Scores by kind of claim
Each kind of claim is scored on its own, out of 100: the share that held up, with the range it likely falls in. Fewer resolved claims give a wider range.
-
Factual claims
Statements about the present or the past, checked against the record: did they hold up?
81 76 87 -
Predictions
Statements about events outside the person's control, checked once they fell due: did they come true?
Insufficient record: 2 predictions resolved, 5 needed for a score.
-
Promises
Commitments to the person's or organisation's own future action, with something to deliver: was it delivered?
65 45 83
All types combined: 80 (likely 75 to 85, high confidence), from 84 claims: 78 factual claims, 6 promises. It mixes kinds of claim in different proportions for different people, so compare people kind by kind.
Demis Hassabis's claims
37 claims
“Excited that our #AlphaFold work, which uses deep neural networks to predict protein structures, is published in @Nature today!”
Demis Hassabis, DeepMind CEO, on X 15 January 2020 Post by @demishassabis on X, 2020-01-15 (X)
DeepMind's AlphaFold work, using deep neural networks to predict protein structures, was published in Nature on 2020-01-15.
Resolved true The record shows it happened or was so.
The second judge would not resolve this claim; the verdict shown is the first judge's.
Said for an organisation they lead or work for; it counts on both records.
What would count as true or false
The claim: a DeepMind AlphaFold paper using deep neural networks to predict protein structure was published in Nature on 2020-01-15. Online publication counts as publication. TRUE: Nature's record shows a DeepMind AlphaFold paper of this kind (for example, Senior et al., on structure prediction with deep-learning potentials) published online, or in an issue dated, 15 January 2020. MOSTLY TRUE: the paper appeared in Nature and the date is off by only a day or two (for example, the issue date differs slightly from the online date). MOSTLY FALSE: the AlphaFold work appeared in Nature, but the publication date is more than a few days off, or the paper does not use deep neural networks for structure prediction. FALSE: no such DeepMind AlphaFold paper appeared in Nature around then, or it appeared in a different journal.
Evidence
- Improved protein structure prediction using potentials from deep learning - Nature Nature (Springer Nature) · 15 January 2020
“Here we show that we can train a neural network to make accurate predictions of the distances between pairs of residues, which convey more information about the structure than contact predictions.”
- Improved protein structure prediction using potentials from deep learning - Nature Nature (Springer Nature) · 15 January 2020
“Senior, A.W., Evans, R., Jumper, J. et al. Improved protein structure prediction using potentials from deep learning. Nature 577, 706–710 (2020). https://doi.org/10.1038/s41586-019-1923-7”
- AlphaFold: Using AI for scientific discovery DeepMind · 15 January 2020 Own record
“Both of these methods relied on deep neural networks that are trained to predict properties of the protein from its genetic sequence.”
- Artificial intelligence used to predict 3D structure of proteins UCL School of Life and Medical Sciences · 16 January 2020
“The new study, published in Nature, found that a deep learning system called AlphaFold could model protein structure from scratch”
Reasoning
Nature's own article page (E1, E2), dated 2020-01-15, is the paper by Senior, Evans, Jumper et al., "Improved protein structure prediction using potentials from deep learning", Nature 577, 706–710 (2020). It describes training a neural network to predict distances between residue pairs in order to predict protein structure. UCL's write-up dated 2020-01-16 (E4) says a deep learning system called AlphaFold was the subject of a "new study, published in Nature", and DeepMind's blog post dated 2020-01-15 (E3) says its methods relied on deep neural networks. The 15 January date comes from the Nature page's own date rather than a labelled date in the quoted text. The UCL item dated the next day (E4) is consistent with that date, so the claim resolves true.
Cite this claim
Demis Hassabis, “DeepMind's AlphaFold work, using deep neural networks to predict protein structures, was published in Nature on 2020-01-15.” (said 15 January 2020). Verdict: resolved true. Reliability API, claim c_85a384b9ced0, method 0.1.0+c876030c. https://epistacklabs.com/claims/c_85a384b9ced0
Claims without a verdict
Is a verdict here wrong, or do you speak for Demis Hassabis? Reply on the claim's own page, with the form at its foot (how replies work). Verdicts are made automatically; how verdicts and scores are made.