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?
78 72 83 -
Predictions
Statements about events outside the person's control, checked once they fell due: did they come true?
64 43 83 -
Promises
Commitments to the person's or organisation's own future action, with something to deliver: was it delivered?
Insufficient record: 0 promises resolved, 5 needed for a score.
All types combined: 77 (likely 71 to 82, high confidence), from 84 claims: 79 factual claims, 5 predictions. It mixes kinds of claim in different proportions for different people, so compare people kind by kind.
Yann LeCun's claims
23 claims
“However in many problems labeled data is scarce whereas unlabeled data is abundant.”
Yann LeCun, co-author of paper 2010 Convolutional Networks and Applications in Vision (yann.lecun.com)
In many problems labeled data is scarce whereas unlabeled data is abundant.
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.
What would count as true or false
Judge as of 2010. True: the machine learning literature of that time (for example, surveys of semi-supervised and unsupervised learning) widely recognises that in many practical problems labelled examples are costly and scarce, while unlabelled data is plentiful. Mostly true: this holds broadly, but 'many problems' slightly overstates how common it was. Mixed: the situation applied in some important domains but not in roughly as many others, so it is only half right as a general statement. Mostly false: it was true only in a few niche cases and is presented as general. False: the evidence shows that labelled data was generally not scarce compared with unlabelled data in the problems the field dealt with.
Evidence
- Semi-supervised Learning Google Books (listing for Chapelle, Schölkopf, Zien)
“Interest in SSL has increased in recent years, particularly because of application domains in which unlabeled data are plentiful, such as images, text, and bioinformatics.”
- IR-Lab Project of Yanjun Qi (Fall 2004) Carnegie Mellon University (Yanjun Qi, IR-Lab project page) · 3 February 2004
“Large amounts of unlabeled data are available in many real-life data-mining tasks.”
- IR-Lab Project of Yanjun Qi (Fall 2004) Carnegie Mellon University (Yanjun Qi, IR-Lab project page) · 3 February 2004
“unlabeled data does not help classification when the unlabeled data distribution does not relate to the classification task”
- [PDF] Semisupervised learning - learning from labeled and unlabeled data Indian Statistical Institute, Bangalore (Saroj K. Meher slides)
“Labeled data are expensive and limited that motivates to use”
Reasoning
E2 is a 2004 literature review of semi-supervised learning. It states that large amounts of unlabeled data are available in many real-life data-mining tasks. E1, the Google Books description of the semi-supervised learning book by Chapelle, Schölkopf and Zien, ties growing interest in the field to application domains where unlabeled data are plentiful, such as images, text and bioinformatics. E4 says labeled data are expensive and limited, though it is undated and so carries less weight. Taken together, these independent sources support the claim that in many problems labeled data is expensive and limited while unlabeled data is plentiful [E1][E2][E4], though only E2 is dated before 2010. E3's caveat is about whether unlabeled data helps, not how scarce labeled data is.
Cite this claim
Yann LeCun, “In many problems labeled data is scarce whereas unlabeled data is abundant.” (said 2010). Verdict: resolved true. Reliability API, claim c_ef76e8af4793, method 0.1.0+c876030c. https://epistacklabs.com/claims/c_ef76e8af4793
Claims without a verdict
- Unresolvable 110
- Not yet due 3
- Not researched 309
- Not checkable 669
Is a verdict here wrong, or do you speak for Yann LeCun? 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.