Curious work explaining why deep neural networks work. https://t.co/ygOmd5fFDM
RT @hpcprogrammer: Still working through this paper today on why neural networks work https://t.co/agBqVwBmr1
Reading: "Why does deep and cheap learning work so well?" https://t.co/GmK4jwXdeI via @techreview #ML #Learning #fb
"Why does deep and cheap learning work so well?" https://t.co/gLBFUy17ph
Why does #DeepLearning work so well? Physics. https://t.co/E5TPRDJJ5U via @Harvard and @MIT by Lin et al. https://t.co/RfIR1ZbSeX
RT @fperez_org: Q. for deep learning experts: is Thm in p.4 of https://t.co/bXdEGbQ2Xe really a new result? Neat... Notebook implem: https:…
RT @kshameer: Why does deep and cheap learning work so well? It's all in the math! https://t.co/cyJfHtzumy #bioinformatics https://t.co…
#elearning #edtech - Why does deep and cheap learning work so well? https://t.co/LszOxx43Hi
RT @fperez_org: Q. for deep learning experts: is Thm in p.4 of https://t.co/bXdEGbQ2Xe really a new result? Neat... Notebook implem: https:…
繰り込み群と機械学習の関係って結構研究されてるみたいだなぁ。。"the connection between RG and machine learning has been studied...repeatedly" https://t.co/UxIlzs5EFd
RT @tmaehara: 紹介文がアレなだけかと思ったが,元論文https://t.co/llvpeQk3oxもアレだWe show how the success of deep learning depends not only on mathematics but al…
こいつの論文他で読んだところだった。 https://t.co/rrPzaVQ16l
RT @fperez_org: Q. for deep learning experts: is Thm in p.4 of https://t.co/bXdEGbQ2Xe really a new result? Neat... Notebook implem: https:…
RT @fperez_org: Q. for deep learning experts: is Thm in p.4 of https://t.co/bXdEGbQ2Xe really a new result? Neat... Notebook implem: https:…
RT @tegmark: Why does #deeplearning work so well? Part of the answer is physics! So say Henry & Max: https://t.co/uejC7mJcJ1 https://t.co/…
RT @moorejh: Why does #deeplearning work so well? https://t.co/e1oj3s10Ix #machinelearning #datascience https://t.co/Pm336GS39m
RT @moorejh: Why does #deeplearning work so well? https://t.co/e1oj3s10Ix #machinelearning #datascience https://t.co/Pm336GS39m
RT @moorejh: Why does #deeplearning work so well? https://t.co/e1oj3s10Ix #machinelearning #datascience https://t.co/Pm336GS39m
RT @fperez_org: Q. for deep learning experts: is Thm in p.4 of https://t.co/bXdEGbQ2Xe really a new result? Neat... Notebook implem: https:…
RT @fperez_org: Q. for deep learning experts: is Thm in p.4 of https://t.co/bXdEGbQ2Xe really a new result? Neat... Notebook implem: https:…
Lin and Tegmark argue #deeplearning works because it admits a Hamiltonian formulation. https://t.co/7X0WqWgjpM https://t.co/nG249kdwsz
Why does #deeplearning work so well? https://t.co/e1oj3s10Ix #machinelearning #datascience https://t.co/Pm336GS39m
RT @fperez_org: Q. for deep learning experts: is Thm in p.4 of https://t.co/bXdEGbQ2Xe really a new result? Neat... Notebook implem: https:…
Still working through this paper today on why neural networks work https://t.co/agBqVwBmr1
Properties of "physics-based functions... translate into... simple #neuralnets" https://t.co/YppQU6oUjb -@tegmark https://t.co/lVdQs6j53v
[1608.08225] Why does deep and cheap learning work so well? https://t.co/dijdMKOpcZ
RT @AlisonBLowndes: Henry Lin & @tegmark propose #deeplearning's success comes down to physics #symmetry #locality #compositionality... htt…
Why does deep and cheap leaning work so well? https://t.co/NO5kMbkVx7
RT @fperez_org: Q. for deep learning experts: is Thm in p.4 of https://t.co/bXdEGbQ2Xe really a new result? Neat... Notebook implem: https:…
RT @fperez_org: Q. for deep learning experts: is Thm in p.4 of https://t.co/bXdEGbQ2Xe really a new result? Neat... Notebook implem: https:…
RT @fperez_org: Q. for deep learning experts: is Thm in p.4 of https://t.co/bXdEGbQ2Xe really a new result? Neat... Notebook implem: https:…
RT @cynicalsecurity: "The Extraordinary Link Between Deep Neural Networks and the Nature of the Universe" - arXiv https://t.co/r5uA7AI1l1 h…
https://t.co/6wKBaLXeLQ Lin & Tegmark's arXiv: I thought no big deal there. But then people kept sharing this the whole week. Wasn't this …
