代写Machine learning代做R编程

Machine learning

General guidelines:

●   This assignment is tentatively due on Tuesday, 5pm, week 8 (May 20); as always, the final deadline is as posted on gradescope

●    Note the slightly unusual deadline: I want to avoid a clash with the HW4 deadline, and I want the leaderboard to close while people are awake

   Also note that there is no late period for this assignment: please submit on time!

   This assignment is worth 20% of the final grade (noting grade scaling as posted on the course webpage)

   Assignment 1 must be completed individually

Data and baselines

Assignment 1 data and baselines:

https://drive.google.com/drive/folders/1U_Wb7Q1JhrljRiG9dQ6oyCSPvPCbynm9?usp=sharing

(data is also posted here in case drive gets rate-limited:

https://cseweb.ucsd.edu/classes/sp25/cse253-a/data/)

There are two files in the above folder:

student_files.tar.gz: A compressed folder containing:

   Training datasets for each of the three tasks below, with labels

●   Testing datasets for each of the three tasks, without labels baselines.ipynb: Working baselines for each of the three tasks

Note that testing labels are not provided: rather you have to make your own predictions on the test set and upload them to gradescope.

Also note that there is no stub provided: it’s easiest to just edit baselines.ipynb directly to form the basis of your solution. You are not required to follow any of the functions outlined in that file if you don’t want to.

Tasks

You are required to build classifiers for each of the following three tasks:

Task 1: Composer classification (symbolic, multiclass). For this task you are required to process a midi file and predict which composer wrote the piece of music. This task is evaluated based on accuracy (percentage of correct predictions).

Task 2: Next sequence prediction (symbolic, binary). For this task you are given two bars of music (really, two midi files) and you are required to guess whether the second bar would follow the first in a real piece of music (or whether they are unrelated). I.e., predict “True” if the two bars are neighbors and “False” otherwise. This task is evaluated based on accuracy.

Task 3: Music tagging (continuous, multilabel, multiclass). Predict a set of tags (e.g. “electronic”, “chill”) associated with an audio file. The input audio files are outputs of a generative model that synthesized music based on the given tags. Note that for this task an example can have multiple tags. This task is evaluated based on the mean average precision (mAP).

Baselines for the three tasks are fairly straightforward, mostly just to show you how to process the data and generate valid outputs that can be submitted to gradescope:

Task 1: Logistic regression; features are computed based on the average pitch and duration.

Task 2: No ML is used: this baseline just measures whether the two bars have a similar average pitch.

Task 3: CNN based on MelSpectrograms; this baseline also shows you how to build your own validation set using a sample of the training data.

Submission

Running the baseline code will produce a valid submission that you can submit to gradescope. It will generate three outputs:

predictions1.json, predictions2.json, predictions3.json

These contain the predictions for the three tasks on the test set. Although you are not given the test labels, the evaluation code in the baselines file shows you how your predictions will be compared to the test labels by the autograder.

The autograder will evaluate your test performance and print your results on a leaderboard.  Note that there is both a “public” and “private” leaderboard, each of which contains a random half of the test set. The latter will only be visible after the deadline to prevent people from overfitting to the public half.

You are also required to submit two additional files:

assignment1.py: the code you used to train your model. The autograder does not currently run your code, though this should be runnable in the event that there is any doubt about your solution.

writeup.txt: 1-2 sentences describing your solution to each task. I just want this so that I can describe solutions after the assignment finishes.

Grading

The assignment is worth 20 marks. The grade breakdown is as follows:

●   Submit valid writeup.txt and assignment1.py files (2 marks). The autograder doesn’t currently check these (and will just give out the marks) but could be modified e.g. if people are submitting empty files.

●    Each of the three tasks is worth 6 marks (3 for the public leaderboard and 3 for the private)

The breakdown of the 3 marks per leaderboard & task is TBD but planned guide is as follows (same for all three tasks):

   0.5 marks: Submit a valid solution (just run the baseline code and submit it!)

