FREQUENTLY ASKED QUESTIONS (FAQ)
Below are some of the requently asked questions. We have provided answers to the questions which are categorized according to topic. We will continue to answer more question and you can also engange with us on our slack channel.
Registration
Q1: Can you provide some information on how to register for the challenges?
A1: Registration is via: https://www.itu.int/en/ITU-T/AI/challenge/2020/Pages/default.aspx (Please create guest account and register as a non-member if your institution/university is not avaiable as an ITU member).
Q2: How does one apply to a specific challenge? Right now the registration is general
A2: You could apply via the survey link or via directly to the regional hosts listed in our website. You are encourage to subscribe to the mailing list related to the problem statement chosen (for example, ai5gchallenge@listas.ufpa.br for the ML5G-PHY challenge in Brazil).
Q3: Sharepoint files are not accessible to all (please give us weblinks).
A3: Follow the registration link on the ITU AI/ML Challenge website in order to register for the challenge. You need to create an ITU guest account. The creation of a guest account is automatic and immediate. To create a guest account, please go to <https://www.itu.int/en/ties-services/Pages/login.aspx>. If you are already registered using guest account, please login using the Username and Password you provided when creating your guest account. If you are facing problems creating the guest account, please email us at ai5gchallenge@itu.int.
Problem Statements
Q4: Will you provide data for the problems tagged as "more problems to come"?
A4: We expect all problem statements and data to be available by 15 June 2020.
Q5: When will the other problem statements will be available?
A5: We expect all problem statements and data to be available by 15 June 2020.
Q6: Can a group formulate a restricted question in a different environment?
A6: Yes. Please contact us to submit the problem statement and corresponding data.
Q7: Some problem statements e.g. "Radio signal coverage analysis and prediction based on UE measurement report" is marked in the original catalog document as "This problem statement is open to all participants." but in the latest version you presented as a restricted one. Is it then possible to work on that or not?
A7: The latest version of the problem statement document has the updated classification. The updated document can be found here.
Q8: What is the reasoning behind the restricted problem statements? Doesn't that defeat the purpose of an open collaborative forum?
A8: Data availability is a subject of local rules and regulations and classification by data providers. We have problem statements with open and restricted data. We are also working with multiple partners to make open data available.
Q9: What next after picking a problem statement (I have sent an email to the contact person but no reply).
A9: Analyse the problem statements currently available, pick one or wait for the deadline to baselined problem statements (June 15, 2020). Data will be made available by that deadline. After selecting a problem statement we recommend to fill in the survey link and read up on any references. Once data becomes available, you can also do data exploration. We also have a Slack Channel set up to engage and guide you.
Q10: Could participants also suggest some Problem Statements?
A10: Yes. Please contact us to submit the problem statement and corresponding data.
Q11: Are you going to host a URL where we can access the info, and send our problem statements?
A11: Yes, you can send via: follow link here or just email us at ai5gchallenge@itu.int
Q12: Should participants contact the contacts listed in the problem statements directly?
A12: For technical details of the problem statements, yes, you can. You can also email us at ai5gchallenge@itu.int. For any other logistics, better to email ai5gchallenge@itu.int first. We also have a Slack Channel set up to engage and guide you..
Q13: I'm unable to find the problem statement listed BNN.UPC (Spain) , UFPA (Brazil)
A13: You can find them via the regional host links: https://bnn.upc.edu/challenge2020 and http://ai5gchallenge.ufpa.br/
Also, here is the link to the latest problem statement document
Q14: If we propose different problem statement, would ITU help for providing data?
A14: we could discuss this based on your problem statement. Please submit your problem statements to us via ai5gchallenge@itu.int
Teams / Team Formulation Q15: Is there a limit for the number of members on a team?
A15: Yes, we would limit it at 4.
Q16: Is it possible to take part as a mix of graduate students and Profs (as a group) and get the privileges of Students?
A16: Yes. It would also be our pleasure to work with Profs to guide the students along with our experts.
Q17: Can multiple teams pick same problem statement?
A17: Yes. Your university or lab could have multiple teams who pick the same problem statement, in which case you will be competing with your colleagues. Globally, of course there will be many teams which pick the same problem statement and compete to win.
Q18: Can you please clarify whether you can have teams that mix students and professionals and how they are classified?
A18: This depends on how you declare the team as, based on your estimated contribution by students and professionals. If you declare the contribution of students to be high, we would treat it as primarily a students' team and vice versa.
Q19: Is it possible to participate in teams with members from different countries in the open challenges?
A19: Yes. You could tell us about your team structure via the survey link which will be shared with you.
Q20: Can we team up with students who have registered but are also open to team up with others?
A20: Yes, You could tell us about your team structure via the survey link which will be shared with you.
Evaluation
Q21: Is there going to be a leaderboard of some sort like in Kaggle?
A21: Yes. We will maintain a leaderboard which will be updated regularly.
Q22: What is the due date for contestants to submit solutions? and the format of the documents that should be delivered?
A22: These are mentioned in our website and for each problem statement.
Prizes
Q23: What are the awards and prizes?
A23: We are working on 2 types of prizes. Regional hosts may provide prizes for the "global round", for the top winners. Also ITU may give certificates and prizes if you are decided winner during the final event. This is still in progress. Please check our website to be updated with this information.
Publications
Q24: Will there be any restrictions to publication of the results obtained during the challenge (including, of course, proper citation to the dataset sources and related works)?
A24: In general, no restrictions as long as you provide citation, but we will confirm with contacts in the problem statements that you are selecting.
Other
Q25: Can we use pretrained networks for Graph NN challenge?
A25: No, solutions must be trained only with samples included in the training dataset we provide. It is not allowed to use additional data extracted from other datasets or synthetically generated. Please, see the "rules" section in the website. https://bnn.upc.edu/challenge2020
