From b17ab2d5dce9f07f677700a71a43535237f8e545 Mon Sep 17 00:00:00 2001 From: Simon Billinge Date: Sun, 19 Jan 2025 13:08:07 -0500 Subject: [PATCH] start to rationalize and clean projects --- db/projects.yml | 365 ++++++++++++++++++++++++------------------------ 1 file changed, 185 insertions(+), 180 deletions(-) diff --git a/db/projects.yml b/db/projects.yml index ad85dc7e..088c085d 100644 --- a/db/projects.yml +++ b/db/projects.yml @@ -1,35 +1,37 @@ algebra_of_synthesis: active: true - description: 'Modern materials genomics aims to discover and bring to applications - novel materials with improved properties on an accelerated time-scale using "genomic" - methods based on data mining, materials informatics, and AI/ML. It has been very - successful at predicting previously undiscovered atomic arrangements ("materials") - that are stable or nearly stable using computational methods. But the ability - to subsequently make those materials is a greater challenge. This project aims - to tackle not the materials prediction problem, but the materials synthesis problem: - given a material, what is the recipe, or protocol, for making it. We are interested - in taking a data analytic view of this problem with a view to apply generative - AI to the problem, but initially aim to find a simplifying framework for the problem + description: + 'Modern materials genomics aims to discover novel materials and bring + them into to applications on an accelerated time-scale using "genomic" methods + based on data mining, materials informatics, and AI/ML. It has been very successful + at predicting previously undiscovered atomic arrangements ("materials") that are + stable or nearly stable using computational methods. But the ability to subsequently + make those materials is a greater challenge. This project aims to tackle not + the materials prediction problem, but the materials synthesis problem: given a + material, what is the recipe, or protocol, for making it. We are interested in + taking a data analytic view of this problem with a view to apply generative AI + to the problem, but initially aim to find a simplifying framework for the problem by understanding the basic non-commutative algebra and topology of synthesis operations.' group: bg highlights: [] name: The algebra and topology of synthesis - summary: 'This project aims to tackle the materials synthesis problem: given a material, + summary: + "This project aims to tackle the materials synthesis problem: given a material, what is the recipe, or protocol, for making it. The approach is to develop generative AI/ML approaches but also in parallel to understand underlying rules that govern - the interactions between different synthesis operations.' + the interactions between different synthesis operations." team: - name: sbillinge - position: mentor + position: PI begin_year: 2022 type: research - website: https://thebillingegroup.com complex_liquids: active: true - description: Complex liquids, and liquid-liquid interactions, are at the heart of + description: + Complex liquids, and liquid-liquid interactions, are at the heart of separation technologies, thermo-responsive behavior, solubility and life itself. Extracting structural information from liquids is a challenge due to the extensive averaging - that takes place. This is a new project to apply latest generation local structural + that takes place. This project aims to apply latest generation local structural methods and modeling, including machine learning, to gain greater insights into the complex liquid, property, relationship grants: [] @@ -37,23 +39,25 @@ complex_liquids: highlights: - year: 2024 month: 2 - description: Submitted paper to Matter with robust experimental support on molecular + description: + Submitted paper to Matter with robust experimental support on molecular clustering in amine-water mixtures that explains the lower critical solution temperature (LCST) phenomenon that has not had robust experimental support to date. name: Structure-property relationships in complex liquids and solutions - summary: This project aims to understand the structure-property relationships in + summary: + This project aims to understand the structure-property relationships in complex molecular liquids and solutions using high-throughput, high-sensitivity measurements and latest developments in modeling team: - name: sbillinge - position: mentor + position: PI begin_year: 2016 type: research - website: https://thebillingegroup.com dmref15: active: false - description: Advancing applied mathematics and computational approaches to the study + description: + Advancing applied mathematics and computational approaches to the study of the nanostructure of materials grants: - dmref15 @@ -61,88 +65,55 @@ dmref15: highlights: - year: 2021 month: 1 - description: Launched PDF in the cloud (PDFitc) website, the first atomic pair + description: + Launched PDF in the cloud (PDFitc) website, the first atomic pair distribution function data analysis website. Instead of downloading group software and using it on their local computers, users