![Taylor Sparks](/img/default-banner.jpg)
- Видео 681
- Просмотров 4 339 519
Taylor Sparks
США
Добавлен 23 апр 2008
I'm an Associate Professor of Materials Science & Engineering at the University of Utah. I'm passionate about engineering education. I love making tutorials and example problems. Just comment an any video with a request for a new problem or explanation and I'll make it happen.
33. Large Language Models in Materials Science
Welcome back to our Materials Informatics playlist! In today's episode, we explore the transformative impact of Large Language Models (LLMs) on materials science. David Sparks from the University of Utah's Materials Science and Engineering department takes us through the exciting possibilities that LLMs bring to the field.
Here's a brief overview of what we'll cover:
Introduction to Large Language Models (LLMs): Understanding their significance and historical context in AI and natural language processing.
Pre-Neural Network NLP Techniques: Overview of methods like Bag of Words, Term Frequency-Inverse Document Frequency (TF-IDF), and N-Gram Models.
Word Embeddings and Neural Networks: Introduc...
Here's a brief overview of what we'll cover:
Introduction to Large Language Models (LLMs): Understanding their significance and historical context in AI and natural language processing.
Pre-Neural Network NLP Techniques: Overview of methods like Bag of Words, Term Frequency-Inverse Document Frequency (TF-IDF), and N-Gram Models.
Word Embeddings and Neural Networks: Introduc...
Просмотров: 294
Видео
32. Bayesian Optimization
Просмотров 1692 часа назад
Welcome back to our Materials Informatics series! In today's episode, we delve into Bayesian Optimization, a critical tool for incrementally improving processes and designs in materials research. Bayesian Optimization leverages Bayes' theorem to make informed decisions with minimal data, making it particularly valuable in material science, where property trade-offs are common. Here's a brief ov...
31. Gaussian Processes
Просмотров 3444 часа назад
Welcome back to our Materials Informatics playlist! In this video, we dive into the fascinating world of Gaussian Processes, building on our previous discussion about Naive Bayes theorem. Gaussian Processes are a powerful tool in probabilistic machine learning, especially for modeling and predicting complex data. They can be tricky to understand, but don't worry; we've got you covered! Here's w...
30a Coding a Naive Bayes classifier by hand
Просмотров 2817 часов назад
Welcome back to our Materials Informatics series! In this video, we continue exploring Bayesian and probabilistic machine learning by diving deeper into Naive Bayes classification. In this video, we cover: A brief recap of Naive Bayes, highlighting its simplicity and the assumption of feature independence. A practical demonstration of Naive Bayes using materials project data to classify whether...
30. Bayes Theorem and Naive bayes classifier
Просмотров 2529 часов назад
Welcome back to our Materials Informatics series! In this video, Taylor Sparks from the University of Utah delves into the fundamentals of Bayes Theorem and Naive Bayes classifiers, essential concepts in probabilistic machine learning. In this video, we cover: An introduction to Bayesian and frequentist approaches in machine learning, highlighting key differences. The basics of conditional prob...
Episode 91: High Entropy Alloys
Просмотров 1,1 тыс.21 час назад
A new class of material doesn't show up often. In this episode, we dive into the revolutionary discovery of high entropy alloys (HEAs) that revitalized the field of metallurgy. We dive into their simultaneous discovery at both Oxford and Tsinghua National University. Learn how they break all the typical rules we know. We explore how their composition gives them exceptional strength, hardness, t...
29. Diffusion Models
Просмотров 25721 час назад
Welcome back to our Materials Informatics series! In this video, we're diving into the fascinating world of diffusion models, the cutting-edge technology in generative models that is outperforming GANs and VAEs. In this video, we cover: Introduction to diffusion models and their superiority in generative tasks. Real-world applications of diffusion models, including popular tools like Midjourney...
28. Image segmentation
Просмотров 210День назад
Welcome back to our Materials Informatics playlist! In our previous video, we explored generative models like GANs. Today, we’re taking a brief interlude to discuss the fascinating world of image segmentation because tools from image segmentation are crucial to our next generative model! In this video, we cover: The basics of image segmentation and its importance in materials science. Different...
Episode 90: The Big Dig Incident
Просмотров 400Месяц назад
Choosing the wrong material can have dire consequences. In this episode of our failure series, we discuss how the incorrect choice of epoxy led to a catastrophic failure and a tragic death. Discover the series of poor decisions that turned a new highway plan into one of the costliest public works projects ever. Join us as we uncover the lessons learned from this devastating event and the change...
