The discussions show strong enthusiasm for the article, with posts labeling it as a groundbreaking development in materials science analysis and AI's role in scientific discovery. Words like 'must read', 'imagine what is possible', and references to its potential impact illustrate the high interest and perceived significance of this research, demonstrating deep engagement and acknowledgment of its large-scale scientific implications.
AgreementStrong agreement
Many discussions praise the article, highlighting its significance and potential for advancing science, reflecting strong support.
InterestHigh interest
The discussions show high curiosity, mentioning specific methods like unsupervised embeddings, and implications such as AI assisting in scientific discovery.
EngagementModerate engagement
Comments reference detailed concepts, example applications, and cite specific studies, indicating active engagement.
ImpactHigh impact
Participants express a belief that this research could revolutionize how scientific knowledge is mined and utilized, emphasizing its transformative potential.
Abstract in brief
Unsupervised word embeddings can effectively extract complex materials science knowledge from research literature without needing human-labeled data, capturing concepts like the periodic table and structure-property relationships.
These embeddings can even predict future material discoveries several years in advance, indicating that valuable latent knowledge about upcoming research is embedded in existing publications.
The study suggests a scalable, generalized approach to mining scientific literature, enabling better understanding and use of the vast amount of unstructured scientific text beyond traditional databases.]
Paper you have to read and share - imagine what is possible in life a science - Unsupervised word embeddings capture latent knowledge from materials science literature | Nature https://t.co/0sVJyR03fL
Oct 25, 2019 · 15 likes · 2 shares · 447 followersView post →
TP
The TWIML AI Podcast
@twimlai · X
Check out episode 291 with Anubhav Jain, Staff Scientist & Chemist at Lawrence Berkeley National Lab (@BerkeleyLab) to discuss his paper, ‘Unsupervised word embeddings capture latent knowledge from materials science literature’. https://t.co/K1qFW3qxPe #materialscience https://t.co/7sda3q4VHC
Aug 17, 2019 · 3 likes · 14,213 followersView post →
AK
Analytic Kant
@KantNot1 · X
@bayesianboy this paper is cool:
“Unsupervised word embeddings capture latent knowledge from materials science literature”
https://t.co/nNiBOJCYWD
Unsupervised word embeddings capture latent knowledge from materials science literature (Nature) - https://t.co/Ph5z04EfvW
Jan 28, 2020 · 2 likes · 3,713 followersView post →
KK
Kohitij Kar
@KohitijKar · X
Using knowledge from past publications to hypothesize and test new stuff — pretty much what we do as scientists. Now by a machine! https://t.co/xj7kaDdidv (might be early days but I bet a version of this will be an off-the-shelf utility in the future!)
Javier Garc�a discusses how automating experiments with robotic chemists can revolutionize material design and research efficiency, highlighting advancements like autonomous testing for photocatalysts and AI-driven…
This video explores how artificial intelligence and robotics are transforming chemistry, offering opportunities for sustainable development and advancements in chemical research. It highlights the use of autonomous…
@EarthML1 @ykilcher @theshawwn @zacharylipton Unsupervised word embeddings capture latent knowledge from materials science literature
https://t.co/mugCUAvh5a
@tprstly @SwipeWright The preservation of element property groupings by word embeddings alone is what first got me interested in this space (extended fig from same paper https://t.co/z2gS2A1xBb). https://t.co/15KlyNA50k
Unsupervised word embeddings capture latent knowledge from materials science literature (Nature) - https://t.co/Ph5z04EfvW
Jan 28, 2020 · 2 likes · 3,713 followersView post →
AS
Andrei Savu
@andreisavu · X
Unsupervised word embeddings capture latent knowledge from materials science literature https://t.co/Y6Gywj3bm7
Nov 24, 2019 · 1 like · 6,656 followersView post →
EN
Elias Najarro
@EIiasNajarro · X
New materials cosily hiding in the latent space of the scientific literature. Word embeddings as AI assistants for science.