can now upload their measured PDFs and search for candidate structures that are pulled from large structural databases. - year: 2021 month: 1 - description: Published a series of papers that applied artificial intelligence + description: + Published a series of papers that applied artificial intelligence (AI) and machine learning (ML) to the study of nanostructure inverse problems, including using a convolutional neural net (CNN) to predict the space group of a material given an atomic pair distribution function (PDF) measured from x-ray or neutron powder diffraction data. - year: 2020 month: 3 - description: Launched PDF in the cloud (PDFitc) website, the first atomic pair + description: + Launched PDF in the cloud (PDFitc) website, the first atomic pair distribution function data analysis website. Instead of downloading group software and using it on their local computers, users can now upload their measured PDFs and search for candidate structures that are pulled from large structural databases. - year: 2018 month: 6 - description: Discovered a novel approach to model metallic nanoparticle structures + description: + Discovered a novel approach to model metallic nanoparticle structures by building discrete models algoritmically for databases. published in Acta Crystallographica A. - name: Using data analytics to regularize the nanostructure inverse problem, and + name: + Using data analytics to regularize the nanostructure inverse problem, and NSF funded DMREF center team: - name: sbillinge - position: mentor + position: PI begin_year: 2016 type: research - website: https://thebillingegroup.com dmrefcheme16: active: false - description: Apply AI and machine learning to get better gas separation membranes + description: + Finding polymer membranes that are both selective to separating particular + gas species, and also have high permeability, is a frontier challenge in chemical + engineering. Thid project sought to apply AI/ML to search for novel gas membrane + candidates grants: - dmrefcheme16 group: bg - highlights: - - year: 2024 - month: 2 - description: In this period, I won one award, published one book, 14 peer reviewed - articles, gave 7 plenary/keynote, 15 invited talks, and 9 seminars and colloquia - - year: 2024 - month: 2 - description: 'Co-edited a perspective on Machine Learning in Materials Science - published in Acta Crystallographica: Section A' - - year: 2024 - month: 2 - description: 'The book "Atomic Pair Distribution Function Analysis: A Primer - authored by Simon Billinge and Kirsten Jensen, is published' - - year: 2024 - month: 2 - description: Submitted a paper describing the first ultra-fast picosecond time - resolution PDF study. This opens the door in the future for fully quantitative - femtosecond movies of atomic motions after a laser pulse is applied. - - year: 2024 - month: 2 - description: Developed new data analysis theory and computational tools for - obtaining more reliable PDFs from laboratory diffractometers, opening the - method to a much broader community of chemists, materials scientists and engineers - - year: 2023 - month: 10 - description: Major new funding secured to study in situ and operando nanostructure - at neutron sources at ORNL. - - year: 2022 - month: 8 - description: Three key papers came out describing novel use of machine learning, - including generative ML, to extract structural information from scattering - data. One used interpretive ML to find chemically plausible clusters that - fit low-information local structural measurements, another used a convolutional - variable autoencoder to do structure solution from atomic pair distribution - function (PDF) data for the first time, albeit on a limited class of close-packed - structures. - - year: 2019 - month: 12 - description: Columbia files a provisional patent for the MOSY concept which - is a graph theoretic way of representing tasks accomplished by human actions. - name: Solutions to the nanostructure problem - summary: 'This is an ongoing project to develop methods for solving the nanostructure - problem: the problem of determining the positions of atoms in nanosized clusters, - that have many applications in advanced technologies.' + highlights: [] + name: AI/ML search for improved gas separation membranes + summary: Apply AI and machine learning to get better gas separation membranes team: - name: sbillinge position: Co-PI @@ -169,10 +140,11 @@ dmrefcheme16: position: pi begin_year: 2018 type: funded - website: '' + website: "" education_outreach_inclusion: active: true - description: Activities in the group in education, outreach as well as diversity, + description: + Activities in the group in education, outreach as well as diversity, equity and inclusion. grants: - collab22 @@ -180,77 +152,29 @@ education_outreach_inclusion: highlights: - year: 2023 month: 6 - description: Ran the 4th JUAMI school in Nairobi Kenya. JUAMI is an NSF funded + description: + Ran the 4th JUAMI school in Nairobi Kenya. JUAMI is an NSF funded capacity building activity between the US materials research community and the