Checklist NOAI China 2024 Round 2 (China Stage) · Task 1
Predicting the Shooting Percentage of Basketball Stars
English title: 球星投篮命中率预测任务
Train a small constrained PyTorch MLP that predicts whether a basketball star's shot is made from the shooting position.
The task
The data record a basketball star's shots with normalised features loc_x, loc_y (court position), minutes_remaining and shot_distance, the label shot_made_flag (1 = made) and an identifier shot_id. The training set has 20,000 rows and the hidden test set about 5,000.
The contestant must implement a multi-layer perceptron in PyTorch whose input is only the two features loc_x and loc_y and whose output is shot_made_flag. Only linear layers followed by activation functions may be used, with at most 3 linear layers and at most 8 neurons per linear layer; the network must not be wrapped in nn.Sequential, because the grader cannot inspect it. Activation functions must be chosen from the list in the statement (nn.ReLU, nn.Sigmoid, nn.Tanh, nn.ELU, nn.LeakyReLU, nn.PreLU [sic]); loss, optimiser and learning rate are free.
Abridged by SOTA from the official materials. The official statement has the exact rules, and it wins wherever this summary differs.
At a glance
- You get
data_train.csv(20,000 rows) from the Bohrium datasets tab; the test set (about 5,000 rows) is hidden.- You submit
submission.zipcontainingsubmission_model.py(model class and imports) andsubmission_dic.pth(trained parameters), as in the baseline.- Scoring
- Test-set accuracy if the architecture constraints are met; otherwise 0. Leaderboard B is final.
- Rules
- Inputs limited to
loc_xandloc_y. - At most 3 linear layers, at most 8 neurons each; no nn.Sequential.
- Activation functions only from the permitted list.
- Inputs limited to
- Format
- NOAI 2024 Round 2 (China Stage), a one-day practical round in Beijing on 10 June 2024 (date and place from a third-party overview page). Republished on Bohrium as the 'NOAI2025 teaching test (NOAI2024 real problems)' and, in English, as the 'APOAI2025 Mock Competition' (10 Nov 2024 - 31 May 2025). Leaderboard A uses 50% of the test set during the contest; leaderboard B (the remaining 50%) is final. Question 1.