I know it is an excessive demand to directly ask others to share their knowledge, and kind of rude also.
\n\nBut when I go through videos and papers explaining what constant coupling is, there comes just more relative work that I should know to understand the stuff I'm looking at. And the worst thing is I cant tell if I should go further because of lack of domain knowledge.
\n\nOf course there are some feature engineering trick to kick off the model as a black box, but I dont think this is what the competition meant to be.
\n\nSo I decide to give it a try here to see if someone will kindly share some information.\nI'll be really grateful if someone can provide some intuitive explaination of NMR, coupleing constant, or a proper roadmap to get to know the backgroud knowledge.
\n\nI'm sorry for my rudeness, but plz give me a hand If you might.
", "messages": [ { "id": "555627", "postDate": "06/19/2019 07:06:11", "content": "I know it is an excessive demand to directly ask others to share their knowledge, and kind of rude also.
\n\nBut when I go through videos and papers explaining what constant coupling is, there comes just more relative work that I should know to understand the stuff I'm looking at. And the worst thing is I cant tell if I should go further because of lack of domain knowledge.
\n\nOf course there are some feature engineering trick to kick off the model as a black box, but I dont think this is what the competition meant to be.
\n\nSo I decide to give it a try here to see if someone will kindly share some information.\nI'll be really grateful if someone can provide some intuitive explaination of NMR, coupleing constant, or a proper roadmap to get to know the backgroud knowledge.
\n\nI'm sorry for my rudeness, but plz give me a hand If you might.
", "rawMarkdown": "I know it is an excessive demand to directly ask others to share their knowledge, and kind of rude also.\n\nBut when I go through videos and papers explaining what constant coupling is, there comes just more relative work that I should know to understand the stuff I'm looking at. And the worst thing is I cant tell if I should go further because of lack of domain knowledge.\n\nOf course there are some feature engineering trick to kick off the model as a black box, but I dont think this is what the competition meant to be. \n\nSo I decide to give it a try here to see if someone will kindly share some information.\nI'll be really grateful if someone can provide some intuitive explaination of NMR, coupleing constant, or a proper roadmap to get to know the backgroud knowledge.\n\nI'm sorry for my rudeness, but plz give me a hand If you might.", "votes": null }, { "id": "555840", "postDate": "06/19/2019 14:02:24", "content": "In order to understand the coupling constant at a basic level, I would suggest starting with the Khan Academy section on NMR spectrography: https://www.khanacademy.org/science/organic-chemistry/spectroscopy-jay#proton-nmr
\n\nEssentially what we're trying to predict is the difference between NMR readings on a spectrogram for different atoms in a molecule. The Khan Academy lays out pretty well (a) how a spectrogram works, (b) how the magnetic field works, and (c) what the constant represents. Hope it helps!
", "rawMarkdown": "In order to understand the coupling constant at a basic level, I would suggest starting with the Khan Academy section on NMR spectrography: https://www.khanacademy.org/science/organic-chemistry/spectroscopy-jay#proton-nmr\n\nEssentially what we're trying to predict is the difference between NMR readings on a spectrogram for different atoms in a molecule. The Khan Academy lays out pretty well (a) how a spectrogram works, (b) how the magnetic field works, and (c) what the constant represents. Hope it helps!", "votes": null }, { "id": "556025", "postDate": "06/19/2019 18:32:29", "content": "Why not to build a team, that is the aim of allowing teams. If you don't mind we can make a team together. I some domain knowledge as subsidiary part of my study, and I have little programming knowledge, but I am new to python 1 month knowledge.
\nMessage me If you agree
I'm happy to take a stab at explaining this at a rudimentary level. The atoms inside of a molecule have a core nucleus (that contains protons and neutrons) and are then surrounded by electrons which are delocalized and shared across all the atoms. The protons and neutrons inside atoms have a quantum mechanical feature called spin. For many nuclei, the net spin of all the protons and neutrons in the nucleus end up being zero, but several nuclei end up with a net permanent spin (like protons (H), C13 and N14).