community in East Africa. It is explained in the video at https://www.juami.org/about/ - year: 2024 month: 2 - description: Curriculum development of a course in machine learning in materials + description: + Curriculum development of a course in machine learning in materials science The course is being given for the first time in spring 2024 as MSAE4990. name: Education, outreach, diversity, equity and inclusion - summary: Activities in the group in education, outreach as well as diversity, equity + summary: + Activities in the group in education, outreach as well as diversity, equity and inclusion. team: - name: sbillinge - position: mentor + position: PI begin_year: 2016 type: outreach - website: https://thebillingegroup.com -fwp17: - active: false - description: Searching for local structural effects in strongly correlated electron - materials using x-ray and neutron scattering - grants: - - fwp17 - group: bg - highlights: - - year: 2021 - month: 1 - description: Found our previously discovered phenomenon of orbital degeneracy - lifting (ODL), which results in local symmetry breaking in quantum materials, - in a series of new materials establishing its ubiquity. - - year: 2020 - month: 1 - description: Paper published in Science Advances reporting the observation of - a novel melting mechanism for gold from an ultrafast time-resolved diffraction - experiment. A femto-second laser pulse dumps energy into a gold thin film - and we probed how it responded on the ps - ns timescale using ultrafast diffraction - pulses. - - year: 2019 - month: 12 - description: Two editor's choice papers in Phys. Rev. B that report ODL symmetry - breaking in the "hot" superconducting material iron-selenide (FeSe). - - year: 2019 - month: 9 - description: Paper published in Nature Communications that reports the discovery - of symmetry breaking due to orbital degeneracy lifting (ODL) in a transition - metal system in a completely unexpected way. - - year: 2019 - month: 3 - description: A paper was accepted in Physical Review Letters that was the first - measurement of phonons in a well characterized semiconducting nanoparticle - system. This experimnet was a tour de force and a collaboration with Jon - Owen in Chemistry at Columbia, as well as researchers at Argonne National - Laboratory. - name: 'DOE funded project at Brookhaven National Laboratory: Local structure of - complex transition metal systems' - team: - - name: sbillinge - position: PI - begin_year: 2008 - end_year: 2022 - - name: ebozin - position: co-pi - begin_year: 2017 - end_year: 2022 - - name: rkoch - position: postdoc - begin_year: 2018 - end_year: 2020 - type: funded + website: "" gtasr: active: true - description: Pretty much everything is a big thing made of small things. This project + description: + Pretty much everything is a big thing made of small things. This project explores fundamental properties that this structure imbues on a system. Because of its ubiquity (that pretty much everything has this structure) any properties that come from this general structure will need to be satisfied in a broad range @@ -258,7 +182,8 @@ gtasr: the results to a broad range of situations. highlights: [] name: A general theory of abstract system responses (gTASR) - summary: This project is to develop a fundamental understanding of complex system + summary: + This project is to develop a fundamental understanding of complex system responses to forces and generalize the observations to a wide range of situations in human life. team: @@ -269,28 +194,32 @@ gtasr: website: https://github.com/diffpy local_symmetry_breaking: active: true - description: 'Quantum materials are at the heart of emerging technologies from quantum + description: + "Quantum materials are at the heart of emerging technologies from quantum computing to sustainable energy. Using local-structural methods we have developed in the group, we have found that many quantum materials exhibit local symmetry breaking: the atomic arrangements at the local scale are different to those implied by the average crystal structure. In many systems this response is intrinsic, though it can also be driven by extrinsic effects that can be controlled. This project aims to characterize these difficult to measure phenomena and, further, - to understand how the local symmetry breaking affects the material properties.' + to understand how the local symmetry breaking affects the material properties." grants: - dmref19 + - fwp17 group: bg highlights: - year: 2022 month: 9 - description: Demonstrated the ability to study texture (crystalline preferred + description: + Demonstrated the ability to study texture (crystalline preferred orientation) in the atomic PDF for the first time. This was not previously possible. It was done using a novel approach of measuring a full 3D PDF using synchrotron x-radiation and using equations developed in the group in 2018 to separate the signal into a structural part and an orientational part. - year: 2023 month: 3 - description: Developed a novel variable-shutter atomic pair distribution function + description: + Developed a novel variable-shutter atomic pair distribution function (vsPDF) technique that allowed us to use neutrons to image fluctuating local structural domains in important thermoelectric material GeTe. This allowed us to develop a new theory for understanding the origin of the very low thermal @@ -299,50 +228,74 @@ local_symmetry_breaking: - year: 2019 month: 6 description: Received funding + - year: 2021 + month: 1 + description: + Found our previously discovered phenomenon of orbital degeneracy + lifting (ODL), which results in local symmetry breaking in quantum materials, + in a series of new materials establishing its ubiquity. + - year: 2020 + month: 1 + description: + Paper published in Science Advances reporting the observation of + a novel melting mechanism for gold from an ultrafast time-resolved diffraction + experiment. A femto-second laser pulse dumps energy into a gold thin film + and we probed how it responded on the ps - ns timescale using ultrafast diffraction + pulses. + - year: 2019 + month: 12 + description: + Two editor's choice papers in Phys. Rev. B that report ODL symmetry + breaking in the "hot" superconducting material iron-selenide (FeSe). + - year: 2019 + month: 9 + description: + Paper published in Nature Communications that reports the discovery + of symmetry breaking due to orbital degeneracy lifting (ODL) in a transition + metal system in a completely unexpected way. + - year: 2019 + month: 3 + description: + A paper was accepted in Physical Review Letters that was the first + measurement of phonons in a well characterized semiconducting nanoparticle + system. This experimnet was a tour de force and a collaboration with Jon + Owen in Chemistry at Columbia, as well as researchers at Argonne National + Laboratory. name: Local symmetry breaking in quantum materials - summary: This project studies the under-appreciated role of intrinsic and extrinsic + summary: + This project studies the under-appreciated role of intrinsic and extrinsic local symmetry breaking in quantum materials using advanced x-ray, neutron and electron scattering methods. team: - name: sbillinge - position: mentor + position: PI begin_year: 2016 type: research - website: https://thebillingegroup.com + website: "" matsci_colab: - active: true - description: This project will build a one-semester accelerated course in machine - learning (ML) applied to STEM research problems for graduate and senior undergraduate - level students in physical science and engineering. Whilst data analytics, artificial - intelligence (AI) and machine learning have revolutionized many aspects of life - from commerce to politics and human interactivity, its impacts in the physical - sciences have been slower. This is now rapidly changing, with disruptive use - of AI and ML to attack research problems that seemed unreachable just a few years - ago. There is a strong hunger among physical science and engineering students - to attain new research skills enabled by ML and to apply them to problems in their - domain. The current training for this cohort of students includes extensive training - in calculus, differential equations, a bit of linear algebra, rudimentary statistics - and nothing in the way of ML. The students often have an aptitude and some training - in basic programming. Our goal with this project is to design a course that will - build on the solid math skills of this group to accelerate their understanding - and ability to adopt ML in their work. The course will build on this existing - knowledge by combining a highly compressed introduction to ML taught by the Statistics - department, but with a focus on newly developed hands-on projects (that we call - Labs) where real (published) problems that apply different ML methods to materials - science problems. The Labs will be developed using the Google Colab platform - and use the actual datasets from the published examples. The developments will - be made following strict standards for coding and documentation and maintained - in a central GitHub repository. As such, it will form a platform that will allow - further Labs to be developed in the aread of Materials Science but also more broadly. Indeed, - this can turn into a widely used community resource that is under continuous development - using standard open-source community development practices. + active: false + description: + This project built a one-semester accelerated course in machine learning + (ML) applied to STEM research problems for graduate and senior undergraduate level + students in physical science and engineering. group: bg highlights: + - year: 2023 + month: 9 + description: + Taught the course for the first time as MSAE4990. Limited to 8 + students, but was well received by the students. + - year: 2023 + month: 4 + description: + Developed a GitHub repo containing the educational examples ("edexes") + used in the course. - year: 2022 month: 5 - description: Proposal funded to develop course to teach ML to physical science + description: + Proposal funded to develop course to teach ML to