Devpost
Participate in our public hackathons
Devpost for Teams
Access your company's private hackathons
Grow your developer ecosystem and promote your platform
Drive innovation, collaboration, and retention within your organization
By use case
Blog
Insights into hackathon planning and participation
Customer stories
Inspiration from peers and other industry leaders
Planning guides
Best practices for planning online and in-person hackathons
Webinars & events
Upcoming events and on-demand recordings
Help desk
Common questions and support documentation
Loosely followed the approach by https://doi.org/10.1038/s41598-022-12201-9
We compared different models on the classification with some feature engineering and achieve over 85% ROC AUC Score on the final model.
As a part of the intact challenge for CxC 2023, I trained a Medical Document classifier using HuggingFace's XLNet NLP Model. The project contains the predictions, and a notebook outlining my process.
Submission For Cyclia
Naive Bayes machine learning model to classify Intact medical data.
Privacy-safe synthetic data and auto-labeling for real-world AI training.
Explored how staffing influences profitability by correlating financial performance with workforce size, determined the required workforce at different times of the day and analyzed tipping trends.
Analysis of North American and International economic sectors with investments.
Error Submission
The provided model can predict the required medical specialty needed (such as Emergency, Surgery) from a text description provided.
Data Comarade: Feiyang Li, Huaye Zeng, Kunling Yang & Tianyi Tang.
Although it probably has the worst accuracy, it is probably the best beginner work :) A for effort?
Researchers now test every query residue to check if they are drug-binding structures. On the other hand, we have developed a classification model that can predict if the residue by it’s features.
Classification with Intact - Predict Medical Specialty from Medical Notes Best Model: with Macro F-1 Score of 0.11 by BERT
CxC submission for the Intact dataset
Predicting Drug-Binding Sites in Proteins using ML models. Model will predict probable drug-binding sites and importance of features for drug-binding. A step towards drug discovery and new medicines.
CxC Group Project
Predicting drug binding sites on AlphaFold2-predicted proteins
Implemented a neural network using Python, Pandas, NumPy, and Tensorflow, to predict 'drug binding' or 'non-drug binding' for any query residue on an AlphaFold2 predicted protein model.
The tree-based method correctly predicted 30% of positive cases and 90% of negative cases. The neural net correctly predicted 90% of positive cases.
Predict drug binding sites on AlphaFold2 predicted 3d protein structures
Intact needed a supervised machine learning algorithm to determine what specialty a patient required based on a medical transcription.
#ds
A neural network that outperforms out-of-the-box models in all regards
1 – 24 of 31
