{
  "fetched_at": "2026-10-20T01:59:31Z",
  "source_url": "https://www.kaggle.com/competitions/gan-based-training-for-binary-classifier/data",
  "pages": [
    {
      "name": "data-description",
      "content": "- The training set consists of 50 VirusTotal-validated malware from the Androzoo database, and their malicious functionalities were confirmed within Android phones. As for the benign applications, 60 trusted apps from Google Play Store and factory/system apps were used.\n- Data consists of fixed time_window samples that show the monograms of system calls\n- There are 128 features and 4465 samples (3000 benign, 1465 malign) for training and GAN seeding purposes in the training set.\n- The validation set (validation.csv) contains malware behavior from 10 unseen malware applications (public) as well as 10 other unseen malware applications (private). \n\n\n## Files\n\n*   **train.csv** - the training set\n*   **Validation.csv** - the test set\n*   **sample_submission.csv** - a sample submission file in the correct format\n\n## Columns\n\n*   `F1-F128` - equivalent to Linux system calls where each sample is the histogram of the system calls at equal time windows.\n*   `Label` -  The class of the sample, 1 for malign and 0 for benign.",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Description",
      "content": "GAN-Based Training for Binary Classifier is a data challenge organized in association with [IEEE BigData 2022 conference](https://bigdataieee.org/BigData2022/BigDataCupChallenges.html). The challenge is to use pictorial-based or raw data-based GAN to generate malign samples suitable for classifier training. The competition could accept 10 papers based on the competition results to be published in the conference proceedings. Those papers should be presented at the associated workshop at the IEEE Big Data 2022 conference organized by each Cup Challenge and will be indexed in the same way as regular conference papers.\n\n## Background\nMalicious software, also known as malware, threatens our cyber services by infecting computational devices like computers and mobile phones. Nowadays, cyber security companies are oriented towards Machine Learning (ML) technologies over static detection techniques as a promising solution to overcome detection limitations and enhance their services. According to a Securelist's article published on 1st of Mar. 2021; the security tools of a single supplier (Kaspersky) detected in a single year 5,683,694 malicious installation packages, 156,710 new mobile banking Trojans, and 20,708 new mobile ransomware Trojans. Today ML-based solutions' reliability is mostly evaluated by collecting real malware inferences within an actual platform environment and not an emulator.\nBehavioral polymorphism is a product of malware developers’ attempts to avoid the existing detection mechanisms by changing the malware’s operational trace while keeping its main tasks intact. Hackers accomplish this by manipulating the malware code using multiple forms of obfuscation like payloads encryption, character code shifting of the variable name and register reassignments. For example, Agobot or Gaobot is a computer worm that was first identified in 2002.  Over 580 versions of this malware were subsequently identified, some of them behaviorally polymorphic. Today, each malware can mutate to generate hundreds of polymorphic versions, adding up to thousands for each malware family. Generative Adversarial Networks (GAN Systems) have been successfully used in solving computer vision, object detection, and classification problems by learning, mimicking, and generating any distribution of data. For example, they produced high-fidelity images visually indistinguishable from real ones or synthetic emails textually close to real messages. In the cyber security research field, one would like to use GANs to generate synthetic malware behavior, avoiding the challenge of manually collecting the training malware traces.\n\n## The Challenge:\nThe challenge is to develop a generative adversarial network to generate a proper training set for the an ML–based malware classifier.  For our competition, you should generate malware behavior samples in the same format as the manually collected ones.\nYou will notice that the manually collected training set, which contains a histogram representation of malware and ordinary software behavior, is unbalanced (3000 benign and 1465 malign samples) hence, the challenge is to:\n- Train an ML model using the unbalanced real dataset (train.csv) and submit the classification outcome on the validation dataset (Validation.csv) as to get the current RMSE feedback as a baseline. Only RMSEs below 0.2 will be considered.\n- Develop a GAN system using the available real dataset as guidance to generate synthetic samples of malware behavior.\n- Retrain the ML model using the now balanced dataset (after integrating the GAN generated malware synthetic dataset in the original dataset).\n- Test the model on the held-out validation set and submit the output. \nThe final evaluation will be carried out by obtaining the GAN trained model RMSE and computing the absolute value of the difference with the baseline model's RMSE. In case of a tie on this difference values, the confusion matrix of the GAN trained model will be considered to asses the quality of malign sample generation.\n\n## Acknowledgment:\nThis work is jointly supported between the [Center for Cyber Physical Systems (C2PS) at Khalifa University](https://www.ku.ac.ae/c2ps); and the [Technology Innovation Institute (TII)](https://www.tii.ae/) under Fund number 8434000379.",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Evaluation",
