brilliant student, uh John Paul Janay, who's now I think uh assistant director at uh AstraZeneca in Sweden, um, running their um inverse design program.
He and I originally talked about all sorts of ways of thinking about materials design, and he very quickly adapted that into training neural networks.
Um that's sort of one, you know, I thought we were in the first sort of hype cycle, the first wave, but I think compared to what's going on right now, it was a tiny baby wave.
I read in your paper that that was actually a class project or something.
Yeah, yeah, that's right.
You know, he just said I have to do something for my homework, and that's how we that's how we got into it.
I've also read in your paper that you've done a lot of work like slightly more recently on active learning and using.
Can you talk a little bit about that?
Yeah, yeah.
So even that Polymer example I was giving that would have been active learning in principle, but we sort of stopped after one generation because um because we had exhausted the space.
But um I think one of the areas where machine learning kind of just with what's out there right now has the most promising chemical sciences is in solving multidimensional challenges.
So right now we're working on a project in um metal organic frameworks where we're trying to solve trade-offs uh relevant for uh direct capture of uh CO2 from the air.
And so, in order to find a material that's good for that, we would worry about its cost, its stability in um, say aqueous humid environments, um, its ability to take in CO2 over other molecules, its mechanical stability, is it gonna hold up um under fours?
Um, is it thermally stable?
Well, can you heat it up and will it be okay?
Um, I'm just naming a few, but in total, um, right now in an active learning campaign, we're working on seven different objectives.
And usually just even for a not so accurate machine learning model, you get, you know, at least a hundred to a thousand fold speed up for every dimension you're optimizing over.
So the real promise is gonna be in searching for that needle in a haystack um with say seven objectives and and doing something where you're not waiting for the models to be accurate before you start doing that optimization.
That's really the promise of active learning.
Yeah, that has an interesting parallel in my mind to the the pharma world where you have a lot of a lot of computational work in in the discovery process, but that actually getting it the drug out to in people's hands is often the bottleneck for a drug.
And also, you know, what happens to the drug when it sits on the shelf for three months, um, that kind of thing.
Yeah.
Are these um medical organic frameworks?
What are the kind of things that um they're they're useful for?
They're used most in gas storage, sensing, um, and separations.
They're used in com combination with polymer composites.
They have really uh strong promise for CO2 capture, especially, but people have looked at them for um catalysis.
The limitation on catalysis has been, you know, how stable are they?
So one of the things we've spent a lot of time on is trying to be able to predict their stability.
Um, but they're used for all sorts of things, even uh drug delivery, um uh, you know, uh what they have the opportunity to do is really place precise chemical groups in specific orientations that can ultimately allow for what's known as host guest interaction.
So basically kind of create a glove to have a targeted interaction with a with a guest molecule in the metal organic framework.
I see.
And just uh for the for the non-chemist metal organic framework, Legos for for chemistry is that Yeah, yeah.
Metal organic frameworks, I think are going to be a little bit more of a household name among some engineers because it they um the discoverers of those uh materials just won the Nobel Prize in chemistry this year.
Um so as as much as that can make something a chemistry a household name, but they're basically um like tinker toys or uh Legos, um, and they have different building blocks that can be combined in basically infinite ways to create very precise chemistry.
I see.
Maybe for context, could we can we step back in like what are the techniques you were using before um you started, or maybe in parallel with machine learning?
And how does machine learning help you um advance those?
Like what are the roles of the two?
So I I started my career studying what's known as transition malcatalysis.
If you look at the periodic table, the middle of it contains a bunch of metals.
A good example would be iron.
And all of those things sitting in the middle of the periodic table, they have uh what's what's referred to as an open shell.
So the electrons in those those materials are not paired and they're not um they they're as a result more reactive.
Uh normally, like the the way that you understand how they're going to behave.
So that for instance, they give rise to, you know, different combinations of these metals give rise to the catalysts that are used in a large number of transformations, including the things that say feed and sustain most of the world's population, such as the Haberbosch process for ammonia synthesis.
