Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Feature: Write-up stats phase 2 #137

Merged
merged 6 commits into from Jan 11, 2019
Merged

Feature: Write-up stats phase 2 #137

merged 6 commits into from Jan 11, 2019

Conversation

tobyspark
Copy link
Owner

Now with

  • analysis of folkrnn.org usage over time
  • analysis of themachinefolksession.org usage

@tobyspark
Copy link
Owner Author

Current write-up output –

24562 tunes were generated from 13410 sessions over 235 days.
Distribution: [(0, 7708), (1, 1929), (2, 1153), (3, 733), (4, 444), (5, 303), (6, 231), (7, 168), (8, 125), (9, 98), (10, 82), (11, 58), (12, 56), (13, 44), (14, 36), (15, 32), (16, 22), (17, 14), (18, 11), (19, 12), (20, 10), (21, 11), (22, 10), (23, 7), (24, 7), (25, 7), (26, 11), (27, 5), (28, 4), (29, 3), (30, 6), (31, 8), (32, 2), (33, 5), (34, 2), (35, 2), (36, 4), (37, 3), (38, 4), (39, 3), (41, 2), (42, 1), (43, 3), (44, 3), (46, 2), (47, 1), (49, 1), (50, 2), (51, 1), (54, 1), (55, 1), (58, 1), (60, 1), (61, 1), (63, 1), (64, 1), (66, 2), (75, 1), (86, 1), (87, 1), (90, 2), (103, 1), (105, 1), (108, 1), (119, 1), (124, 1), (141, 1), (182, 1)]
With mean: 1.8 and standard deviation: 5.4

A session is more-or-less the use of a unique browser, tracked over time. However the number of sessions with no tunes generated may have little value, as some sessions reported as distinct were identifiably not, and not all visits to the site may have been made in good faith (e.g. bots). Our best approximation of good-faith users are then joining sessions that are identifiably continuations of previous ones, and then discounting any that did not generate a tune.
Our best approximation of good-faith users generated on average mean: 4.3 standard deviation: 7.5.
For each tune generated, the frequency of the following generate parameters being used was:
model: 0.058, temp: 0.26, seed: 0.17, key: 0.19, meter: 0.16, start_abc: 0.18
For each tune generated, the frequency of it being played, downloaded or archived was:
play: 0.78, download: 0.081, archive: 0.016
If the tune had no changes to the generate parameters, the frequency of it being played, downloaded or archived was:
play: 0.7, download: 0.11, archive: 0.014, all: 0.78
Whereas if the tune did have changes to the generate parameters, the frequency of it being played, downloaded or archived was:
play: 0.83, download: 0.065, archive: 0.018, all: 0.87

Usage over time –
- archive_download_compose_histogram__daily.pdf and cropped version saved
- archive_download_compose_histogram__weekly.pdf and cropped version saved
Usage data over time shows three features. Overall use for the first 18 weeks was similar, with a median of 154.5 tunes generated each week. In the subsequent 15 weeks to the time of writing, overall use increased, with a median of 667.0 tunes generated each week. This period also features usage spikes. One week, correlating to an interview in Swedish media, shows 2.7x the median tunes generated. The largest, correlating to a mention in German media, shows an 18.4x increase.
The Machine Folk Session – 
As of 2019-01-09T05:42:25.275000, the website themachinefolksession.org has 551 contributed tunes. Of these, 82 have had further iterations contributed in the form of ‘settings’; the site currently hosts 92 settings in total. 69 tunes have live recordings contributed; the site currently hosts 64 recordings in total (a single performance may encompass many tunes).
Of the 551 contributed tunes, 406 were generated on, and archived from, folkrnn.org. Of these entirely machine-generated tunes, 32 have had human edits contributed; themachinefolksession.org currently hosts 37 settings of folkrnn generated tunes in total.
30 Registered users have selected 136 tunes as being noteworthy enough to add to their tunebooks. 46% of these are actually settings of the tune, rather than the original tune. Per the algorithm used by the home page of themachinefolksession.org to surface ‘interesting’ tunes, “Why are you and your 5,599,881 parameters so hard to understand?” is the most, with 4 settings and 5 recordings.
Most content-affecting activity has been from the administrators, however. Sturm accounts for 70% of such activity.

@tobyspark
Copy link
Owner Author

Write-up as my final contribution, incorporating some of Bob's text / framing.

During the first 235 days of activity at folkrnn.org, 24562 tunes were generated by – our heuristics suggest – 5700 users. Activity for the first 18 weeks averages a median of 155 tunes weekly. In the subsequent 15 weeks to the time of writing, overall use increased, with a median of 665 tunes generated weekly. This period also features usage spikes. One week, correlating to an interview in Swedish media, shows 2.7x the median tunes generated. The largest, correlating to a mention in German media, shows an 18.4x increase. These results show making our tool available to users of the web has translated into actual use, and that use is increasing. Further, media attention brings increased use, and this use is similarly engaged, judged by similar patterns of downloading MIDI and archiving tunes to themachinefolksession.org.

Of the fields available for users to influence the generation process on folkrnn.org, the temperature was used more often then the others (key, meter, initial ABC, and random seed). Perhaps this is because changing temperature results in more obviously dramatic changes in the generated material. Increasing the temperature from 1 to 2 will often yield tunes that do not sound traditional at all. If changes were made to the generate parameters, the frequency of the resulting tune being played, downloaded or archived increased from 0.78 to 0.87.

Over the same period since launch, themachinefolksession.org has seen tunes 551 contributed. Of these tunes, 82 have had further iterations contributed in the form of ‘settings’; the site currently hosts 92 settings in total. 69 tunes have live recordings contributed; the site currently hosts 64 recordings in total (a single performance may encompass many tunes). These results show around 100 concrete examples of machine-human co-creation have been documented.

Of the 551 contributed tunes, 406 were generated on, and archived from, folkrnn.org. Of these entirely machine-generated tunes, 32 have had human edits contributed; themachinefolksession.org currently hosts 37 settings of folkrnn generated tunes in total. These examples in particular attributable human iteration of, or inspiration by, machine produced scores.

Further value of machine produced scores can be seen by the 30 registered users who have selected 136 tunes or settings as being noteworthy enough to add to their tunebooks. Per the algorithm used by the home page of themachinefolksession.org to surface ‘interesting’ tunes, “Why are you and your 5,599,881 parameters so hard to understand?” is the most, with 4 settings and 5 recordings.

While these results are encouraging, most content-affecting activity on themachinefolksession.org has been from the administrators; co-author Sturm accounts for 70% of such activity. To motivate the use of the websites, we are experimenting with e.g. ‘tune of the month’, see above, and have organised a composition competition.

The composition competition was open to everyone but targeted primarily at music students. Submission included both a score for a set ensemble and an accompanying text describing how the composer used a folkrnn model in the composition of the piece. The judging panel - the first author was joined by Profs. Elaine Chew and Sageev Oore - considered the musical quality of the piece as well as the creative use of the model. The winning piece Gwyl Werin by Derri Lewis was performed by the New Music Players at a concert organised in partnership with the 2018 O’Reilly AI Conference in London. Lewis said he didn’t want to be ‘too picky’ about the tunes, rather he selected a tune to work from after only a few trails. He describes using the tune as a tone row and generating both harmonic, melodic and motivic material out of it. Though the tune itself, as generated by the system, does not appear directly in the piece.

@tobyspark tobyspark merged commit 127e2bc into master Jan 11, 2019
@tobyspark tobyspark deleted the feature-write-up-stats-2 branch January 11, 2019 10:53
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

None yet

1 participant