RT @fperez_org: Q. for deep learning experts: is Thm in p.4 of https://t.co/bXdEGbQ2Xe really a new result? Neat... Notebook implem: https:…
RT @AlisonBLowndes: Henry Lin & @tegmark propose #deeplearning's success comes down to physics #symmetry #locality #compositionality... htt…
"The Extraordinary Link Between Deep Neural Networks and the Nature of the Universe" - arXiv https://t.co/r5uA7AI1l1 https://t.co/bgCQlNqBJW
pdf https://t.co/RG6dukIHif "simplicity [bc of] properties such as symmetry, locality, compositionality&polynomial log-probability" @mpesce
Success of DNNs is due to laws of physics, not mathematics alone https://t.co/crLBdNKOVl
RT @fperez_org: Q. for deep learning experts: is Thm in p.4 of https://t.co/bXdEGbQ2Xe really a new result? Neat... Notebook implem: https:…
RT @fperez_org: Q. for deep learning experts: is Thm in p.4 of https://t.co/bXdEGbQ2Xe really a new result? Neat... Notebook implem: https:…
Q. for deep learning experts: is Thm in p.4 of https://t.co/bXdEGbQ2Xe really a new result? Neat... Notebook implem: https://t.co/ilNj4SPUGg
RT @gkay92: Why does deep and cheap learning work so well? - Answer in Physics? https://t.co/s3ucx7xyJ6 Full paper: https://t.co/AEXfWjrh…
Why does #deeplearning work so well? https://t.co/ERVjSqjB0G
TL;DR neural networks are more about physics than mathematics https://t.co/Jgvd7z9AtI
Why does Deep and Cheap learning work so well?Explores the link betwn #DeepNeuralNetworks & Physical laws https://t.co/Vp3FUHHVqd
Deep neural networks good at complex problems, b/c the universe governed by a tiny subset of all possible functions https://t.co/hIVTAX9zeL
Some really cool perspective on neural networks by Henry Lin and @tegmark. Definitely a must read. https://t.co/V3wg78mOYS
Some really cool perspective on neural networks by @tegmark. Defnitely a must read. https://t.co/V3wg78mOYS
RT @deliprao: Two physicists walk into a Machine Learning bar, and confuse everyone with their terminology https://t.co/OXICtHdkfK https://…
RT @AlisonBLowndes: Henry Lin & @tegmark propose #deeplearning's success comes down to physics #symmetry #locality #compositionality... htt…
RT @AlisonBLowndes: Henry Lin & @tegmark propose #deeplearning's success comes down to physics #symmetry #locality #compositionality... htt…
Henry Lin & @tegmark propose #deeplearning's success comes down to physics #symmetry #locality #compositionality... https://t.co/zadTEm3YHv
RT @kshameer: Why does deep and cheap learning work so well? It's all in the math! https://t.co/cyJfHtzumy #bioinformatics https://t.co…
RT @kshameer: Why does deep and cheap learning work so well? It's all in the math! https://t.co/cyJfHtzumy #bioinformatics https://t.co…
RT @kshameer: Why does deep and cheap learning work so well? It's all in the math! https://t.co/cyJfHtzumy #bioinformatics https://t.co…
RT @kshameer: Why does deep and cheap learning work so well? It's all in the math! https://t.co/cyJfHtzumy #bioinformatics https://t.co…
RT @kshameer: Why does deep and cheap learning work so well? It's all in the math! https://t.co/cyJfHtzumy #bioinformatics https://t.co…
Why does deep and cheap learning work so well? It's all in the math! https://t.co/cyJfHtzumy #bioinformatics https://t.co/ExvNkdgPUC
Why does deep and cheap learning work so well? #deeplearning Henry W. Lin and Max Tegmark https://t.co/YVJ6z7lth9
Deep-learning only needs to approximate a finite set of properties imposed by physics to work well https://t.co/02hUUjlHbL
RT @gkay92: Why does deep and cheap learning work so well? - Answer in Physics? https://t.co/s3ucx7xyJ6 Full paper: https://t.co/AEXfWjrh…
Why does deep and cheap learning work so well? - Answer in Physics? https://t.co/s3ucx7xyJ6 Full paper: https://t.co/AEXfWjrhNg
Today's read: Why does deep and cheap learning work so well? https://t.co/Nls3FZ1TKj
Why does #deeplearning so well? #MIT #Harvard https://t.co/n1CEGIFrxD
Why does deep learning work so well? Another step towards a better understanding of deep learning networks. https://t.co/U7kjwTMpzh
RT @jackhidary: @tegmark been too long since we both spoke @DLDConference stimulating paper on #deeplearning and #physics https://t.co/xi58…