physical science and engineering students in Columbia Engineering - name: Applying machine learning to valuable data in the physical sciences and engineering + name: Machine Learning for Materials Science course development team: - name: sbillinge position: PI @@ -354,17 +307,18 @@ matsci_colab: position: Co-PI begin_year: 2022 type: teaching - website: '' + website: "" mrsec14: active: false - description: Study of nanoparticle assemblies and superatom systems + description: Study of nanoparticle assemblies and superatom systems using PDF analysis grants: - mrsec14 group: bg highlights: - year: 2018 month: 6 - description: Published a Nature Communications paper where we characterized + description: + Published a Nature Communications paper where we characterized the structure of nanoparticles arranged into a metal-organic expanded structure. name: Assembly of superatoms team: @@ -407,14 +361,15 @@ mrsec14: type: funded nanostructure_problem: active: true - description: 'Material properties depend sensitively on the arrangements of the + description: + "Material properties depend sensitively on the arrangements of the atoms. The atomic structure. For around 100 years we have had x-ray based tools and algorithms that allow us to solve the structure of crystalline materials: crystallography. Increasingly, we want to engineer materials at the nanoscale to give us different properties, but the nanostructure problem, the ability to solve the structure of nanomaterials, remains an unsolved problem in general. This decadal project develops experiments, algorithms, and computational methods to - tackle the nanostructure problem.' + tackle the nanostructure problem." grants: - doeneutron23 - efrc18 @@ -422,14 +377,63 @@ nanostructure_problem: highlights: - year: 2024 month: 3 - description: 'Publication in ChemComm was selected to feature in an online collection + description: + "Publication in ChemComm was selected to feature in an online collection highlighting 60 pioneering historic papers from North America. The paper is Beyond crystallography: the study of disorder, nanocrystallinity and crystallographically - challenged materials with pair distribution functions (DOI:10.1039/b309577k)' + challenged materials with pair distribution functions (DOI:10.1039/b309577k)" - year: 2024 month: 3 - description: 3.3.1 Release of diffpy.utils software with new diffraction software + description: + 3.3.1 Release of diffpy.utils software with new diffraction software objects, new precise interpolation function, and bugfixes. + - year: 2024 + month: 2 + description: + In this period, I won one award, published one book, 14 peer reviewed + articles, gave 7 plenary/keynote, 15 invited talks, and 9 seminars and colloquia + - year: 2024 + month: 2 + description: + "Co-edited a perspective on Machine Learning in Materials Science + published in Acta Crystallographica: Section A" + - year: 2024 + month: 2 + description: + 'The book "Atomic Pair Distribution Function Analysis: A Primer + authored by Simon Billinge and Kirsten Jensen, is published' + - year: 2024 + month: 2 + description: + Submitted a paper describing the first ultra-fast picosecond time + resolution PDF study. This opens the door in the future for fully quantitative + femtosecond movies of atomic motions after a laser pulse is applied. + - year: 2024 + month: 2 + description: + Developed new data analysis theory and computational tools for + obtaining more reliable PDFs from laboratory diffractometers, opening the + method to a much broader community of chemists, materials scientists and engineers + - year: 2023 + month: 10 + description: + Major new funding secured to study in situ and operando nanostructure + at neutron sources at ORNL. + - year: 2022 + month: 8 + description: + Three key papers came out describing novel use of machine learning, + including generative ML, to extract structural information from scattering + data. One used interpretive ML to find chemically plausible clusters that + fit low-information local structural measurements, another used a convolutional + variable autoencoder to do structure solution from atomic pair distribution + function (PDF) data for the first time, albeit on a limited class of close-packed + structures. + - year: 2019 + month: 12 + description: + Columbia files a provisional patent for the MOSY concept which + is a graph theoretic way of representing tasks accomplished by human actions. name: The nanostructure inverse problem team: - name: sbillinge @@ -438,7 +442,8 @@ nanostructure_problem: type: funded pytentiostat: active: false - description: Outreach project to develop an affordable potentiostat based on the + description: + Outreach project to develop an affordable potentiostat based on the arduino platform, and hands-on workshops to train users in East Africa to use it. logo: https://github.com/juami.png @@ -447,7 +452,7 @@ pytentiostat: repo: https://github.com/juami/pytentiostat team: - name: sbillinge - position: mentor + position: PI begin_year: 2016 type: research - website: https://thebillingegroup.com + website: ""