Page MenuHomePhabricator

Design how Jade judgments feed into ORES model training
Closed, DeclinedPublic

Description

Jade human judgments should have some kind of influence on the error function we use during ML training. How is that feedback structured? Is it a strong signal? Do we respect all judgment equally? Do we need another level of human meta-meta-curation of feedback??

Event Timeline

Harej renamed this task from Spike: Design how Meta ORES feedback plugs back into the ORES models to Spike: Design how JADE feedback plugs back into the ORES models.Jul 6 2018, 3:00 AM
Harej updated the task description. (Show Details)
awight renamed this task from Spike: Design how JADE feedback plugs back into the ORES models to Design how JADE judgments feed into ORES model training.Dec 19 2018, 10:57 PM
awight updated the task description. (Show Details)
Harej renamed this task from Design how JADE judgments feed into ORES model training to Design how Jade judgments feed into ORES model training.Apr 15 2019, 10:16 PM
Aklapper lowered the priority of this task from Low to Lowest.May 23 2021, 10:54 PM