Jesús F. Cevallos QwertyJacob

👋 Hey, I'm Jesús F. Cevallos-Moreno

🧠 I am postdoc researcher and lecturer @ University of Insubria

New Tutorial Out Now!

Deep Active Inference in PyTorch

👉 Check the Tutorial Now!

🔬 Research Focus

Algorithmic Reasoning Inductive Biases for Cybersecurity

Human-centred Language models

🎯 Recent Publications

🛡️ Ethical-AI

🤖 Open-Ethical AI: Advances in Human-Centric NLMs

ACM Computing Surveys, Vol 57(4), 2024
Survey on open-source ethical AI developments

🧠 Ethical Treatment of Language Models Against Harmful Inference-Time Interventions

Engineering Applications of Artificial Intelligence, Vol 172, 2026
Can LLMs resist harmful representation engineering???

⚙️ Neural Algorithmic Reasoning for Cybersecurity

💡 NERO: NEural algorithmic reasoning for zeRO-day attack detection in the IoT: A hybrid approach

Computers & Security (2024)
Extreme value theory, the algorithmic way!

🦸🏻 HERO: From High-dimensional network traffic to zERO-Day attack detection

Computer Networks (2025) Full firepower of the Neural Algorithmic Reasoning paradigm for zero-day attack detection!

🏙️ SmartVille: A Framework for Realistic Deep Learning-Based Online Network Intrusion Detection

Journal of Network and Systems Management (2026)
The open, live testbed behind NERO, HERO and ASAP: online, open-world intrusion detection.

🧪 Inductive Biases for cybersecurity

🧠 ACID: beta-testing ACtive Inference for active cyber-Defence

Journal of Reliable Intelligent Environments (Springer), 2026

A powerful optimal-control framework needs a powerful task to be tested against!

🐯 TIGER: an open-source cyber-Threat Intelligence Game Environment for Reinforcement learning

IEEE Transactions on Network and Service Management, 2026

A game environment to train and benchmark RL agents on cyber-threat-intelligence tasks.

🧠 ASAP: Automatic Synthesis of Attack Prototypes

Computer Networks (Elsevier), 2024
End-to-end online-learning approach using the relational-bottleneck Inductive Bias for Prototype attack synthesis from raw network traffic.

🚀 Collaborations

🛡️ Deep RL for vCDN Service Function Chain Deployment

Online optimization for live-streaming CDNs using dueling DDQN. Published in Future Internet (MDPI).

🧬 DeepReGraph

Heterogeneous Graph Learning for Gene Regulation

Deep learning approach for understanding embryonic regulatory mechanisms through graph representation learning.

🚄 SERENA-IIoT

Collaborative research on secure and reliable IoT systems. Check out our project website!

🛡️ SereBench: Adversarial Robustness Benchmarking for Security and Reliability Event Classification in Networked Railway Cyber-Physical Systems

Computer Networks (Elsevier), 2026

How far can adversarial perturbations push ML-based security & reliability event classifiers in the railway IIoT?

🚆 SereWay: Toward SEcurity and REliability Benchmarking for the RailWAY IIoT

Journal of Network and Systems Management (2026)

Open-source testbed benchmarking ML-based security & reliability solutions for the railway IIoT.

🚉 Open-FARI: An Open-source testbed for Federated Anomaly detection in the Railway IIoT

IWCMC 2025

First open-source testbed for federated-learning benchmarking in the railway domain.

🧬 Learning at the nano-scale: how to dynamically protect data in nano-network transmissions

Annals of Telecommunications (2026)

DNA-based bio-molecular cryptography meets ML-driven security configuration for the Internet of Nano-Things.

🧑‍🏫 Teaching

🎲 Probability for computer scientists - Insubria University

A collection of Google Colab notebooks to support the teaching of Probability and Statistics for Computer Science students.

🤔 Active Inference made (almost) easy

Tired of seaching for Active Inference complete tutorials without missing math proofs? Here's a Google Colab notebook that makes a full journey in pytorch and algebra!

🧑‍🏫 Guest Lectures

IN MY FREE TIME, I like to give talks on AI and its intersection with cybersecurity, ethics and philosophy for broad audiences. Here are some of my side projects:

📚 AI and the Anthropology of Fernando Rielo

Universidad Técnica Particular de Loja (Ecuador), 2024 - Guest Lecture for broad audience of the thought of Fernando Rielo

📚 An extremely gentle introduction to Active Inference

Insubria University (Italy), 2025 - Guest Lecture for the undergraduate course on Legal Informatics

📚 Inductive Biases 4 meta-learning

Federico II University (Italy), 2024 - A Research Retrospective

📚 Machine Learning meets the Internet of Things

Insubria University (Italy), 2023 - Guest Lecture on the Innovative IoT PhD course.

📚 AI working principles

Insubria University (Italy), 2024 - Guest Lecture for the undergraduate course on Legal Informatics

📚 AI and Cybersecurity

Insubria University (Italy), 2023 - Guest Lecture for the undergraduate course on Legal Informatics

💬 Let's Connect

I'm always open to discussing research collaborations, interesting AI problems, or just chatting about the latest developments in deep learning and cybersecurity.

📧 Email: jf.cevallosmoreno@uninsubria.it
🏢 Office: Department of Theoretical and Applied Science (DiSTA), University of Insubria, Varese, Italy

"Creativity, communication and commitment for the team. That's leadership."