updated news

Author: andremht

_bibliography/uscl_publications.bib

@@ -56,7 +56,7 @@ @inproceedings{Seifullaev_IFACWC2026
 	author = {R. Seifullaev and A. M. H. Teixeira},
 	title = {Impact analysis of hidden faults in nonlinear control systems
       using output-to-output gain},
-	booktitle = {IFAC WC (Submitted)},
+	booktitle = {IFAC World Congress (Submitted)},
 	published={0},
 	year={},
 	tag={10005}

_data/news.yml

@@ -1,3 +1,6 @@
+- date: 15 December 2025
+  headline: "Read the latest popular science piece about [our research funded by the Knut and Alice Wallenberg foundation.](https://kaw.wallenberg.org/en/research/smart-machines-learning-withstand-cyberattacks)"
+
 - date: 27 November 2025
   headline: "André was awarded a [WASP Industrial PhD project grant](https://wasp-sweden.org/opportunities/calls/funded-projects/), together with Zenseact, for the project titled ''Federated Reinforcement and Imitation Learning for Safe and Scalable Autonomous Driving''."  
 

_data/project.yml

@@ -5,7 +5,7 @@
   period: "2026-2030"
   ongoing: 1
   lead: "*André Teixeira* (PI-UU), together with Mina Alibeigi (PI-Zenseact)"
-  member: "Adrian Gheorghiu"
+  member: "Adrian Gheorghiu (PhD student at Zenseact)"
   selected: 1
   summary: "This project introduces a Federated Reinforcement and Imitation Learning (FRL+FIL) framework for safe, scalable, and privacy-preserving training of planning policies in (semi-)autonomous fleets. The approach enables vehicles to collaboratively learn planning policies without sharing raw data, thus preserving privacy while exploiting the diversity of large fleets. Planning policies will be bootstrapped from expert demonstrations and refined locally through shielded reinforcement learning, with runtime safety mechanisms—such as coarse planners and learned safety filters—ensuring safe exploration. Model updates will be aggregated using robust federated techniques that tolerate faulty or malicious clients with low-quality driving behaviors and support personalization for region-specific adaptations. The framework will be validated on both simulated and real-world driving datasets, paving the way for the industrial deployment of next-generation AD/ADAS systems that are safer, more robust, and human-like in their driving behavior."
 

_data/students.yml

@@ -8,11 +8,11 @@
 #   education3:
 #   education4:
 
-- name: Kartikey Sharma
-  photo: rock.jpg
-  info: Master student, Spring 2026 - "Client Selection in Federated Learning"
-  number_educ: 0
-  education1: 
+# - name: Kartikey Sharma
+#   photo: rock.jpg
+#   info: Master student, Spring 2026 - "Client Selection in Federated Learning"
+#   number_educ: 0
+#   education1: 
 
 - name: Teame Hailu Gebru
   photo: rock.jpg