ML Competition Winner Questionnaire

JavaScript isn't enabled in your browser, so this file can't be opened. Enable and reload.

Congratulations for winning a competition!

We really appreciate you taking the time to fill in this form.

Your answers will be used as part of a study into winning methods for data science competitions across major platforms including Kaggle, DrivenData, AIcrowd, CodaLab and others, to be published on mlcontests.com. Your name will be listed in the acknowledgements section of the final public report, if you would like it to be.

The key questions on this form should take you only a few minutes. There are 14 questions on the first page, mostly multiple choice, which will should take you 2-5 minutes to complete. The second page has 14 more optional questions - fill these in if you'd like to, but feel free to skip if you don't have time!

To compensate you for your time we have partnered with Genesis Cloud to provide you with $100 in free credits, to use on their cloud GPUs powered by 100% sustainable energy.

The email address you provide at the top of the form won't be shared. There is a question in the form specifically about which contact details you want us to share with Genesis Cloud to arrange these credits.

Email *

Which competition did you win? (put a link to the competition page, or the name of the competition) *

How many people were in your team for this competition? *

Other:

Have you or anyone in your team won a data science competition before? *

No, this was my first

Yes, one

Yes, multiple

Would you share the code for your solution with us? Please include a link to your source code if public, or let us know if you would prefer to share privately and we will contact you by email.

What was the main programming language used in your solution? *

Python

Matlab

Julia

C++

Rust

Other:

Any additional languages used?

Python

Matlab

Julia

C++

Rust

Other:

Which package did you use for your main model? *

scikit-learn

TensorFlow

PyTorch

Fast.ai

Keras

XGBoost

LightGBM

Catboost

Other:

What kind of computing hardware was primarily used for training your solution? *

Personal GPU

Personal computer (not using GPU)

Rented cloud GPU

Rented cloud TPU

Work/university compute cluster

Other:

Which other software packages did you use? (including modelling, data prep, etc)

scikit-learn

TensorFlow

PyTorch

Fast.ai

Keras

XGBoost

LightGBM

Catboost

NumPy

Pandas

Dask

tidyr

Other:

What is the primary modelling technique you used? *

Simple feed-forward deep neural net (MLP)

Gradient boosted trees

Transformer-based neural net

Convolutional neural net

Recurrent neural net

Support Vector Machines (SVMs)

Linear regression

Logistic regression

Random forest

Other:

Which part of your solution do you think was most important in beating other teams?

Feature design

Modelling approach

Data augmentation

Domain knowledge

Computing power

Other:

Clear selection

What kind of cross-validation method did you use?

K-fold cross-validation

Fixed evaluation set

Other:

Clear selection

Which of these best describes your final solution?

One main model

An ensemble of multiple slightly different versions of the same model (e.g. several neural nets with roughly the same architecture)

An ensemble of multiple different types of models (e.g. one neural net, one linear model, one decision tree)

Other:

Clear selection

We'd love to do a more in-depth interview with you, that we can publish in written form. Would you be ok with us contacting you for this?

Yes, through the email address submitted in this form

Other:

Clear selection

Please enter your name(s) as you would like them to be listed in the acknowledgements section of our public report (or leave blank if you would prefer to stay anonymous!)

Please list the email addresses you would like us to share with Genesis Cloud, to be used to assign your $100 free credits to. (one per team member)

Clear form

Never submit passwords through Google Forms.

This content is neither created nor endorsed by Google. Report Abuse - Terms of Service - Privacy Policy

Forms