The founding 1998 DjVu document-compression paper was co-authored by Yann LeCun and Yoshua Bengio — two future 2018 Turing Award recipients — alongside Léon Bottou at AT&T Labs

The founding paper of DjVu — "High Quality Document Image Compression with DjVu," Journal of Electronic Imaging, 7(3):410–425, 1998 — lists six authors: Léon Bottou, Patrick Haffner, Paul G. Howard, Patrice Simard, Yoshua Bengio, and Yann Le Cun. DjVu is a compression format for scanned documents, developed at AT&T Labs where Bottou worked from 1996. Two of the six co-authors, LeCun and Bengio, would go on to share the 2018 ACM A.M. Turing Award with Geoffrey Hinton for deep learning — an award for neural networks, not for the document-scanning side project the three of these researchers were engineering together years earlier.

The claim's value is as a bridge. It connects two otherwise-separate vault clusters: the Bottou/Perceptrons thread (claim-bottou-2017-foreword-perceptrons-chapters-0-13-are-1969) and the LeCun handwriting-recognition thread (claim-lecun-1989-first-practical-recognition, claim-lecun-1989-two-papers-publication-order) — through a shared history in image compression rather than in neural nets. See moc-backpropagation-origins for the LeCun/Bottou cluster this extends.

Mechanically, DjVu is described (Wikipedia, unverified against the primary paper) as separating a page into background, foreground, and text-mask layers, compressing the continuous-tone layers with an IW44 wavelet codec and the bitonal text mask with JB2 shape-matching. That mechanism is a lead to the 1998 paper itself, not a claim resting on this note's source; it is recorded here only as context.

The roster is a bibliographic fact, but it currently rests on Bottou's own publication page alone — an authoritative but single source. An independent catalog cross-check is routed at question-verify-djvu-1998-paper-roster.