Um, and the way going back 20, 30, 50 years that people understood these materials and could enable their rational design is through quantum mechanical modeling.
Um, quantum mechanical modeling by using approximations to the Schrödinger equation, it can be very accurate, but it's very computationally costly.
And so a single quantum mechanical prediction, depending on the level of fidelity used, could take hours to days to weeks.
And that's what I would have normally been doing before I got started in AI.
Some of what we do these days is accelerating those quantum mechanical predictions as well as looking at, you know, an area that I'm particularly excited about is that not all quantum mechanical approximations are equal.
And you can actually use ML models to kind of predict what the best approximation to use is depending on the material studied.
Is that like closer is better, or is it some is it it's not really distance related?
In terms of which method is the right method to use?
Yeah.
So we it actually turns out to be quite complex.
You can't just determine it from heuristics.
So we actually, in one area, use the quantum mechanical wave function as inputs to uh neural networks to actually predict what is the right method to use and learn that mapping.
I see.
That's probably gonna be in the soundbooker challenge.
But cool like 22 Adam Wigan challenge go.
I have a spicy question I want to ask.
So there's a school of thought that why should I bother to learn chemistry or physics or whatever when ChatGPT no has PhD level understanding of that anyway, and shouldn't I just focus on being really good at using AI for stuff?
So I want to hear your thoughts.
My personal experience is that um, and this will date itself immediately, is that is that uh Chat GPT is super good at Wikipedia level chemistry knowledge.
But one of my favorite things to actually throw at GPT as as an anecdote is I'm really interested in molecular design.
Like how do you find a new ligand um that can go into a transition male complex?
And what that means is it's some combination of atoms and it's gonna bind to the metal and it's gonna change its properties.
And so the thing I constantly do every time an LLM is updated is I just ask it, please design me a ligand that has uh 22 atoms.
So the first time I've I've done that, there are many ligands out there that have 22 atoms.
And I say, I wanted to bind to the metal with two nitrogen atoms.
I can never get an answer that has 22 atoms.
So then you can try a range and see how many times you can get a range.
And so that's that's maybe a trivial thing, but that's something that an expert chemist, you know, could do in a do in a second.
Um, so there are really good introductions to chemistry that I think you can get through conversations with an LLM.
Um you can get a lot of insight into an area you're unfamiliar with.
And for sure, things have improved a lot.
Like when I first tried typing in, you know, which exchange correlation functional should I use for this type of chemistry?
The answers were completely wrong.
They looked right, but they were completely wrong.
I think things have gotten better because that knowledge is out there on the internet, it's in the training data.
But I think there's a lot of things that um probably backing up a moment, um you should learn chemistry well enough to know when when these models are right or wrong um and if you don't know any chemistry at all it's hard to know if you're if you're assessing correctly but I think that there are a lot of things that you don't have time to do a deep dive into that you can now get from say an LLM that can augment knowledge but I think you have to start from somewhere and then use it as a tool rather than starting from zero and relying blindly on what an LLM will say but one of my favorite things if if someone can get in one shot an LLM to generate me a 22 atom ligand I I would I would love to see it.
What do you think the biggest gaps that machine learning has from your experience that like if you were an aspiring um ML engineer with looking to take on a new problem from the machine learning side that what do you think someone could work on which would really help the chemistry side?
There are a lot of challenges out there where the data sets aren't large enough or diverse enough and so I think they've attracted less interest.
So the ones closest to my heart are uh reactivity predictions, so predicting which reactions will occur and and why, especially in complex uh phenomena uh like in um you know multiple elements and and sort of predicting those transformations uh another thing that I think um there's not enough data on is just more diverse chemical bonding and more diverse chemistry.
For me, that's transition metals, but there's also questions of warm, dense materials, sort of exotic phenomenon.
We have really good data sets out there for really boring chemistry.
So we have, you know, probably, even if you're not a chemist, you're familiar with organic molecule data sets and organic molecules binding to proteins.
Those are the common data sets out there.
There's lots of challenges out there where the physics is much more complex, and the things like how does matter behave when you shine light on it and you excite it into excited states, all sorts of things like that receive relatively little attention because you know there may not be a benchmark or a leaderboard yet for that.