Episode 89: Special Applications of Microscopy Technologies
Просмотров 580Месяц назад
Electron microscopy is almost a century old, but it continues to play a role in exciting new developments that extend its use well beyond its original purpose. We sit down with Professor Sergei Kalinin from the University of Tennessee-Knoxville to discuss these exciting new applications of this older technology. Learn how electron microscopy, originally developed for imaging, is now used for at...
Episode 88: Accelerating Materials Discovery with Microsoft
Просмотров 9702 месяца назад
The discovery of new materials is an immense challenge, with a vast design space and numerous success criteria. Microsoft has recently demonstrated an advanced approach to machine learning-assisted material discovery, particularly in the realm of lithium-ion battery electrolytes. They began by exploring all possible structure types, decorating these structures with various atoms, leading to a p...
Episode 87: Stories of a Materials Salesman
Просмотров 3203 месяца назад
Designing a great material is only half the battle, now you need to sell it. In this episode we sat down with Dan Wilson from Sintx and took a dive into the surprisingly nuanced world of materials sales and the challenges and pitfalls of marketing materials solutions. Learn more about Sintx and the great materials they are developing by visiting their website at: www.sintx.com/ The Materialism ...
Lecture in Materials 9: Anil Virkar "Thermodynamic and kinetic considerations of fuel cell catalyst"
Просмотров 2283 месяца назад
State-of-the-art PEMFC use platinum-based catalysts as cathodes. The high cost of Pt and tendency for catalyst degradation at the cathode is the principal reason for devising ways of minimizing the amount of Pt used and ways to enhance catalyst activity and durability. Approaches used to decrease the Pt loading include alloy formation and forming core-shell catalyst with non-noble metal as the ...
Lecture in Materials 8: Angie Richardson & Lindsay Fuoco "Piezoelectric ceramic processing"
Просмотров 2703 месяца назад
Piezoelectric ceramic industry is a 1.4 billion dollar market with applications as varied as sonar and navigation systems for the US Navy to medical ultrasonics to watch alarms. The ceramic formulation and processing not only requires structural integrity, but also a high efficiency electrical characteristic that can convert electrical signal to mechanical energy, or mechanical to electrical si...
Lecture in Materials 7: Dan Belnap "HPHT diamond sintering and nanodiamond manufacturing"
Просмотров 2293 месяца назад
Lecture in Materials 7: Dan Belnap "HPHT diamond sintering and nanodiamond manufacturing"
Use GitHub Copilot Chat to interactively create documentation, add type hints, and make unit tests
Просмотров 6183 месяца назад
Use GitHub Copilot Chat to interactively create documentation, add type hints, and make unit tests
Interactive sample transfer workflow with slack notifications
Просмотров 2663 месяца назад
Interactive sample transfer workflow with slack notifications
Human-in-the-loop Bayesian Optimization
Просмотров 4073 месяца назад
Human-in-the-loop Bayesian Optimization
Episode 86: PHAs and Biodegradable Plastic
Просмотров 5123 месяца назад
Episode 86: PHAs and Biodegradable Plastic
Materialism Podcast Ep 85: Electron Backscatter Diffraction
Просмотров 5444 месяца назад
Materialism Podcast Ep 85: Electron Backscatter Diffraction
Max Balandat (Meta Adaptive Experimentation) "Bayesian Optimization for Sustainable Concrete"
Просмотров 4654 месяца назад
Max Balandat (Meta Adaptive Experimentation) "Bayesian Optimization for Sustainable Concrete"
Martin Fitzner "Industrial view on Bayesian optimization A perfect match for the low/no-data regime"
Просмотров 2414 месяца назад
Martin Fitzner "Industrial view on Bayesian optimization A perfect match for the low/no-data regime"
Discovering materials twice as fast at a fraction of the cost through Bayesian optimization
Просмотров 8234 месяца назад
Discovering materials twice as fast at a fraction of the cost through Bayesian optimization
Lecture in Materials 6: Ian Harvey "Rocket nozzle materials"
Просмотров 3134 месяца назад
Lecture in Materials 6: Ian Harvey "Rocket nozzle materials"
Lecture in Materials 5. S. Elango Elangovan "Mission to Mars: One small step before a giant leap"
Просмотров 1794 месяца назад
Lecture in Materials 5. S. Elango Elangovan "Mission to Mars: One small step before a giant leap"
Lecture in Materials 4: Willard A Cutler "Materials Science-Based Innovations to Pollution Control"
Просмотров 2324 месяца назад
Lecture in Materials 4: Willard A Cutler "Materials Science-Based Innovations to Pollution Control"
26a. Application of GANs in Materials Science
Просмотров 7014 месяца назад
26a. Application of GANs in Materials Science
Materialism Podcast Ep 84: The ICME Method with QuesTek
Просмотров 3844 месяца назад
Materialism Podcast Ep 84: The ICME Method with QuesTek
Materialism Podcast Ep 83: Computed Tomography at Zeiss
Просмотров 3135 месяцев назад
Materialism Podcast Ep 83: Computed Tomography at Zeiss
I paused in the middle of the video to subscribe 👌
@@basharb5215 hero :)
Can you please make a tutorial for new updated material project API and building various machine learning models?