\n\nThe name spin is confusing and misguided, but it's been around for 95 years so it's what we go with. On a basic physical level, spin is an intrinsic angular momentum (for a quick refresher, angular momentum tells you how much an object is spinning about a particular axis). If we place an object with spin in a static magnetic field (think placing a sample in a high-powered superconducting magnetic coil), its overall energy will depend on whether the spin (which is a vector) points in the same direction as the applied magnetic field or opposite to it. This creates two distinct energy levels (technically it just makes the two energy levels which were always there have 2 separate energies). This energy difference is proportional the total magnetic field felt at the nucleus, and can be measured very precisely by noting what wavelength of light (which happens to be in the radio frequency) the nucleus will absorb. This energy difference is the primary featured measured by an NMR or MRI machine.
\n\nNow, according to the laws of electricity and magnetism, a charged particle with angular momentum produces a magnetic field. That small magnetic field is what's important here. Let's imagine now that instead of a single nucleus in a magnetic field, we have two nuclei, both with 2 non-zero spins, and both in the same external magnetic field. To a zeroth order approximation, they would absorb the same wavelength of light. However, remember that each nucleus produces its own tiny magnetic field depending on whether its spin is aligned or antialigned with the large applied magnet. So if I'm Nucleus A, the magnetic field I feel is not just the big magnetic field B0, it is the sum of the big field plus the little field produced by nucleus B (B=B0+B_B). Similarly, the magnetic field at Nucleus B is B0 + B_A. So the wavelength of a photon that nucleus A will absorb depends on, or is coupled to, the state of Nucleus B. That is the fundamental relationship when we say the A-B Coupling constant. It's a measure of how strongly the state of nucleus A affects the energy levels of Nucleus B. In all cases its by changing the local, microscopic electric field. The amount of field that that nucleus A feels from nucleus B depends on a lot of factors. Some are geometric (How far is A from B). Some are electronic in nature (the electrons in the molecule can shield or protect Nucleus A from feeling the full magnetic field of Nucleus B).
\n\nIs that helpful at all? I'm happy to answer questions. \nCheers,\nJames
", "rawMarkdown": "I'm happy to take a stab at explaining this at a rudimentary level. The atoms inside of a molecule have a core nucleus (that contains protons and neutrons) and are then surrounded by electrons which are delocalized and shared across all the atoms. The protons and neutrons inside atoms have a quantum mechanical feature called spin. For many nuclei, the net spin of all the protons and neutrons in the nucleus end up being zero, but several nuclei end up with a net permanent spin (like protons (H), C13 and N14). \n\nThe name spin is confusing and misguided, but it's been around for 95 years so it's what we go with. On a basic physical level, spin is an intrinsic angular momentum (for a quick refresher, angular momentum tells you how much an object is spinning about a particular axis). If we place an object with spin in a static magnetic field (think placing a sample in a high-powered superconducting magnetic coil), its overall energy will depend on whether the spin (which is a vector) points in the same direction as the applied magnetic field or opposite to it. This creates two distinct energy levels (technically it just makes the two energy levels which were always there have 2 separate energies). This energy difference is proportional the total magnetic field felt at the nucleus, and can be measured very precisely by noting what wavelength of light (which happens to be in the radio frequency) the nucleus will absorb. This energy difference is the primary featured measured by an NMR or MRI machine.\n\n Now, according to the laws of electricity and magnetism, a charged particle with angular momentum produces a magnetic field. That small magnetic field is what's important here. Let's imagine now that instead of a single nucleus in a magnetic field, we have two nuclei, both with 2 non-zero spins, and both in the same external magnetic field. To a zeroth order approximation, they would absorb the same wavelength of light. However, remember that each nucleus produces its own tiny magnetic field depending on whether its spin is aligned or antialigned with the large applied magnet. So if I'm Nucleus A, the magnetic field I feel is not just the big magnetic field B0, it is the sum of the big field plus the little field produced by nucleus B (B=B0+B_B). Similarly, the magnetic field at Nucleus B is B0 + B_A. So the wavelength of a photon that nucleus A will absorb depends on, or is coupled to, the state of Nucleus B. That is the fundamental relationship when we say the A-B Coupling constant. It's a measure of how strongly the state of nucleus A affects the energy levels of Nucleus B. In all cases its by changing the local, microscopic electric field. The amount of field that that nucleus A feels from nucleus B depends on a lot of factors. Some are geometric (How far is A from B). Some are electronic in nature (the electrons in the molecule can shield or protect Nucleus A from feeling the full magnetic field of Nucleus B). \n\nIs that helpful at all? I'm happy to answer questions. \nCheers,\nJames", "votes": null }, { "id": "556229", "postDate": "06/20/2019 02:35:38", "content": "That's really clear explaination helps a lot, thanks!