And so maybe it's on us chemists to generate more data sets.
So those leaderboards are out there, but there's definitely, you know, a lot of interest in chemistry for which there has been less attention.
So in the protein world, there's CASP, right?
And people have been working on that for a while, and this led to AlphaFold, like kind of without CASP, AlphaFold probably wouldn't exist.
Is there like an equivalent to CASP in the material science world?
So there are all sorts of repositories of fairly low fidelity DFT data on crystalline materials.
So materials project, um, opid catalyst project.
These do provide good leaderboards, but some of the limitations of that are the data comes from not very high fidelity density functional theory.
So I'd say that's a second um challenge is that we're all the smartest ML engineers right now are learning on data that is not going to be reflective of experiment.
There aren't big experimental data sets, for example.
Um, one of the advantages of things like CASP is that it comes from an experimental ground truth, um, whereas uh that aspect just isn't available in materials as much.
We were talking about CASP and you know the role of CASP and AlphaFold.
Do you think that there is um like a problem, a way of phrasing this, that we could start collecting data at scale, that we could um you know really have a community challenge which breaks open some open problem in your mind.
And maybe like uh maybe actually even stepping back beyond that, what what would you want to have if there was like an alpha fold um for materials?
What would you want it to do?
One kind of murky area, so maybe I'm not gonna directly answer answer this question.
One murky area for us is electronic structure calculations are expensive, and uh they should in principle give you the right answer.
They should from first principles give you the right answer of how a material is going to behave.
And a lot of people are scaling these up right now with uh machine-learned interatomic potentials on training data.
And every time someone comes out with kind of a new data set trained on a and they call it a foundation potential, foundation model, it looks really good.
And then you get it into your lab and you say, okay, I want to use it for this problem, I'm really excited about, and it starts doing kind of wacky things like molecules fall apart.
I won't name names, but um there was one that made a huge splash this summer and people started declaring, oh, this method is dead, this method is dead.
We're all gonna just use these neural network models now.
Um it's only in my hands, uh, the one I'm still not naming is only about five times faster than my fastest DFT calculation on a GPU, and it also doesn't work all the time.
So I would say we need a more transparent way of of trying to figure out um if these models can really replace conventional physics-based modeling.
Um, if they could, if if I could just give up ever doing a DFT calculation again and just rely on machine learned potentials and if they were, you know, two orders of magnitude faster than the traditional approach, that would change that would change how we're doing science.
But there needs to be a little more rigor on what we consider, you know, just fitting data when that data maybe lacks quality, or there needs to be a little bit tougher requirement for for how we say this this model can really replace the physics-based modeling.
Yeah, so I one of our theses is that the interface between bits and atoms is really the bottleneck, right?
Where you have to the actual activity of trying things in the lab is the bottleneck, and you've addressed that to some extent in active learning.
Um, but I think that there's also an extent to which that just pure uh process and automation, good operational practice, those are important things.
So that if you you can push to automation on the one side, but on the other side, that creates brittleness.
So, how do you think about kind of bridging that gap to experimental chemistry and and um using that as sort of as a as a nature's computer um to figure out things for your design process?
Yeah, so so there are a lot of really smart people working in um high throughput synthesis and experimentation and autonomous labs.
I think the thing that uh that's interesting to me in that space, at least, is that there are some types of experiments that, at least as of the last conference I went to on this, are really hard for autonomous uh high throughput experimentation, but are really easy for a human and vice versa.
And then there's, you know, the serendipity that a human might experience in the lab that a couple of people have tried to think about like, well, how do you how do you introduce that noise into high throughput experimentation?
So I think that's a challenge.
Your question also brought to mind another point that I'm by no means an expert on, but most people who actually work on getting materials to the device scale, say something that would um be in your television or something like that, is they will tell you that it's not just the material, it's the process.
And I think we're at we're at ground zero.
We're we're nowhere when it comes to like, well, how do we machine learn not just the structure and the properties, but also the role that processing plays?