      "content": "## Track 1: Pictorial-based GAN for generating Malign Samples\nThe pictorial dataset is a data representation format that has shown good performance in generating artificial samples by using the GAN technique. If you choose this track, you will be required to transform the provided raw dataset into a pictorial form of your choice, use it to train the ML classifier and then extract some data items you will use to train the GAN. The comparison between the original ML model and the one trained with GAN data will be based on your chosen representation format.\n\n## Track 2: Raw Data-based GAN for generating Malign Samples\nIf you choose this track, you will directly use the provided raw data to extract the seeds for the GAN training. In addition, the comparison between the baseline and model trained on GAN data will be based on this representation format.\n\nThe solutions for track 1 and 2 will be evaluated based on the accuracy on the validation set (public) and test set (private) as well as on the quality of the generated malware samples.\n\n## Evaluation Metric\nThe evaluation metric for this competition is the RMSE (root mean squared error). \n\n## Tracks 1 and 2 Submission Format:\nAs given in Sample_submission.csv file\n\n```\nID, Labels\n1,0\n```",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "rules",
      "content": "## Getting Started Honor Code\nMake sure to use GAN generated samples and not samples created manually or another technique like SMOTE.\n\n## One account per participant\nYou cannot sign up to Kaggle from multiple accounts and therefore you cannot submit from multiple accounts.\n\n## Team Limits\nThe maximum team size is 3.\n\n## Submission Limits\nYou may submit a maximum of 2 entries per day. The first submission will be considered the baseline regardless of when it is done.\n\n## Final Leaderboard Score\nThe Kaggle leaderboard is to assist your performance and the final decision will be made based on your final report and the functionality of the code. You need to submit a zip file containing the report, source code, submission.csv and generated malware samples via email to hussam.alhamadi@hotmail.com , martina_nd@hotmail.com & sivaprasad.nandyala@tii.ae  by September 5th 2022 in order fairly evaluate each teams' submission.",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Timeline",
      "content": "- August 31, 2022: deadline for submitting the solutions.\n- September 5, 2022: deadline for sending the reports, end of the competition.\n- September 22, 2022: online publication of the final results, sending invitations for submitting papers to the associated workshop at the IEEE Big Data 2022 conference.\n- October 20, 2022: deadline for submitting invited papers.\n- November 8, 2022: notification of paper acceptance.\n- November 22, 2022: camera-ready versions of accepted papers due.\nAll deadlines are at 11:59 PM UTC on a corresponding day unless otherwise noted. The competition organizers reserve the right to update the contest timeline if they deem it necessary.",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Organizers",
      "content": "Prof. Ernesto Damiani, C2PS, Khalifa University, ernesto.damiani@ku.ac.ae\nDr. Hussam Al Hamadi, C2PS, Khalifa University, hussam.alhamadi@ieee.org\nEng. Martina Morcos, C2PS, Khalifa University, martina.morcos@ku.ac.ae\nDr. Sivaprasad Nandyala, SSRC, TII, sivaprasad.nandyala@tii.ae\nDr. Brian McGillion, SSRC, TII, brian.mcgillion@tii.ae",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Code Requirement",
      "content": "The participants will submit the source code to the organizers for validation to hussam.alhamadi@hotmail.com , martina_nd@hotmail.com & sivaprasad.nandyala@tii.ae using your Kaggle 'User Name' as email title and a zip file containing the following:\n1. Submission.csv (2 files: classification outcome when using unbalanced set to train the ML model, classification outcome when using balanced set to train the ML model)\n2. Source code \n3. Generated samples from GAN\n4. Report",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Prizes",
      "content": "Track 1\n\n1st Place - $1,000 + one free IEEE BigData 2022 conference registration\n\n2nd Place - $500 + one free IEEE BigData 2022 conference registration\n\nTrack 2\n\n1st Place - $1,000 + one free IEEE BigData 2022 conference registration\n\n2nd Place - $500 + one free IEEE BigData 2022 conference registration",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Important Announcement ",
      "content": "##Important News: Both tracks are now closed. You are encouraged to submit your code and report by email to [our team](https://www.kaggle.com/competitions/gan-based-training-for-binary-classifier/overview/organizers) if you haven't done so already. Please don't forget to include your name and affiliation in the report. The submissions made after 24th of Sept. 12:00 am (UAE Time) will not be considered.  ",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    }
  ]
}