\n\nFrom my understanding, all nucleus in a molecular effect each other, it can be costly to measure or infer coupling constant on each pair of particles in a big molecular, while atom type and structure of molecular is easier to maintain. So it can be helpful to infer coupling constant via those data provided in this competition. Is that the point of holding this competition I suppose?
", "rawMarkdown": "That's really clear explaination helps a lot, thanks!\n\nFrom my understanding, all nucleus in a molecular effect each other, it can be costly to measure or infer coupling constant on each pair of particles in a big molecular, while atom type and structure of molecular is easier to maintain. So it can be helpful to infer coupling constant via those data provided in this competition. Is that the point of holding this competition I suppose?", "votes": null }, { "id": "556231", "postDate": "06/20/2019 02:38:26", "content": "Thanks a lot for sharing the course, I'll go through it.
", "rawMarkdown": "Thanks a lot for sharing the course, I'll go through it.", "votes": null }, { "id": "556237", "postDate": "06/20/2019 02:47:28", "content": "I'm glad to, but here's my condition you should know about.
\n\nI'm just starting up and not an experienced kaggler(although i register 2 years ago, afk all these days).\nI'm an algorithm engineer so I can do the programming part, but it may take one or two week for me to get familiar with kaggle way. \nThats what i got, so let me know if you still want to team up.
", "rawMarkdown": "I'm glad to, but here's my condition you should know about.\n\nI'm just starting up and not an experienced kaggler(although i register 2 years ago, afk all these days).\nI'm an algorithm engineer so I can do the programming part, but it may take one or two week for me to get familiar with kaggle way. \nThats what i got, so let me know if you still want to team up.", "votes": null }, { "id": "556838", "postDate": "06/20/2019 18:27:53", "content": "No problem @littlebeandog , sure I still want to team up, I like teams :) \"TWO BRAINS ARE ALWAYS BETTER THAN ONE BRAIN\" But we have to find an easy contact way to share our thoughts I prefer facebook messenger \" mohamed.a.gad \"
", "rawMarkdown": "No problem @littlebeandog , sure I still want to team up, I like teams :) \"TWO BRAINS ARE ALWAYS BETTER THAN ONE BRAIN\" But we have to find an easy contact way to share our thoughts I prefer facebook messenger \" mohamed.a.gad \"", "votes": null }, { "id": "557277", "postDate": "06/21/2019 03:40:12", "content": "Too many mohamed.a.gad, i'm not sure which is you. I havn't used facebook before, you may message me on this newly register account
\n\n\n", "rawMarkdown": "Too many mohamed.a.gad, i'm not sure which is you. I havn't used facebook before, you may message me on this newly register account\n> https://www.facebook.com/profile.php?id=100038234835750\n> Rick Sanchez", "votes": null }, { "id": "557356", "postDate": "06/21/2019 04:55:37", "content": "https://www.facebook.com/profile.php?id=100038234835750\n Rick Sanchez
\n
Thanks for asking the question that I'm going to ask myself. =)
", "rawMarkdown": "Thanks for asking the question that I'm going to ask myself. =)", "votes": null }, { "id": "558152", "postDate": "06/21/2019 22:24:10", "content": "Hi, @littlebeandog I have sent you friend request
", "rawMarkdown": "Hi, @littlebeandog I have sent you friend request", "votes": null } ], "comments": [ { "id": 555840, "author_name": "sharkbait1223", "author_url": "", "post_date": "06/19/2019 14:02:24", "content": "In order to understand the coupling constant at a basic level, I would suggest starting with the Khan Academy section on NMR spectrography: https://www.khanacademy.org/science/organic-chemistry/spectroscopy-jay#proton-nmr
\n\nEssentially what we're trying to predict is the difference between NMR readings on a spectrogram for different atoms in a molecule. The Khan Academy lays out pretty well (a) how a spectrogram works, (b) how the magnetic field works, and (c) what the constant represents. Hope it helps!
