Atlas record: https://www.aitimeline.dev/en/events/evt-0768/
Research · November 17, 2014
Show and Tell: a caption for an image
On 17 November 2014 Oriol Vinyals and colleagues at Google described a network that composes a sentence about an image: a convolutional network encodes the picture and an LSTM writes the description. On Pascal the BLEU score was 59 against a previous best of 25 and human performance around 69.
Why it matters
Describing an image became a single model trained end to end on image-sentence pairs, just as translation had become sentence to sentence. The line of models joining vision and language starts here.
By the first version of the preprint, BLEU on Flickr30k rose from 55 to 66 and on SBU from 19 to 27. The MSCOCO training split is 82,783 images. The record does not claim the 27.7 on COCO: it is not in the first version.
Event record
- Event date
- November 17, 2014
- Timeline date
- Event date
- Verification
- Sources gathered automatically · September 25, 2026
- Lines
- ID
- evt-0768
The day the first version of the preprint was submitted; the figures were read in it.
Related events
Earlier
- Builds on Sequence-to-sequence 2014
The model is carried over from encoder-decoder machine translation (Sutskever, Vinyals, Le): an image is encoded in place of a source sentence.
- Enables MS COCO: objects in context 2014links to this record
Show and Tell trains and tests on MSCOCO among five datasets.
- Related Words and pictures: a joint model 2003links to this record
Thematic: Show and Tell in 2014 also writes words for an image, but cites other predecessors; no line read connects them.
Later