LLMs vs. the World

Context

LLMs are revolutionizing the way in which we think, build and run code. However, there is increasing concern in their characteristincs overweight their usability. In this thesis we are going to explore different LLMs to execute software engineering tasks and evaluatem to answer the question, is it really worth it?

Project proposal

Everyone thinks LLMs are great, we don’t. Let’s explain to everyone why this is the case.

The idea behind the project is to choose a specific application domain from software engineering and evaluate it against existing solutions using LLMs. Among the topics to choose from we have (but others could be proposed):

The implementation of this thesis is thought out to be between multiple students, each working in an independent topic, but all trying to prove the same, LLMs are less than great.

Implementation plan

The implementation plan for the projec consists on:

  1. Initial task identifying state-of-the-art algorithms to solve the problem
  2. Creation/extraction of a data set
  3. Execution of the identified solution on the dataset
  4. Selection of LLM(s) to use
  5. Execution of the LLM on the dataset
  6. Comparison and Analysis

Background and Literature

LLMs and your topic of choice

Contact

n.cardozo


Universidad de los Andes | Vigilada Mineducación
Reconocimiento como Universidad: Decreto 1297 del 30 de mayo de 1964.
Reconocimiento personería jurídica: Resolución 28 del 23 de febrero de 1949 Minjusticia
Edificio Mario Laserna Cra 1Este No 19A - 40 Bogotá (Colombia) | Tel: [571] 3394949 Ext: 2860, 2861, 2862 | Fax: [571] 3324325
© 2017-2026 - Departamento de Ingeniería de Sistemas y Computación