Dear organizers, Would you please provide a baseline solution for participants to download and submit? I would like to ensure that I could submit a baseline model which could be scored correctly by your platform. Yours, Tsai-Min

Created by TSAI-MIN (在民) CHEN (陳) chentsaimin
Hello Tsai-Min, All participants should now be able to create up to 3 separate projects. Cheers.
Dear organizers, Thanks for your help. I could see my team score now in the Leaderboard of “500 Bootstraps”, but I still could not see it of “10 Bootstraps.” Although I could use the "Save and History" features to track different models/submissions, it's much straight and convenient for me to create different projects for different models with different causal rationale. I hope you could consider about releasing the limitation of maximum number of projects. Yours, Tsai-Min
Hello Tsai-Min, 1. A fix for issue 1 is deploying now and should be active in ~1 hour, we apologize for the inconvenience. 2. Multiple models can be submitted from the same project. You can use the Save and History features to track different models/submissions. Cheers.
Dear organizers, Thanks for your help. I could run G-Formula Estimation and receive successful estimation results after I sent the mail to "'challenges@cstructure.io." However, I encountered another 2 issues: ``` 1. I could not see my team score in the Leaderboard of either “10 Bootstraps” or “500 Bootstraps.” 2. I could not create more than one model project for building up 3 models for Mid-point Validation. It said "I have reached the maximum number of projects." ``` Would you please help me fix them? Yours, Tsai-Min
Hi Haoze, Glad to hear you can submit a model. I mentioned during the webinar that there are certain features that fill fail. Binary features with high missing rates will cause convergence problems for the g-formula. Please consult the Data Profile to assess whether a feature is likely to fail to due to very high missing rates. Quantitative features are generally less likely to cause problems. The leaderboard now includes an example model with age and a quantitative feature (egfr) as an example. Models that fail to generate a result should return an error message - this will be fixed by Monday.
Dear Organizers, I have a follow-up question regarding the G-formula estimation. After updating my email address to match the one used during registration, I am now able to initiate the estimation. However, the job remains in the "Queued" status and does not return a result—unless the model used is exactly the same as the baseline model. Could you please advise on how to resolve this issue? Best, Haoze
Dear organizers, I encountered the same issue to @talhz 's: ``` I followed the instructions provided to run the baseline model. However, after clicking "Run G-Formula Estimation," I encountered the following error message: "Estimation Error. Apologies, something went wrong". ``` I also have confirmed that my email address for the interest form is the same to my email address for creating my cStructure account. However, I still could not solve the issue. Could you please help me here? Yours, Tsai-Min
Hi Scottie, Thank you so much for your prompt reply! The problem has been solved after I change my email address. Best, Haoze
Hello Haoze, Can you email challenges@cstructure.io? I suspect that the email address you used for the interest form does not match the email used to create your cStructure account. These must match for a valid analysis request to be submitted because we use the Team Name to identify the model submission. You can fill out a new interest form with the email address you used to create your cStructure account. Alternatively, you can create a cStructure account using the interest form email. Also, please ensure that STEROIDS is used as a time-varying covariate.
Dear Organizers, I followed the instructions provided to run the baseline model. However, after clicking "Run G-Formula Estimation," I encountered the following error message: "Estimation Error. Apologies, something went wrong". Could you please let me know if I might be missing any steps? Best, Haoze
Hi Tsai-Min, 1. Registration occurs at https://cstructure.net. There is a Registration button (Register for the challenge) that opens a form. Each participant on your team should fill out the Registration form, then click the 'Sign up for an account' button and use the same email address as the Registration form, and finally sign in to the platform. 2. Two separate teams will be fine. If you will be sharing knowledge, then you should cap your per weekly model submissions to 6 across the two teams. 3. All are welcome to participate.
Dear organizers, Thanks for your update. I would like to confirm the team registration procedure in Participation Requirements and Process: ``` Eligibility Team size limit: 1-4 members per team Multiple team memberships not permitted No specific prior experience with causal inference required Registration Process Pre-registration opens on April 15th, 2025 at https://cstructure.net (orange button) Complete team registration form with all member information Accept challenge rules and data use agreements Receive platform access credentials ``` 1. When and where should I complete team registration form with all member information? 2. If my team members are more than 4 people, should I separate them to form 2 independent teams with shared knowledge in between to participate the challenge? 3. Does it mean that a person who has prior experience with causal inference is not allowed to participate the challenge? Yours, Tsai-Min
Hello Tsai-Min, The challenge no longer has this residency restriction. I have updated the wiki language, thank you for bringing this oversight to our attention. Cheers
Dear organizers, Thanks for your clarification. I just noticed your updated information in Participation Requirements and Process: ``` Eligibility Open to individuals and teams residing in the United States and Canada only ``` Does it mean that anyone, for example, living in Japan, can not participate in this challenge? Yours, Tsai-Min
Hello Tsai-Min, Thank you very much for the question. All submissions occur through the Estimate sidebar in the cStructure interface. A valid submission will require the action node, outcome node, and all adjusted nodes to have a feature and feature category assigned in the Data sidebar tab. A model with features already layered onto the Age, 28d Survival, and Steroids nodes can be downloaded from this linked page: https://www.synapse.org/Synapse:syn66534452. This model can be imported into cStructure using the following process: Model Sidebar -> Import SCM -> select ccddc_age_base_model.json -> Open -> clear. Then click the Age node and press the 'a' key to change the node status from 'unadjusted' to 'adjusted'. To submit this causal diagram, open the Estimate Sidebar Step 1: Name the model, and select whether a feature should be modeled as a static (i.e. AGE) or time-varying (i.e. STEROIDS) covariate using the checkboxes, then click Generate R Script. Step 2: Click Run G-Formula Estimation. This step can take as little as a few seconds if the model has been previously run. If this is a novel and complex model submission, it can take several hours to run the bootstraps, though we try to return a n=10 bootstrap result within 10 minutes. Step 3: Wait and review the results. Screenshots of the submission process can be found in the same folder as the base model: Step 1: https://www.synapse.org/Synapse:syn66534453 Step 2: https://www.synapse.org/Synapse:syn66534455 Step 3: https://www.synapse.org/Synapse:syn66534454 There was an error that prevented the results for the base age model to render in the Results section that has now been resolved. We will be holding a webinar on Tuesday May 20th to answer questions and provide an in-depth tutorial for the challenge. Materials from the webinar will be provided on synapse.

(solved)[Submission] Asking for Baseline solution page is loading…