", "votes": null, "replies": [ { "id": 556231, "author_name": "littlebeandog", "author_url": "", "post_date": "06/20/2019 02:38:26", "content": "Thanks a lot for sharing the course, I'll go through it.
", "votes": null, "replies": [] } ] }, { "id": 556025, "author_name": "mohamed0gad", "author_url": "", "post_date": "06/19/2019 18:32:29", "content": "Why not to build a team, that is the aim of allowing teams. If you don't mind we can make a team together. I some domain knowledge as subsidiary part of my study, and I have little programming knowledge, but I am new to python 1 month knowledge.
\nMessage me If you agree
I'm glad to, but here's my condition you should know about.
\n\nI'm just starting up and not an experienced kaggler(although i register 2 years ago, afk all these days).\nI'm an algorithm engineer so I can do the programming part, but it may take one or two week for me to get familiar with kaggle way. \nThats what i got, so let me know if you still want to team up.
", "votes": null, "replies": [] }, { "id": 556838, "author_name": "mohamed0gad", "author_url": "", "post_date": "06/20/2019 18:27:53", "content": "No problem @littlebeandog , sure I still want to team up, I like teams :) \"TWO BRAINS ARE ALWAYS BETTER THAN ONE BRAIN\" But we have to find an easy contact way to share our thoughts I prefer facebook messenger \" mohamed.a.gad \"
", "votes": null, "replies": [] }, { "id": 557277, "author_name": "littlebeandog", "author_url": "", "post_date": "06/21/2019 03:40:12", "content": "Too many mohamed.a.gad, i'm not sure which is you. I havn't used facebook before, you may message me on this newly register account
\n\n\n", "votes": null, "replies": [] }, { "id": 558152, "author_name": "mohamed0gad", "author_url": "", "post_date": "06/21/2019 22:24:10", "content": "https://www.facebook.com/profile.php?id=100038234835750\n Rick Sanchez
\n
Hi, @littlebeandog I have sent you friend request
", "votes": null, "replies": [] } ] }, { "id": 556070, "author_name": "jjohnsumn", "author_url": "", "post_date": "06/19/2019 19:59:17", "content": "I'm happy to take a stab at explaining this at a rudimentary level. The atoms inside of a molecule have a core nucleus (that contains protons and neutrons) and are then surrounded by electrons which are delocalized and shared across all the atoms. The protons and neutrons inside atoms have a quantum mechanical feature called spin. For many nuclei, the net spin of all the protons and neutrons in the nucleus end up being zero, but several nuclei end up with a net permanent spin (like protons (H), C13 and N14).
\n\nThe name spin is confusing and misguided, but it's been around for 95 years so it's what we go with. On a basic physical level, spin is an intrinsic angular momentum (for a quick refresher, angular momentum tells you how much an object is spinning about a particular axis). If we place an object with spin in a static magnetic field (think placing a sample in a high-powered superconducting magnetic coil), its overall energy will depend on whether the spin (which is a vector) points in the same direction as the applied magnetic field or opposite to it. This creates two distinct energy levels (technically it just makes the two energy levels which were always there have 2 separate energies). This energy difference is proportional the total magnetic field felt at the nucleus, and can be measured very precisely by noting what wavelength of light (which happens to be in the radio frequency) the nucleus will absorb. This energy difference is the primary featured measured by an NMR or MRI machine.
