2023 Tech Arena UK

Discovering New Algorithms for Predictors in Computer Architecture!


Overview of the tasks

Details of the tasks HERE

Before the release of the detailed descriptions of the 2 tasks, here is an overview of the requirements and what will be asked for you to prepare ! 
More details will be shared with you during the 2 webinars
  • Task 1: You will receive a sequence comprising integer elements. Your objective is to create an algorithm that accurately predicts the subsequent element in the sequence while observing the elements of the sequence one by one, leveraging the context provided by previous elements. The accuracy of the algorithm is calculated as a ratio between the number of sequence elements predicted correctly between the total number of elements in the sequence. The challenge is that your prediction algorithm should work very fast and use very little memory.
  • Task 2: For the subsequent task, your goal continues to be maximizing the accuracy of the algorithm for predicting the upcoming element in the provided sequence under memory and compute resource constraints. However, a new dimension is introduced: you are empowered to manipulate the given input sequence by introducing additional numeric values between its original elements. Your goal is to identify the optimal numeric values that should be introduced within the original sequence, with the primary aim of enhancing prediction accuracy to its maximum potential.
  • In both assignments, you are required to develop two distinct functions: one for memorizing the currently observed element and another for predicting the subsequent element within the sequence. These functions should be implemented using the C language (or Cython) and must be submitted in the files named predictor.h and predictor.c. The templates for these files will be provided upon the commencement of the challenge. Additionally, you will receive the input sequence and the necessary infrastructure to evaluate and validate the accuracy of your solution.

If you are:

  • An undergraduate, Master's, or PhD student 
  • Registered at a higher education institution in the United Kingdom
  • Studying engineering, mathematics, computer science, or other related fields
>> Then you are the perfect candidate!