Um I don't think we know anything about how to do that.
Maybe for non-experts, like with protein structure, it's really easy to imagine, like, oh, you can see these proteins, like, and we can run some simulations and see them wiggling around and uh the structures look really pretty.
What does the data look like for material science?
Is there's the computations like DFT, I think, gives you something which looks like a crystal structure, you can imagine.
But then there's also like, is there experimental data where you can observe that crystal structure, or is this mostly sort of like kind of probes where you're measuring individual properties, which are kind of collective and with not fine-grained?
So experimental structures are available, and the example I was giving is something we know is stable and we've seen a structure of it before, and it will fall apart with some of these models.
The challenge here is that um what Alpha Fold has done really well is is predict structures of globular proteins primarily with 20 um natural amino acids.
I could actually point to lots of cases where alpha fold fails too for more interesting chemistry.
And so there's lots of different ways to think about chemical bonding and right now no um potentials are really robustly encoding all of that bonding especially with respect to um metal organic bonding.
Yeah maybe a different way of saying it is like with alpha fold I mean alpha cold is solving ground state structures like it's not looking at dynamics um which is I think consistent with like some of your statements about needing quantum mechanics for catalytic enzymes um so um but even you're saying even at like just kind of ground state properties you're saying that just there are too many uh parameters and there's not like a clear set of interactions which is limited to a small number of building blocks.
The bonding is is highly variable across all of material space.
Now there's simple regions of material space.
You can pick aluminum.
Aluminum is very boring, and you can write down people in the 60s could write down on on paper, you know, how you need to model aluminum.
That's something that is pretty easy to fit a neural network potential to um but then if you want to get over to iron oxide, and then if you want to get over to um high entropy alloys, there are definitely cases where people are using these methods um but I'd say a big challenge is that there's no real way to know if when you go to bigger lane scales and time scales, there's no real way to know if you're right or wrong.
The experimental data is not there.
Experiment, even interpreting, say looking at an image of an experimental uh surface, which you would want to do.
Um it requires some degree of an interpretation of that image.
Um, so it's just it's just hard to know from experiment or from other computations if these types of models are are correct.
And they're certainly not correct across all of chemical space.
Um, and I'd say they could fail more catastrophically than Alpha Fold obviously fails, though they're definitely failures of Alpha Fold too.
Switching gears a little bit.
I I read in your paper also that you had done some work with integrating textual information from papers and into your um, so it's kind of the AI that we all know and love right now.
Um, can you talk about what kind of lift that that gives uh the models and how did how did you actually do that integration?
Yeah, so we started um, I guess about five years ago.
So when we first started doing it, we were just doing sort of standard natural language processing and graph digitization.
Um these days we use LLMs, um, but just to try to extract from the literature data sets of properties, uh, wherever, wherever people are widely reporting properties, and what we noticed is that there's a lot you can learn from these models.
So you can you can even on the scale of a few thousand data points, you can then do things like predict the temperature at which a moth will break apart based on experimental reports.
But one of the funniest things I think we noticed is that you can get the temperature at which a material will break down two ways.
One, you can get it from the graph, and two, you can get it from what the authors say about how they interpret the graph.
And those two things do not line up.
So people, you know, one of the challenges I think with literature extraction from papers is one would be the obvious mistakes people make, you know, no one's perfect.
But the other would be just, you know, people interpret their results in different ways.
And so if we're building models based on those interpretations, that's a challenge.
In terms of LLMs, um, they've come a long way in terms of literature extraction, but they're still definitely sensitive to false positives.
And I think the amount of time we spend uh checking on LLMs to make sure that the data we're ingesting is is um accurate, definitely is an overhead on those types of workflows.
I see.
And and what about um the way that it might bias the discovery process, right?
Because you have this known literature, your job as a chemist kind of sort of is to find new stuff.
But if so, if you're emphasize if your computational method is pulling in literature, then maybe it's biasing you towards the previously reported results instead of something, yeah.
You know, one of the ways we try to address that is we try to train a model on that literature, but then apply it to new structures that have never been seen before and try to really look at how far we can extend the model.