\n\nNow, according to the laws of electricity and magnetism, a charged particle with angular momentum produces a magnetic field. That small magnetic field is what's important here. Let's imagine now that instead of a single nucleus in a magnetic field, we have two nuclei, both with 2 non-zero spins, and both in the same external magnetic field. To a zeroth order approximation, they would absorb the same wavelength of light. However, remember that each nucleus produces its own tiny magnetic field depending on whether its spin is aligned or antialigned with the large applied magnet. So if I'm Nucleus A, the magnetic field I feel is not just the big magnetic field B0, it is the sum of the big field plus the little field produced by nucleus B (B=B0+B_B). Similarly, the magnetic field at Nucleus B is B0 + B_A. So the wavelength of a photon that nucleus A will absorb depends on, or is coupled to, the state of Nucleus B. That is the fundamental relationship when we say the A-B Coupling constant. It's a measure of how strongly the state of nucleus A affects the energy levels of Nucleus B. In all cases its by changing the local, microscopic electric field. The amount of field that that nucleus A feels from nucleus B depends on a lot of factors. Some are geometric (How far is A from B). Some are electronic in nature (the electrons in the molecule can shield or protect Nucleus A from feeling the full magnetic field of Nucleus B).
\n\nIs that helpful at all? I'm happy to answer questions. \nCheers,\nJames
", "votes": null, "replies": [ { "id": 556229, "author_name": "littlebeandog", "author_url": "", "post_date": "06/20/2019 02:35:38", "content": "That's really clear explaination helps a lot, thanks!
\n\nFrom my understanding, all nucleus in a molecular effect each other, it can be costly to measure or infer coupling constant on each pair of particles in a big molecular, while atom type and structure of molecular is easier to maintain. So it can be helpful to infer coupling constant via those data provided in this competition. Is that the point of holding this competition I suppose?
", "votes": null, "replies": [] } ] }, { "id": 557356, "author_name": "dexter6855", "author_url": "", "post_date": "06/21/2019 04:55:37", "content": "Thanks for asking the question that I'm going to ask myself. =)
", "votes": null, "replies": [] } ], "raw_markdown_by_id": { "555627": "I know it is an excessive demand to directly ask others to share their knowledge, and kind of rude also.\n\nBut when I go through videos and papers explaining what constant coupling is, there comes just more relative work that I should know to understand the stuff I'm looking at. And the worst thing is I cant tell if I should go further because of lack of domain knowledge.\n\nOf course there are some feature engineering trick to kick off the model as a black box, but I dont think this is what the competition meant to be. \n\nSo I decide to give it a try here to see if someone will kindly share some information.\nI'll be really grateful if someone can provide some intuitive explaination of NMR, coupleing constant, or a proper roadmap to get to know the backgroud knowledge.\n\nI'm sorry for my rudeness, but plz give me a hand If you might.", "555840": "In order to understand the coupling constant at a basic level, I would suggest starting with the Khan Academy section on NMR spectrography: https://www.khanacademy.org/science/organic-chemistry/spectroscopy-jay#proton-nmr\n\nEssentially what we're trying to predict is the difference between NMR readings on a spectrogram for different atoms in a molecule. The Khan Academy lays out pretty well (a) how a spectrogram works, (b) how the magnetic field works, and (c) what the constant represents. Hope it helps!", "556025": "Why not to build a team, that is the aim of allowing teams. If you don't mind we can make a team together. I some domain knowledge as subsidiary part of my study, and I have little programming knowledge, but I am new to python 1 month knowledge. \nMessage me If you agree", "556070": "I'm happy to take a stab at explaining this at a rudimentary level. The atoms inside of a molecule have a core nucleus (that contains protons and neutrons) and are then surrounded by electrons which are delocalized and shared across all the atoms. The