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
“In computer-vision applications, the trend is to apply ConvNets to an entire image so as to detect, segment, and recognize objects of any size at any location”
Yann LeCun, paper author 2019 ISSCC 2019 Plenary: Deep Learning Hardware: Past, Present and Future (yann.lecun.com)
In computer vision, the trend is to apply ConvNets to an entire image to detect, segment and recognize objects of any size at any location.
Resolved true The record shows it happened or was so.
What would count as true or false
Judge as of 2019 against computer vision literature of about 2015–2019. True: dominant detection and segmentation methods then in use applied ConvNets to the whole image (fully convolutional or single-pass, e.g. FCN, Faster R-CNN, YOLO, SSD, Mask R-CNN, feature pyramids) and handled objects of varied size and location. Mostly true: this was a major trend, but crop- or region-based pipelines were still about equally common. Mixed: whole-image application was only one of several competing approaches, with no clear trend. Mostly false: it was a niche approach then. False: the field was moving away from whole-image ConvNets.
Evidence
- Deep Learning for Generic Object Detection: A Survey - International Journal of Computer Vision Springer (International Journal of Computer Vision) · 31 October 2019
“RPN is, in fact, a kind of Fully Convolutional Network (FCN) (Long et al. 2015; Shelhamer et al. 2017); Faster RCNN is thus a purely CNN based framework without using handcrafted features.”
- Deep Learning for Generic Object Detection: A Survey - International Journal of Computer Vision Springer (International Journal of Computer Vision) · 31 October 2019
“This approach has been found to be effective for segmentation (Long et al. 2015; Shelhamer et al. 2017) and human pose estimation (Newell et al. 2016), has been widely exploited by both one-stage and two-stage detectors to alleviate problems of scale variation across object instances.”
- Recent Advances in Object Detection in the Age of Deep Convolutional Neural Networks arXiv (rendered via ar5iv) · 1 June 2018
“All pipelines are now fully convolutional, which brings structure (regularization), simplicity, speed and elegance to the detectors.”
- Image Segmentation Using Deep Learning: A Survey arXiv · 31 January 2026
“This work is considered a milestone in image segmentation, demonstrating that deep networks can be trained for semantic segmentation in an end-to-end manner on variable-sized images.”
Reasoning
A 2018 review of object detection states that "all pipelines are now fully convolutional" [E3]. A 2019 IJCV survey describes Faster R-CNN's region proposal network as a fully convolutional network, which makes Faster R-CNN a purely CNN-based framework [E1]. The same survey says an approach found effective for segmentation and pose estimation has been widely used by both one-stage and two-stage detectors to deal with scale variation across objects [E2]. A segmentation survey calls a work a milestone because it showed that deep networks could be trained end-to-end for semantic segmentation on variable-sized images [E4]. Together these sources show that fully convolutional, whole-image designs had become the standard in object detection in that period [E1][E3], and that end-to-end segmentation on variable-sized images was regarded as a milestone [E4], which meets the criteria for true.
Cite this claim
Yann LeCun, “In computer vision, the trend is to apply ConvNets to an entire image to detect, segment and recognize objects of any size at any location.” (said 2019). Verdict: resolved true. Reliability API, claim c_0364a61afaa9, method 0.1.0+c876030c. https://epistacklabs.com/claims/c_0364a61afaa9
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.