8 July, 2021

2nd HACKATHON

MUSKETEER 2nd hackathon scheduled to September, 28 & 29 2021

MUSKETEER 2nd Hackathon: Attacking federated learning scenarios – Come and try to develop attacks capable of penetrating our defences in a federated learning environment.

About this Event

Date

September 28 & 29, 2021

Registration link

MUSKETEER | 2nd HACKATHON: Attacking Federated Learning Scenarios – Sep 28 | Hopin

Description

You will have the opportunity to attack our MUSKETEER platform and try to penetrate our defences and affect our federated learning training in 3 different scenarios.

SCENARIO 0

No defences. Participants will work on developing and implementing poisoning attacks on their own for a scenario with a number of honest clients and malicious users in a no defences scenario.

SCENARIO 1

Participants will work on developing and implementing poisoning attacks on their own for a scenario with honest clients and malicious users. In front there will be a defence method performed by the consortium members.

SCENARIO 2

Participants will work on developing and implementing poisoning attacks on their own. The participants will perform a black box attack against our system.

COVID-19 Update: the current crisis does not allow us to properly welcome the participants in a physical meeting. Therefore, this hackathon will be 100% online. We know the importance of meeting in person, especially for such a format of event. Therefore, we’ll do our best to organize and foster interaction among participants during the two days.

Entry skills

  • Solid Python 3 programming skills, experience with training simple classifier models in Keras, basic understanding of federated learning.
  • Familiarity with federated learning will be ideal.
  • Participants are required to bring their own laptop.

Agenda

Hackathon Agenda 1st day – Date: Tuesday, September 28, 2021

10:00-10:10 – Welcome and introduction (TBD) 

Objectives: Welcome and high level MUSKETEER introduction. 

10:10-10:30 – Technical Talk: Introduction to federated learning in MUSKETEER 

Objectives: Organisers provide a general introduction to Federated Learning in general (concepts, roles etc.) and the research under the MUSKETEER project in particular. General talk, intended to motivate the importance of FL, introduce general concepts / terminology, provide intuition about the working of the platform. 

10:30-10:50 – Technical Talk: Introduction to MUSKETEER attacks and defences

Objectives: Organisers provide a general introduction to poisoning attacks and defences in general (concepts, roles etc.) and the research under the MUSKETEER project in particular. General talk, intended to motivate the importance of a robust environment, introduce general concepts / terminology, provide intuition about the working of the platform. 

10:50-11:00 – Hackathon rules, guidelines, general instructions, Q&A

Objectives: Rules with details on number of teams, overview of the agenda, details of communication channels, a walk though of the instructions, and logistics of evaluation, the assignment of group mentors and details on breakout rooms will be provided. By the end of this session, it is expected that participants should be ready to dive into the code and organisers should be ready with their breakout room setups. 

11.00-11:15 – Coffee Break 

11:15-13:00 – [Breakout rooms] Hacking phase 1 (Mentors & participants) 

Objectives: This session will begin in the breakout rooms where assigned mentors will work with the teams to help them run the FL setups.  

  • Each group executes a complete FL task (Scenario 0 – No defences).
  • The mentors will act as the aggregator and 8 honest clients. The participants will act as 2 malicious clients and will run FL via pycloudmessenger using fresh credentials provided to them by their mentor.  
    Any remaining technical issues (e.g. missing dependencies) can be resolved here (if participants attended the prep session the week before there should not be any)   
  • Participants will work on finishing poisoning attacks on their own for a scenario with 8 honest clients and 2 malicious users. For local development and testing, they can either connect among themselves through pycloudmessenger.

13:00-14:00 – Lunch break 

14:00-15:45 – [Breakout rooms] Hacking phase 2 (Mentors & participants) 

Objectives: Scenario 1: (Defence method, 8 honest users, 2 malicious users) 

  • To evaluate the developments the mentors will run the aggregator provided with specific defence and 8 honest clients. The participants will run 2 malicious clients. 
  • One organizer/mentor will be available for each group, through the platform..  
  • Organizers internally communicate among themselves (how are all the groups doing, do we need to lower / raise the bar, provide hints etc.) via private Slack channel and/or dedicated call.  
  • The developments can be evaluated any time the participants want in scenario 1
  • The required, End-of-Day-1 evaluation (still on Scenario 1) will be conducted by the mentors. 

15:45-16:00 – Summary of Day 1. 

  • Participants can share impressions, lessons learned. 
  • Q&A. Qualitative feedback 
  • Brief participants on what to expect on Day 2 (more challenging scenarios that would penalize overfitting to specific attacks) 

Hackathon Agenda 2nd day – Date: Wednesday, September 29, 2021 

10:00-10:15 – Recap from Day 1 and outlook on the day

10:15-12.00 – [Breakout rooms] Hacking phase 3 & Final Evaluation. (Mentors & participants) 

Objectives: Scenario 2: (Defence method, 8 honest users, 2 malicious users) 

  • To evaluate the developments the mentors will run the aggregator and 8 honest clients. The participants will run 2 malicious clients. 
  • One mentor will be available for each group, also on Slack and check-in once per hour. 
  • Organizers communicate among themselves to know how are all the groups doing, do we need to lower / raise the bar, provide hints etc. through private Slack channel and/or dedicated call.
  • The developments can be evaluated any time the participants want in scenario 1
  • Collect final evaluations on Scenario 0, 1, 2. [Decision on how to decide the final leader boards based on the results]. Also ask the participants to walk us through their solution (explain their algorithm(s)). 

12.00-12:15 – Coffee Break 

12:15-12:45 – Technical talk on robustness and description of Defence Scenarios used in the Hackathon

Objectives: technical talk on robustness of federated machine learning and details about the actual defences that were used. 

12:45-13:00 – Assembly, winner ceremony, virtual group photo, and closing remarks

Organizers

MUSKETEER aims to develop an industrial data platform with scalable algorithms for federated and privacy-preserving machine learning techniques, detection and mitigation of model capable of fairly monetizing datasets according to their real data value. MUSKETEER is an H2020 project that has received funding from the European Union’s Horizon 2020 research and innovation programme under grant agreement No 824988.