protons and neutrons inside atoms have a quantum mechanical feature called spin. For many nuclei, the net spin of all the protons and neutrons in the nucleus end up being zero, but several nuclei end up with a net permanent spin (like protons (H), C13 and N14). \n\nThe name spin is confusing and misguided, but it's been around for 95 years so it's what we go with. On a basic physical level, spin is an intrinsic angular momentum (for a quick refresher, angular momentum tells you how much an object is spinning about a particular axis). If we place an object with spin in a static magnetic field (think placing a sample in a high-powered superconducting magnetic coil), its overall energy will depend on whether the spin (which is a vector) points in the same direction as the applied magnetic field or opposite to it. This creates two distinct energy levels (technically it just makes the two energy levels which were always there have 2 separate energies). This energy difference is proportional the total magnetic field felt at the nucleus, and can be measured very precisely by noting what wavelength of light (which happens to be in the radio frequency) the nucleus will absorb. This energy difference is the primary featured measured by an NMR or MRI machine.\n\n Now, according to the laws of electricity and magnetism, a charged particle with angular momentum produces a magnetic field. That small magnetic field is what's important here. Let's imagine now that instead of a single nucleus in a magnetic field, we have two nuclei, both with 2 non-zero spins, and both in the same external magnetic field. To a zeroth order approximation, they would absorb the same wavelength of light. However, remember that each nucleus produces its own tiny magnetic field depending on whether its spin is aligned or antialigned with the large applied magnet. So if I'm Nucleus A, the magnetic field I feel is not just the big magnetic field B0, it is the sum of the big field plus the little field produced by nucleus B (B=B0+B_B). Similarly, the magnetic field at Nucleus B is B0 + B_A. So the wavelength of a photon that nucleus A will absorb depends on, or is coupled to, the state of Nucleus B. That is the fundamental relationship when we say the A-B Coupling constant. It's a measure of how strongly the state of nucleus A affects the energy levels of Nucleus B. In all cases its by changing the local, microscopic electric field. The amount of field that that nucleus A feels from nucleus B depends on a lot of factors. Some are geometric (How far is A from B). Some are electronic in nature (the electrons in the molecule can shield or protect Nucleus A from feeling the full magnetic field of Nucleus B). \n\nIs that helpful at all? I'm happy to answer questions. \nCheers,\nJames", "556229": "That's really clear explaination helps a lot, thanks!\n\nFrom my understanding, all nucleus in a molecular effect each other, it can be costly to measure or infer coupling constant on each pair of particles in a big molecular, while atom type and structure of molecular is easier to maintain. So it can be helpful to infer coupling constant via those data provided in this competition. Is that the point of holding this competition I suppose?", "556231": "Thanks a lot for sharing the course, I'll go through it.", "556237": "I'm glad to, but here's my condition you should know about.\n\nI'm just starting up and not an experienced kaggler(although i register 2 years ago, afk all these days).\nI'm an algorithm engineer so I can do the programming part, but it may take one or two week for me to get familiar with kaggle way. \nThats what i got, so let me know if you still want to team up.", "556838": "No problem @littlebeandog , sure I still want to team up, I like teams :) \"TWO BRAINS ARE ALWAYS BETTER THAN ONE BRAIN\" But we have to find an easy contact way to share our thoughts I prefer facebook messenger \" mohamed.a.gad \"", "557277": "Too many mohamed.a.gad, i'm not sure which is you. I havn't used facebook before, you may message me on this newly register account\n> https://www.facebook.com/profile.php?id=100038234835750\n> Rick Sanchez", "557356": "Thanks for asking the question that I'm going to ask myself. =)", "558152": "Hi, @littlebeandog I have sent you friend request" }, "source": "meta" }