Canonical text (the exact bytes fingerprinted)

TRUVACE RECORD VERSION
record: TRV-2026-0209
version: 4
kind: sources_changed
reason: Source set updated
timestamp: 2026-09-04T06:06:53.848401Z
status: published
lens: g_space
sector: science
headline: Deep learning
dek: Deep learning allows computational models that are composed of multiple processing layers to learn representations of data with multiple levels of abstraction. These methods have dramatically improved the state-of-the-art in speech recognition, visual object recognition, object detection and many other domains such as drug discovery and genomics. Deep learning discovers intricate structure in large data sets by using the backpropagation algorithm to indicate how a machine should change its internal parameters th…
gain_title: Deep learning models composed of multiple processing layers improved state-of-the-art performance in speech recognition, visual object recognition, object detection, drug discovery and genomics by 2015
problem_title: (none)
trace_subject: (none)
gain_reading: Deep learning models composed of multiple processing layers improved state-of-the-art performance in speech recognition, visual object recognition, object detection, drug discovery and genomics by 2015
gain_evidence: dramatically improved the state-of-the-art in speech recognition, visual object recognition, object detection and many other domains such as drug discovery and genomics | Deep convolutional nets have brought about breakthroughs in processing images, video, speech and audio
problem_reading: (none)
problem_evidence: (none)
quick_read: Published in May 2015, this peer-reviewed overview describes deep learning as models with multiple processing layers that learn multi-level representations of data, trained via backpropagation to adjust parameters between layers

The significance by that date was reported as dramatic state-of-the-art gains in speech recognition, visual object recognition and detection, with extensions to drug discovery and genomics, but the text provides no specific deployment, population, measured clinical or societal outcome, or quantified risk assessment
limitation: 
tag: Evidence-backed gain
key_points: Article defines deep learning as computational models composed of multiple processing layers that learn representations with multiple levels of abstraction | Method uses backpropagation algorithm to indicate how a machine should change its internal parameters that are used to compute the representation in each layer from the representation in the previous layer | Recurrent nets are described as effective for sequential data such as text and speech, while convolutional nets are linked to images, video, speech and audio
rundown: The source describes the mechanism as discovering intricate structure in large data sets by using backpropagation to adjust internal parameters layer by layer

It distinguishes two architectural families, noting convolutional nets for images, video, speech and audio and recurrent nets for sequential data such as text and speech
sources:
- peer_reviewed | Briefings in Bioinformatics | https://doi.org/10.1093/bib/bbx044 | 2017-04-05
- peer_reviewed | Frontiers in Artificial Intelligence | https://doi.org/10.3389/frai.2020.00004 | 2020-02-28
- peer_reviewed | Genomics, Proteomics & Bioinformatics | https://doi.org/10.1016/j.gpb.2017.07.003 | 2018-02-01
- peer_reviewed | Nature | https://doi.org/10.1038/nature14539 | 2015-05-01
prev: 994fd315db9cff935848820b2a0bdde4a2f7301b43282385dffe52783d227712

sha256

1af50eb11040abf2ac8030f33b0e36574b0d84d33f2da20f7e3289909f43eb17

previous

994fd315db9cff935848820b2a0bdde4a2f7301b43282385dffe52783d227712