Um, but we are trying to answer this in general.
Uh there are repositories out there of experimental data where you can have a sense of when it was published, what the structure is, um, what it was used for, and we're really trying to build generative models on top of that now to try to be able to say, well, well, if I know about the first 30 years of a field, can a model trained on that predict the next 20?
I think that's an that's an open question.
And what model is best?
And you know, and maybe it won't get all of them, but maybe some of those discoveries that we think are new in the most recent 20 years, maybe maybe some of them are trivial for a model to generalize to, whereas others are are not.
I think in an ideal case where where we have the available literature data and we don't know, we could use uncertainty quantification to then identify, okay, these would be the most interesting materials to get into our data set.
I see.
And those data sets just for people who are interested in getting involved, what what are some of the what's the best ones to get to get started with?
I don't know about the best.
We we've we've curated a few thousand data points of metal organic framework, thermal stability, as well as metal organic framework, activation stability, water stability.
Uh other groups have curated other measures of stability.
They're all out there, they're on our website, um, that kind of thing.
Awesome.
Do you imagine there being uh you useful to like create an initiative or like a multi-institutional like funding source or something which really is trying to get data in a high throughput automated way?
Um what would sort of like your dream be if you could organize something which like which really will drive the field forward in your mind?
I think the National Science Foundation has one initiative.
I've also heard about things with with um uh foundations before, sort of being interested in in putting together cloud labs.
So things that users can on demand make use of high throughput automation.
I definitely think I think having user facilities where a computational researcher like me could design an experiment and have it executed would be awesome.
Having all that data collected in sort of a public way would be great.
You know, the way that research right now gets published into papers, it's very hard to then extract back out.
We spend a lot of energy trying to get it back out.
And so some of this is a need also for, you know, maybe systematization of how results get reported so that uh they can be machine learning ready from from day one when they're published.
Some research sub-fields are trying to do that, but it's it's not really developed across material science.
But for sure, you know, I think there will be more sort of shared facilities where people can uh make use of data from high throughput experimentation and that would be really really awesome.
I don't know if it'll come from companies donating equipment from National Science Foundation or from uh you know private foundations.
Yeah there is a large like philanthropic push in the biotech space um it seems like people haven't quite picked up on this as such an important uh field like especially with things like materials for climate change you can imagine a particular a very important problem that we could use a lot of push on yeah that that kind of brings up the question there's been a ton of very recent materials investment for private companies uh startups um where does that leave in your mind the role of the academic and chemistry I ask myself that all the time it or more recently in the past year so in particular there's a lot of uh compute that companies have access to that academics don't so I ask myself you know what can we do that's uh more creative that doesn't require just brute force compute um and I think there is there is like an a lot of stuff that we can still do um but we have to ask those questions um for sure Microsoft Meta those those ones are are kind of like the companies that have basically infinite resources and as an academic I don't have infinite resources you know but we have an interest in problems that you know haven't crossed the radar of those companies yet and I think as long as we you know whenever someone poses a problem to me now versus a few years ago I try to make sure that we're not just in the process of trying to do something that throwing a lot of compute at it would would would solve it.
Yeah I think we're kind of running out of time but would like to give you an opportunity call to action what what would you uh like our listeners to know about do um what what should they do to get involved or something that you're really passionate about.
I think I will stick to something kind of niche.
Great.
So I I think there is still a place for chemistry.
I will say that um but uh my group develops a code for uh transition malcomplex structure generation metal organic framework screening it's called mole simplifier when we're working on MOFs we call it moff simplify there's website versions of it that you can look up and not install anything but it's also on conda and github.
And if you do have an interest in transition malcomplexes, you know, just try it out.
It includes machine learning predictions, but it also make novel structures.
And I'm just really interested to hear ever if people are using it.
I know a lot of companies are using it, but we sort of find out sort of after the fact so um if you're interested more in this material space I'm I'm definitely interested and open to feedback.
Grateful, awesome.
Get in there both.
Thank you very much.
Thank you, Doctor