Summary: https://t.co/0EePqOiYTI
Paper: https://t.co/M4MRLUOIdJ https://t.co/oNL9bxIgGP
RT @hetzer_martin: Paper you have to read and share - imagine what is possible in life a science - Unsupervised word embeddings capture lat…
Oct 25, 2019 · 2 shares · 2,507 followersView post →
UM
Uri Manor
@manorlaboratory · X
RT @hetzer_martin: Paper you have to read and share - imagine what is possible in life a science - Unsupervised word embeddings capture lat…
Oct 25, 2019 · 2 shares · 15,842 followersView post →
MH
Martin Hetzer
@hetzer_martin · X
Paper you have to read and share - imagine what is possible in life a science - Unsupervised word embeddings capture latent knowledge from materials science literature | Nature https://t.co/0sVJyR03fL
Oct 25, 2019 · 15 likes · 2 shares · 447 followersView post →
PM
Planet MatSci
@planetmatsci · X
Unsupervised word embeddings capture latent knowledge from materials science literature, V. Tshitoyan, J. Dagdelen, L. Weston, A. Dunn, Z. Rong, O. Kononova, K. A. Persson, G. Ceder, and A. Jain, Nature 571, 95 (2019). https://t.co/WkIqd4rE6p
Check out episode 291 with Anubhav Jain, Staff Scientist & Chemist at Lawrence Berkeley National Lab (@BerkeleyLab) to discuss his paper, ‘Unsupervised word embeddings capture latent knowledge from materials science literature’. https://t.co/K1qFW3qxPe #materialscience https://t.co/7sda3q4VHC
Aug 17, 2019 · 3 likes · 14,213 followersView post →
LS
Life Sciences SF
@LifeSciencesSF · X
Unsupervised word embeddings capture latent knowledge from materials science literature. https://t.co/uxe1ZvsDUo
Using knowledge from past publications to hypothesize and test new stuff — pretty much what we do as scientists. Now by a machine! https://t.co/xj7kaDdidv (might be early days but I bet a version of this will be an off-the-shelf utility in the future!)
Machine learning was fed many scientific papers, analyzed connections between terms and concepts, and found new scientific discoveries by itself through the analysis https://t.co/QDmduxXogI
Extended data and supplementary information: https://t.co/MwltQoMDw2.
Code: https://t.co/ZS0fcZI8ab.
From the article: "All other data generated and analysed during the current study are available from the corresponding authors on reasonable request."
AI For Science Report | PDF | Artificial Intelligence | Intelligence (AI ...
News
doi:10.1038/s41586- Features in Volumetric Data. Advances in 019-1335-8 ... Constraints Generative Adversarial 019-1335-8. Networks for Modeling Chaotic ...
Technological Forecasting & Social Change: A A B C D e A | PDF ...
News
https://doi.org/10.1038/ Zammuto, R.F., Griffith, T.L., Majchrzak, A., Dougherty, D.J., Faraj, S., 2007. Information s41586-019-1335-8. technology and the ...
Ntibio Ai Final | PDF | Artificial Intelligence | Intelligence (AI ...
News
https://doi.org/10.1038/s41586-023-06221-2. and Rise of A.I. Large Language ... s41586-019-1335-8. 23. Bo Chen et al., “xTrimoPGLM: Unified 100B-Scale ...
Research 4.0 Interim Report | PDF | Artificial Intelligence ...
News
https://doi.org/10.1038/ Automating methodological and statistical review. s41586-019-1335-8. Retrieved 23 September 2019, from Research in Turovsky, B ...
1 s2.0 S2238785424020192 Main | PDF | Support Vector Machine ...
News
https://doi.org/10.1038/s41586-019-1335-8. [179] Karniadakis GE, Kevrekidis IG, Lu L, Perdikaris P, Wang S, Yang L. Physics- informed machine learning. Nat ...
Discovering new functional materials for clean energy and beyond ...
News
https://doi.org/10.1038/s41586-019-1335-8 Sponsor: SPP, Toyota Research Institute 9. 10. Summary • We arecollaborating with other groups at LBNL (G. Ceder, H ...
L L M M K: U S M S GPT: Arge Anguage Odels As Aster EY Nlocking ...
News
http://dx.doi.org/10.1038/s41586-019-1335-8. [3] Shu Huang and Jacqueline M. Cole. BatteryBERT: A Pretrained Language Model for Battery Database Enhance- ment.
The team, reporting its work in Nature 10.1038/s41586-019-1335-8, now plans to train a model on full texts of scientific articles, rather than just abstracts.
NLP Google Nature 2019-Word-Embeddings PDF | PDF | Teaching ...
News
thermoelectric predictions. V.T., J.D. and L.W. analysed the results and developed 10.1038/s41586-019-1335-8. the software infrastructure for the project. J.D. ...