I have a notebook already prepared for the new API. Here's the link. Does it have what you are looking for? If you have additional questions, let me know and I can update it with additional content. github.com/sp8rks/MaterialsInformatics/blob/main/worked_examples/MP_API_example/new_MPRester_tutorial.ipynb
And this video has a bunch of instructions on cleaning, featurizing, splitting data and building classic models. Feel free to use the bookmarks at the bottom to jump right to the timestamp of interest ruclips.net/video/5fMr4mYuCXI/видео.htmlsi=PDp3NgWM3kEYTZqX
Such a great video. Thanks a lot :) i am working on this so this helps a lot
@@Pingu_astrocat21 glad to help!
Great job explains this!
Stanley has a some brand new notebooks to accompany it coming soon ;) I'm reviewing them tomorrow with him
What I need right now. Please Sensei, is formation energy per atom here related to enthalpy of formation? What is the significance actually?
@@kwamivikolor5276 Formation energy per atom focuses on the energy change per atom in a compound, commonly used in computational materials science. Enthalpy of formation is a broader thermodynamic measure of the heat change when forming one mole of a compound from its elements. Both indicate stability, but formation energy per atom is more specific to atomic-level stability, while enthalpy of formation applies to bulk thermodynamic processes.
@@TaylorSparks Thank you Sensei. Therefore in terms of sign(+/-) of the formation energy per atom what kind of intuition can we develop? I mean when the formation energy per atom is negative or positive or null?
Yessr
Thank you for this video professor. The problem I feel about such programming is that it is difficult to make the program understand science rather than just learning patterns from a large data.
@@8848nepalyt yup. Some machine learning models are black boxes with limited interpretability. Others are more interpretable. Random forests, for example, provide feature weights that can help intuite mechanistic understanding
Very useful video. Please teach how to refine materials with partial occupancy and how to apply constraints on it
this is the greatest explanation of this I've ever gotten
how do we read SN curves after getting hot spot stress
Hi professor, would you please talk on XAS, XPS!
@@Hamza-vk6sc this is an excellent idea.
can you do a playlist for all the phase diagram videos
Greetings and courtesy and respect Please advise where is the metallurgy database in materials? Data such as strain rate, analysis, air humidity, etc
@@AliakbarKhosravi-th3mv unfortunately there isn't a great source for this. Best bet is materials data facility by globus
@@TaylorSparks you mean globus?
@@AliakbarKhosravi-th3mv yes
@@TaylorSparks I do not know what is it globus
@@AliakbarKhosravi-th3mv materialsdatafacility.org/ This is the closest thing we have to a generalized materials data repository. If it's not here I don't know where else you would find it.
Could you cover dual phase metals next
Any alloy in particular? Most alloy systems are multiple phase mixtures.
@@TaylorSparks could you do dielectric and diffractive films like the kind they use on euv mirrors? I've seen something in literature with neutron diffraction
would be cool to have such a breakdown in terms of plastic recyclates and their possible morphologies
I've been struggling to learn Python. I attended your seminar at the Texas A&M material science summer workshop in 2023 and found your videos! They've been a real life-saver! Thank you so much.
@@lakshmiviswanadha3454 this makes me so so happy to hear. I need to reteach this course in the era of LLM-generated code. Total game changer.
At last! Been waiting for a HEA ep. for so long!
Music could be a little lower but a solid episode overall❤
Great podcast! it pretty much covers all the critical issues. I look forward to more on this topic.
Additive manufacturing is pushing the compositional space as well since rapid cooling significantly influence microstructure of multiphase HEAs. Much to talk about there!
