Im stillwaiting on thinhs as far back as august, though the august one is one that gets rejected by people but accepted on appeal lol. The latestni think from when ml got turned off is about september/october time, when there was still sunlight at 6am lol
1 Like
Im very curious to see if what people think ML auto sccepts, and what ML will auto accept moving forward match.
2 Likes
Im hoping trail markers still get auto accepted, cause quite often they get rejected by people unless theres a map
1 Like
I am a retired form a state highway Dept as a highway engineer for 33.3 years.
I worked with little fake points representing the digital land in the 1990’s. I surveyed and the designed 3d roads and surfaces for measurement and payment to federal standards… And built roads, bridges, trails etc… And then I went over into CAD IT support for my state hwy Dept. So I like what Niantic is attempting to do. I also made utilities move their stuff off my highway sections because IT was boring and teaching hwy engineers was annoying.
The reason I like AI or machine learning is that a person with knowledge can ask the right questions and get some interesting results or Tangents.
My last duty was to investigate a Superfund site for public information and to answer technical questions for scary ass’s DOJ lawyers.
But I didn’t ask the lawyers questions, I wanted to retire on my terms.
Fortunately Google Gemini can process old files and make me never regret, retiring from those hwy idiots.
I am not convinced all of the changes are for the better. For example, I had this nomination
which ML allowed to be determined by the community.
When I resubmitted it, following ML’s reboot,
ML rejected it.
Given that parks / outdoor recreation areas without signs are supposed to be eligible:
my conclusion is that ML denying the community the opportunity to vote on the nomination is a backward step.
1 Like
The thing with ML is it isn’t fed the criteria and told to review to it, it’s fed how people review and tries to interpret that. While you changed the description, you used the same title & photo, so I think in this case it actually proves that ML is learning because now it’s rejecting based on how the community handled almost the same nomination.
I think you’ll have to get a new photo and change more text if you want to get past ML, or else rely on appeals after ML rejects it.
I don’t say that to knock your nomination or the particular candidate and I hope you can eventually have success with it.
3 Likes
I spent forever getting pocket parks accepted. And I only got them accepted with explicit proof to council maps, LGA documentation to prove their existence.
Not surprised Emily struggles as the machine is learning from us. And I can say that humans had epic fails on parks - as too many people believed a “park” equals a certain size with x facilities - so rejected - not a local space used by community. And with no objects to hang the nomination - it is a hard ask to get nominated.
Emily learns from us. So I put the blame there to start with
@Liverlouike15 evidence evidence evidence. Because your local council does not list it. Even though it exists and is maintained. I would find every bit of evidence and stuff your supporting info. It is one part of the walking trail from 12 Acre Drive to Peachcroft Park. So i like it. I get its local community importance
1 Like
Your dear daughter Emily has again rejected the “correct title” with an automatic rejection.
She submitted an edit report on 12/20/2024 and was rejected by “our team” on 12/23/2024.
I also submitted additional photos to show that the wayspot had disappeared, and she continued to approve them one by one.
Incidentally, the Wayfarer team has since seen my deletion report and pointed to the photo that Emily approved, and they continue to approve deletion reports.
So the problem is not solved.
It should be stopped immediately.
1 Like
ML takes around 24 hours to make a decision. I believe this was decided by Niantic human staff; it most likely was listed as Under Review in Contribution Management, with the Niantic logo next to it. If it’s in this status for a few days, it was pulled from the regular queue to be reviewed by staff, not ML. “Our team” can either be ML or human staff, but the time frame for resolution is quicker with ML and slower with staff.
You also shouldn’t be submitting new photos at a Wayspot that no longer exists as proof it should be removed. ML is learning from how we as a community review, so if there’s a similar photo already approved at that Wayspot, ML will typically approve it.
If, as you say, it is not machine learning, but rather the judgment of the team staff, then it is even worse.
They are not equipped with the academic knowledge to judge our data.
I’ve had plenty of ML accepts well within 24 hours. Shortest I’ve seen was 3 hours or so.
They are thinking that ML took 3 days to decide, as they submitted on December 20th, but the decision wasn’t made until December 23rd, and the email said “our team.” This sounds more like staff review than ML, as yes, it usually takes ML no more than 24 hours to make a decision.
“Our team” emails can either be for a ML or human staff decision; they do not say which made the decision in the emails. The way to know though is typically by the timeframe of the decision. 24 hours or less, it’s usually ML. A few days, it’s human staff.