19 lines
1.1 KiB
Markdown
19 lines
1.1 KiB
Markdown
---
|
|
title: "Eduardo Cueto-Mendoza"
|
|
image: "images/author.jpg"
|
|
description: "Author introduction"
|
|
sidebar: true
|
|
widgets: ["recommended","categories"]
|
|
draft: false
|
|
---
|
|
|
|
Hello, I'm Eduardo. Currently, pursuing a PhD in Computer Science. Some places where I have previously worked are: Intel, InfolinkEXP, and Indboo, as an AI engineer.
|
|
|
|
My current interests are optimization (combinatorial, convex, etc), neural networks (NN), and systems programming.
|
|
|
|
> Current research on Neural Networks does not consider parameter other than the accuracy obtained by the model on unknown data.
|
|
|
|
Currently I am working on unraveling the relationships that exists between several hyper-parameters of NN's. Current research on this area does not consider this type of hyper-parameters: network size, energy consumed, time, fitting methodology, etc.
|
|
|
|
there is research where those parameters are taken into account, but they are only presented as information, there is no real attempt to use those parameters as a guideline (or model) to produce a training/inference methodology that can optimize any of the parameters being studied.
|