This is an excellent overview of image segmentation and different techniques. We found that each technique has its strengths and weaknesses, and by integrating these different approaches we are able to automate some pretty challenging materials microstructure characterization problems. Regarding SAM, for most images coming off the microscope, we are able to run this model on on-premises workstations, on off the shelf hardware, no need for cloud computing. Thank you for putting together this video, we find that there's still a ton of uncertainty around how these tools can be used to solve real world problems.
@@MiparUs hey guys, I really enjoyed your talk at the TMS specialty conference this summer. Let me know if you ever want to do a dedicated sponsored episode of the materials and podcast. We have a pretty big reach.
I recommend the intro music be turned down, it overshadowed everything you said. Good talk otherwise.
@@Kestrel216 good feedback. I'll have Jared adjust it
I have made a 3D reconstruction of a microstructure and I would like to plot a "Comparison of two-point correlation function curves between reconstructed and real cores" to illustrate the effectiveness of my reconstruction. The horizontal axis of the image is "two-point distance/pixel" and the vertical axis is "two-point correlation function value". Do you have a suitable method?
@@arhejie have you tried pymks??
Sorry, I haven't tried it. I'm not familiar with this area. I don't know where the data I want is stored. For example, "two-point distance/pixel", I don't know where to find these two parameters. I also don't understand what "two-point correlation function value" is. How can such a function that describes a distribution have a specific value?
@@arhejie you will want to look up the original papers by surya kalidindi and David Fulwood and others. The two points statistics do not boil down to a single number, you are correct but they are a distribution. The results will be a matrix or tensor in your case with three-dimensional data
@@TaylorSparks Thanks, your reply is very helpful!
Just wanna say you and this other channel saved my life for a Materials and Manufacturing exam 😅
Heck yeah!! Tell your class mates
Hi. found your channel recently. Why do you have such a high sub count but a low average of views per video? I mean the content is so good and so incredibly unique
I dunno man! Sread the word ;) I'm trying to keep churning out useful materials science content.
What other courses would be good to study for material science? (Currently a working PE in the HVAC field)
@@jeremiahdewitt2072 kind of depends on your interests. I think polymers and ceramics are both fundamental courses for material Science though. Materials processing is also great and semiconductors.
I love these videos. As a mechanical engineer, I wish I could go back and have studied MatSci after watching so many of your lectures and listening to your podcasts. In the future could you talk about Kirkendall voiding, if you haven't already? Thanks!
How does a compressive curve look and how do you analyze it ?
@@giggachad8153 there's no guarantee, but it typically looks a lot like the tension curve but smoother and higher strength values since most materials are going to be stronger under compression than tension. This is especially the case with ceramics.
@@TaylorSparks Do you hppen to have any referenc material showing the compression part of these graphs? most of what i find are examples in tension and I am trying to mathematically prove the Modulus of elasticity for some experimental data i collected on a polymer. my stress (compression) vs strain does not have the shape i expected and it doesnt follow any of the examples ive found. It also does not express linear behavior which is making it difficult to find the modulus. Thanks !
Hello, thank you for this material. Where can I find the periodic table that you're using that shows the naturally occurring charges of elements?
@@mhermovsisyan3342 this one is my favorite because it has the charges and their sizes. mrlweb.mrl.ucsb.edu/~seshadri/Periodic/index.html
@@TaylorSparks Thank you very much!
great video thanks!! how much is the factor used to chnge BH to MPa, for Aluminum
interesting!...
Nice topic, very interesting and well explained
What does it mean if a point is directly on the liquidous or solidus line only?
@@MachanamistMechamations-ux5mc directly on the line simply means that it is in equilibrium with the start of that phase. Equilibrium means the same amount of liquid is beginning to solidify as the solid is beginning to liquefy. It's an equilibrium. So if it's on the liquidus line then it is still 100% liquid but it is beginning to solidify and if it's on the solidus line then it's 100% solid but it's beginning to liquefy
@@TaylorSparks Thanks.
Would love to see a follow-up with Eric at CalNano!
@@user-yi3nj8fx1n we've got a few things in the works!
Very nice presentation, I’m certainly going to link to it when I get around to revise my report writing guide for lab classes! It’s probably too late to ask, but just in case you see this, why work with EndNote Click instead of the native Zotero Connector? Better metadata extraction?
@@claudineallen7174 Yes, at the time of recording I would agree with that statement, but things change so rapidly that that may no longer be the case.
congrats! I owe you a thousand shout out for the knowledge you provide <3
@@Rrrrrrrrr107 🙏🏻
The video is great but not clear
This was really good and informative! Thank you so much!
My pleasure!
How do you estimate the percentages because im lost
Thank you .. the video is great help well done thanks again