Why does #deeplearning work so well? A #physics inspired approach by Lin & Tegmark [PDF] https://t.co/ivIk3r2Rs8 #ml #condmat #arxiv
RT @memotv: how can #DeepLearning solve probs w 1e1000000s DoF w only 1e9s params? PHYSICS says @tegmark https://t.co/aYI3Aj6y1y https://t.…
how can #DeepLearning solve probs w 1e1000000s DoF w only 1e9s params? PHYSICS says @tegmark https://t.co/aYI3Aj6y1y https://t.co/RLJX4zds4A
@NietzscheanHobo and this is on my reading list but idk if it's good yet https://t.co/RljdxlaJUG
Why does deep and cheap learning work so well? https://t.co/W0sjBQLYiF
RT @tegmark: Why does #deeplearning work so well? Part of the answer is physics! So say Henry & Max: https://t.co/uejC7mJcJ1 https://t.co/…
RT @janexwang: Why do neural networks work so well? Because physics. A paper after my own heart :) https://t.co/Wo8Hk6uERR
RT @tegmark: Why does #deeplearning work so well? Part of the answer is physics! So say Henry & Max: https://t.co/uejC7mJcJ1 https://t.co/…
RT @davidjayharris: Interesting argument: machine learning "works" on real data because physics is mostly limited to 4-way interactions htt…
RT @davidjayharris: Interesting argument: machine learning "works" on real data because physics is mostly limited to 4-way interactions htt…
RT @davidjayharris: Interesting argument: machine learning "works" on real data because physics is mostly limited to 4-way interactions htt…
RT @davidjayharris: Interesting argument: machine learning "works" on real data because physics is mostly limited to 4-way interactions htt…
@tegmark been too long since we both spoke @DLDConference stimulating paper on #deeplearning and #physics https://t.co/xi58TaASLl
RT @davidjayharris: Interesting argument: machine learning "works" on real data because physics is mostly limited to 4-way interactions htt…
RT @davidjayharris: Interesting argument: machine learning "works" on real data because physics is mostly limited to 4-way interactions htt…
RT @davidjayharris: Interesting argument: machine learning "works" on real data because physics is mostly limited to 4-way interactions htt…
@bolkerb This paper uses mutual info to quantify insufficiency. Apparently renormalization groups minimize it. https://t.co/QxbpGjFnKu
Interesting argument: machine learning "works" on real data because physics is mostly limited to 4-way interactions https://t.co/QxbpGjFnKu
RT @janexwang: Why do neural networks work so well? Because physics. A paper after my own heart :) https://t.co/Wo8Hk6uERR
RT @janexwang: Why do neural networks work so well? Because physics. A paper after my own heart :) https://t.co/Wo8Hk6uERR
RT @janexwang: Why do neural networks work so well? Because physics. A paper after my own heart :) https://t.co/Wo8Hk6uERR
Back at school! #deeplearning explained using statistical mechanics! Looks fascinating! #ML https://t.co/AzleKKDhZ1 https://t.co/Ncqz9WsD8C
RT @tegmark: Why does #deeplearning work so well? Part of the answer is physics! So say Henry & Max: https://t.co/uejC7mJcJ1 https://t.co/…
Why does deep and cheap learning work so well? https://t.co/4JNKbKN5ij - what depth would be enough? Interactions of sparseness etc.
RT @janexwang: Why do neural networks work so well? Because physics. A paper after my own heart :) https://t.co/Wo8Hk6uERR
RT @janexwang: Why do neural networks work so well? Because physics. A paper after my own heart :) https://t.co/Wo8Hk6uERR
RT @v_vashishta: Why does deep and cheap learning work so well? https://t.co/KEW5JRIjvr #MachineLearning
Why does deep and cheap learning work so well? https://t.co/KEW5JRIjvr #MachineLearning
RT @DonskerClass: Tegmark seems to argue you only ever need 32 parameters since you can just simulate Standard Model to get your data http…
RT @tegmark: Why does #deeplearning work so well? Part of the answer is physics! So say Henry & Max: https://t.co/uejC7mJcJ1 https://t.co/…
RT @brandondamos: Why does deep and cheap learning work so well? https://t.co/VW5wZPeCou https://t.co/k3TjOlElbS
RT @janexwang: Why do neural networks work so well? Because physics. A paper after my own heart :) https://t.co/Wo8Hk6uERR