●    1 mark: Submit a solution that is at least 5% better than the baseline (i.e., >= baseline performance * 1.05)

●    1.5-2.5: Submit a solution that is “significantly” better than the baseline. Tentative thresholds are below (may adjust down or add smaller increments, won’t adjust up)

   Task 1: 0.4 (1.5); 0.5 (2.0); 0.6 (2.25); 0.7 (2.5)

   Task 2: 0.7 (1.5); 0.8 (2.0); 0.85 (2.25); 0.9 (2.5)

   Task 3: 0.3 (1.5); 0.35 (2.0); 0.4 (2.5)

●   Submit a solution that is among the top 10% of submissions: 3 marks (possibly computed “smoothly” rather than given to the top 10% in a binary fashion)

(the above three marks are computed for both the public and private leaderboards, and for each of the three tasks, for a total of 18 marks)

Note: baseline performance, task 1/2/3 (public): 0.251256 / 0.623775 / 0.270429

Getting started

We extracted features from midi files in Homework 1 and built CNN classifiers for audio in Homework 2: you can use those (or the solution files, which I’ll post) as the basis of your  solutions.

Also please watch the Thursday week 4 lecture for discussion of this assignment and some tips about getting started!





热门主题

课程名

mktg2509 csci 2600 38170 lng302 csse3010 phas3226 77938 arch1162 engn4536/engn6536 acx5903 comp151101 phl245 cse12 comp9312 stat3016/6016 phas0038 comp2140 6qqmb312 xjco3011 rest0005 ematm0051 5qqmn219 lubs5062m eee8155 cege0100 eap033 artd1109 mat246 etc3430 ecmm462 mis102 inft6800 ddes9903 comp6521 comp9517 comp3331/9331 comp4337 comp6008 comp9414 bu.231.790.81 man00150m csb352h math1041 eengm4100 isys1002 08 6057cem mktg3504 mthm036 mtrx1701 mth3241 eeee3086 cmp-7038b cmp-7000a ints4010 econ2151 infs5710 fins5516 fin3309 fins5510 gsoe9340 math2007 math2036 soee5010 mark3088 infs3605 elec9714 comp2271 ma214 comp2211 infs3604 600426 sit254 acct3091 bbt405 msin0116 com107/com113 mark5826 sit120 comp9021 eco2101 eeen40700 cs253 ece3114 ecmm447 chns3000 math377 itd102 comp9444 comp(2041|9044) econ0060 econ7230 mgt001371 ecs-323 cs6250 mgdi60012 mdia2012 comm221001 comm5000 ma1008 engl642 econ241 com333 math367 mis201 nbs-7041x meek16104 econ2003 comm1190 mbas902 comp-1027 dpst1091 comp7315 eppd1033 m06 ee3025 msci231 bb113/bbs1063 fc709 comp3425 comp9417 econ42915 cb9101 math1102e chme0017 fc307 mkt60104 5522usst litr1-uc6201.200 ee1102 cosc2803 math39512 omp9727 int2067/int5051 bsb151 mgt253 fc021 babs2202 mis2002s phya21 18-213 cege0012 mdia1002 math38032 mech5125 07 cisc102 mgx3110 cs240 11175 fin3020s eco3420 ictten622 comp9727 cpt111 de114102d mgm320h5s bafi1019 math21112 efim20036 mn-3503 fins5568 110.807 bcpm000028 info6030 bma0092 bcpm0054 math20212 ce335 cs365 cenv6141 ftec5580 math2010 ec3450 comm1170 ecmt1010 csci-ua.0480-003 econ12-200 ib3960 ectb60h3f cs247—assignment tk3163 ics3u ib3j80 comp20008 comp9334 eppd1063 acct2343 cct109 isys1055/3412 math350-real math2014 eec180 stat141b econ2101 msinm014/msing014/msing014b fit2004 comp643 bu1002 cm2030
联系我们
EMail: 99515681@qq.com
QQ: 99515681
留学生作业帮-留学生的知心伴侣!
工作时间:08:00-21:00
python代写
微信客服:codinghelp
站长地图