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
55 claims
“the check reading system my colleague and I built at Bell Labs in the early 1990s used a form of structured prediction on top of convolutional nets that we called “Graph Transformer Networks”.”
Yann LeCun, Quora answer 28 July 2016 Yann LeCun Quora Session Overview (KDnuggets)
The check reading system LeCun and a colleague built at Bell Labs in the early 1990s used structured prediction on top of convolutional nets, called Graph Transformer Networks.
Mostly true True in substance, wrong in a detail.
The second judge said resolved true; the verdict shown is the first judge's.
What would count as true or false
Judge as of 2016-07-28. True: published sources (e.g. LeCun, Bottou, Bengio & Haffner 1998, 'Gradient-Based Learning Applied to Document Recognition', and related papers) show that the Bell Labs check-reading system developed in the early-to-mid 1990s combined convolutional nets with a structured-prediction layer called Graph Transformer Networks. Mostly true: as stated, except for a minor detail such as the dates (e.g. mid-1990s rather than early 1990s) or the number of collaborators. Mixed: the system used convnets, but GTNs were a separate or later development not used in it, or vice versa. Mostly false: the system existed but did not use structured prediction on convnets in the way described. False: no such Bell Labs check-reading system by LeCun existed.
Evidence
- Yann LeCun - A.M. Turing Award Laureate ACM A.M. Turing Award · 1 January 2018
“As well as developing the convolutional approach, LeCun pioneered its application in “graph transformer networks” to recognize printed and handwritten text. This was used in a widely deployed system to read numbers written on checks, produced in the early 1990s in collaboration with Bengio, Leon Bottou and Patrick Haffner.”
- leon.bottou.org Léon Bottou (personal site hosting paper: Bottou, Le Cun, Bengio, CVPR 1997) · 1997 Own record
“A complete check reading system based on these concept is described. The system uses convolutional neural network character recognizers, combined with global training techniques to provides record accuracy on business and personal checks.”
- Gradient-based Learning Applied to Document Recognition Proceedings of the IEEE · November 1998 Own record
“A graph transformer network for reading a bank cheque is also described. It uses convolutional neural network character recognizers combined with global training techniques to provide record accuracy on business and personal cheques.”
Reasoning
The ACM Turing Award biography (E1), an independent source, says LeCun pioneered the application of the convolutional approach in "graph transformer networks", which was used in a widely deployed check-reading system produced in the early 1990s. It dates that system to the early 1990s. The 1998 Proceedings of the IEEE paper (E3) describes a graph transformer network for reading bank cheques, and the 1997 CVPR paper (E2) describes a check-reading system using convolutional character recognizers with global training techniques. It uses convolutional character recognizers with global training, which is a structured-prediction layer on top of convnets. The core claim holds, but E1 credits the work to LeCun with three collaborators (Bengio, Bottou and Haffner), not one colleague, so the verdict is mostly true.
Cite this claim
Yann LeCun, “The check reading system LeCun and a colleague built at Bell Labs in the early 1990s used structured prediction on top of convolutional nets, called Graph Transformer Networks.” (said 28 July 2016). Verdict: mostly true. Reliability API, claim c_4f9a162fcb6d, method 0.1.0+c876030c. https://epistacklabs.com/claims/c_4f9a162fcb6d
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.