Blog | Roberto Santana

This blog is being set up. Posts will cover research findings, discussions of recent papers, technical tutorials, and reflections on AI trends. Check back soon or follow on Twitter/X for updates.

Machine Learning

Coming soon

Adversarial Attacks in Explainable ML: What Can Be Done?

A discussion of our recent survey on adversarial threats against explainable machine learning models and humans. What makes explanations vulnerable, and how can we defend them?

Coming soon

Large Language Models for Feature Selection: First Steps

Can LLMs be useful for feature selection in anomaly classification? We explore stacking LLM predictions as a novel approach to feature engineering.

Coming soon

Domain Adaptation for Brain Decoding

Bridging the gap between visual perception and mental imagery in brain decoding: how domain adaptation with searchlight analysis enables cross-paradigm classification.

Evolutionary Computation & Neural Architecture Search

Coming soon

Neuroevolution in 2025: Where Are We?

A perspective on the state of neuroevolutionary algorithms, their applications in neural architecture search, and the new frontiers opened by large-scale models.

Coming soon

Factorized Models in Neural Architecture Search

How factorized representations in NAS impact computational costs and performance. Summary of our recent work presented at IJCNN 2024.

Optimization & Scheduling

Coming soon

Deep RL for Job-Shop Scheduling: An Overview

Deep reinforcement learning has shown remarkable results for the flexible job-shop scheduling problem. We review our approach combining constraint programming with DRL for real-time scheduling.

Coming soon

Offline RL for Combinatorial Optimization: Lessons Learned

What happens when you apply offline reinforcement learning to industrial scheduling? A discussion of our results and the open challenges in this direction.

Large Language Models in Research

Coming soon

Using LLMs to Enhance Scientific Writing and Code

Reflections on how large language models are changing research workflows: from paper writing to code generation, literature review, and beyond.