We did something similar in a 2014 short paper using uncertainty sampling. We also built a tool using this approach. Paper: https://dl.acm.org/citation.cfm?id=2791022 … Tool:https://github.com/uiuc-ischool-scanr/SAIL …
Very happy & proud to share some research in Deep Bayesian Active Learning from @yaringal, @joost_v_amersf and me at @OATML_oxford 
http://oatml.cs.ox.ac.uk/blog/2019/06/24/batchbald.html …
-
-
-
Yeah, human-in-the-loop is useful for NLP! Indeed, Active Learning could help your tool identify which tweets from the additional data to present to humans, and thus increase data efficiency by a great deal. Thanks for sharing!
- Još 1 odgovor
Novi razgovor -
Čini se da učitavanje traje već neko vrijeme.
Twitter je možda preopterećen ili ima kratkotrajnih poteškoća u radu. Pokušajte ponovno ili potražite dodatne informacije u odjeljku Status Twittera.