Skip to content

Latest commit

 

History

History
44 lines (36 loc) · 2.42 KB

final_phase.md

File metadata and controls

44 lines (36 loc) · 2.42 KB

Final Phase

Welcome to the finals phase for CSAW ESC'22! This document will discuss more about the finals and how to install the Raspberry Pi Desktop VM. A few things to note while solving the challenges:

  1. The cloud server, which runs inference procedures for most challenges, has an access control policy that requires a team key (provided via email by the organizers). To prevent against brute force attacks, each team has a daily quota that will allow them to conduct a reasonable number of inference procedures.
  2. All communication with the server must be done through the Raspberry Pi VM; this is enforced by technical constraints imposed by the provided communication binaries.

Challenge Structure Overview

Throughout the next few weeks several challenges will be released. All points from all challenges will be accumulated in the final score of each team. Challenges will be given a different number of points based upon their difficulties; the criteria for points will be given in the description of each challenge.

Raspberry Pi Desktop VM Overview

This year, we have prepared an ISO to use with VMWare. This VM will act as a client and send various inputs to a cloud server that runs the machine learning algorithms for each challenge. We recommend running the virtual machine in "Run with Persistence" mode and pausing the VM instead of doing a full shutdown; shutdowns may result in loss of data.

Submission Details

See the Final Phase section in deliverables.

Verifying Solutions

The challenge leaders will verify the solutions that each team has submitted before the day of the live finals. For the Technical Track teams, the organizers will run the provided inputs as indicated by each individual challenge description with the precompiled script on the VM and assign correctness points based on the criteria outlined in the challenge description. For the Research Track teams, the organizers will run the provided solutions on a cloud server and use the provided evaluation scripts on the Raspberry Pi VM. Correctness points will be awarded based on the criteria outlined in the challenge description.