news

Sep 25, 2026 Our paper “Disentangling the Good From the Bad: Quantization-Induced Flips Are Not Random” was accepted at NeurIPS 2026 (December 2026, Sydney, Australia, with satellite events in Atlanta and Paris)!!! :tada: :tada: :tada:

Disentangling the Good From the Bad: Quantization-Induced Flips Are Not Random

Authors: Aymen Bouguerra, Alexandra Gomez-Villa, Chokri Mraidha, and Fabio Arnez

NeurIPS 2026, Sydney, Atlanta & Paris.
Aug 31, 2026 Two of our papers were accepted at IEEE ICVES 2026 (November 9-11, 2026, Cochabamba, Bolivia)!!! :tada: :tada: :tada:

Quantization and Corruption Robustness in Deployed Road-Scene Perception

Authors: Aymen Bouguerra, Ansgar Radermacher, Chokri Mraidha, and Fabio Arnez

Drive-SynOOD-OD: A Diffusion-Inpainted Synthetic Out-of-Distribution Object Detection Benchmark for Autonomous Driving

Authors: Daniel Montoya, Mauricio Sayri Espinoza Mayzer, and Fabio Arnez

IEEE ICVES 2026, Cochabamba, Bolivia.
Jul 15, 2026 Our new preprint “The SIGReg Objective as Variational Free Energy: A Theoretical Active-Inference Account of JEPA World Models” is now available on arXiv! :page_facing_up:

The SIGReg Objective as Variational Free Energy: A Theoretical Active-Inference Account of JEPA World Models

Authors: Fabio Arnez and Alexandra Gomez-Villa

Preprint on arXiv.
May 13, 2026 Honored to be recognized as a Gold Reviewer at ICML 2026, placing among the top reviewers for this year’s conference! :tada:
Apr 30, 2026 Our paper “Less Precise Can Be More Reliable: A Systematic Evaluation of Quantization’s Impact on VLMs Beyond Accuracy” was accepted at ICML 2026 (July 6-11, 2026, Seoul, South Korea)!!! :tada: :tada: :tada:

Less Precise Can Be More Reliable: A Systematic Evaluation of Quantization’s Impact on VLMs Beyond Accuracy

Authors: Aymen Bouguerra, Daniel Montoya, Alexandra Gomez-Villa, Chokri Mraidha, and Fabio Arnez

ICML 2026, Seoul, South Korea.
Apr 13, 2026 Our paper “FedSALAT: Adaptive Buffer-Based Active Learning for Federated Data Streams” was accepted at FLICS 2026 (June 9-12, 2026, Valencia, Spain)!!! :tada: :tada: :tada:

FedSALAT: Adaptive Buffer-Based Active Learning for Federated Data Streams

Authors: Prajit T Rajendran, Fabio Arnez, Huascar Espinoza, Agnes Delaborde, Chokri Mraidha

FLICS 2026, Valencia, Spain.
Mar 31, 2026 Our paper “Digital Twins and World Models: A Systematic Taxonomic Disambiguation” was accepted at the MIDas4CS 2026 workshop, part of CAiSE 2026 (June 9, 2026, Verona, Italy)!!! :tada: :tada: :tada:

Digital Twins and World Models: A Systematic Taxonomic Disambiguation

Authors: Kunal Suri, Fabio Arnez

MIDas4CS 2026 Workshop at CAiSE 2026, Verona, Italy.
Nov 03, 2025

New Blog Post: Uncetainty Quantification & Propagation in a DNN-based Navigation System

After a long long pause in writing blog content, I wrote a new post about “Uncertainty Quantification & Propagation in a DNN-based Navigation System”.

The post describes how to quantify and propagate uncertainty in a minimalistic UAV/drone DNN-based navigation system (2 neural networks) and shows how to use uncertainty to improve the navigation system’s performance inside the AirSim simulation environment.

For more details, check the post here.
Sep 10, 2025 Our paper “Oracle-Guided Soft Shielding for Safe Move Prediction in Chess” was accepted at ICMLA 2025!!! :tada: :tada: :tada:

Oracle-Guided Soft Shielding for Safe Move Prediction in Chess

Authors: Prajit T Rajendran, Fabio Arnez, Huascar Espinoza, Agnes Delaborde, Chokri Mraidha

ICMLA 2025 website. ICMLA 2025 accepted papers here.
Aug 26, 2025 Our paper “The Map of Misbelief: Tracing Intrinsic and Extrinsic Hallucinations Through Attention Patterns” was accepted at ATRACC-25, part of AAAI’s FSS-25!!! :tada: :tada: :tada:

The Map of Misbelief: Tracing Intrinsic and Extrinsic Hallucinations Through Attention Patterns

Authors: Elyes Hajji (MSc. student intern), Aymen Bouguerra (Ph.D. student), and Fabio Arnez

ATRACC-25 Symposium accepted papers here.
Jul 01, 2024 Our work was featured in the CEA List Research Report! :sparkles: :smile: Check out the full report here.

A Safety supervision environment for autonomous systems

CEA-List has developed a runtime safety supervision environment for autonomous systems built using AI. It factors in the uncertainties related to the system and to its environment to effectively determine the level of safety and potential risks. The system has been evaluated on an autonomous drone (UAV) use case.