Publications by Milos Vasic | EPFL Graph Search

Beyond fine-tuning: LoRA modules boost near-OOD detection and LLM security

Pierre Vandergheynst, Milos Vasic, Francesco Craighero, Renata Khasanova

Under resource constraints, LLMs are usually fine- tuned with additional knowledge using Parameter Efficient Fine-Tuning (PEFT), using Low-Rank Adaptation (LoRA) modules. In fact, LoRA injects a new set of small trainable matrices to adapt an LLM to a new ...

2024

Beyond fine-tuning: LoRA modules boost near-OOD detection and LLM security

Pierre Vandergheynst, Milos Vasic, Francesco Craighero, Renata Khasanova

Under resource constraints, LLMs are usually fine-tuned with additional knowledge using Parameter Efficient Fine-Tuning (PEFT), using Low-Rank Adaptation (LoRA) modules. In fact, LoRA injects a new set of small trainable matrices to adapt an LLM to a new t ...

2024

Cooperative Perception Algorithms for Networked Intelligent Vehicles

Milos Vasic

The degree of intelligence built-in in today's vehicles in constantly on the rise. The vehicles are being equipped with sensors, with the goal to estimate the state of the vehicle and the environment surrounding it. Intelligent algorithms that process the ...

EPFL2017

Cooperative Multiple Dynamic Object Tracking on Moving Vehicles Based on Sequential Monte Carlo Probability Hypothesis Density Filter

Alcherio Martinoli, Milos Vasic, Jonathan Gan

This paper proposes a generalized method for tracking of multiple objects from moving, cooperative vehicles -- bringing together an Unscented Kalman Filter for vehicle localization and extending a Sequential Monte Carlo Probability Hypothesis Density filte ...

Ieee2016

An Overtaking Decision Algorithm for Networked Intelligent Vehicles Based on Cooperative Perception

Alcherio Martinoli, Milos Vasic, Inaki Navarro Oiza, Gael Lederrey

This paper presents an overtaking decision algorithm for networked intelligent vehicles. The algorithm is based on a cooperative tracking and sensor fusion algorithm that we previously developed. The ego vehicle is equipped with lane keeping and lane chang ...

Ieee2016

A System Implementation and Evaluation of a Cooperative Fusion and Tracking Algorithm based on a Gaussian Mixture PHD Filter

Alcherio Martinoli, Milos Vasic, David Mansolino

This paper focuses on a real system implementation, analysis, and evaluation of a cooperative sensor fusion algorithm based on a Gaussian Mixture Probability Hypothesis Density (GM-PHD) filter, using simulated and real vehicles endowed with automotive-grad ...

Ieee2016

Distributed Graph-Based Control of Convoys of Heterogeneous Vehicles using Curvilinear Road Coordinates

Alcherio Martinoli, Milos Vasic, Inaki Navarro Oiza, Florian Zimmermann

This paper investigates the problem of controlling a heterogeneous group of vehicles with the aim of forming multi-lane convoys. We use a distributed, graph-based control law, implemented in a longitudinal coordinate system parallel to the road. Each vehic ...

Ieee2016

Distributed Graph-based Convoy Control for Networked Intelligent Vehicles

Alcherio Martinoli, Ali Marjovi, Milos Vasic, Joseph Chadi Benoit Lemaitre

This paper presents an approach for formation control of multi-lane vehicular convoys in highways. We extend a Laplacian graph-based, distributed control law such that networked intelligent vehicles can join or leave the formation dynamically without jeopa ...

IEEE2015

A Collaborative Sensor Fusion Algorithm for Multi-Object Tracking Using a Gaussian Mixture Probability Hypothesis Density Filter

Alcherio Martinoli, Milos Vasic

This paper presents a method for collaborative tracking of multiple vehicles that extends a Gaussian Mixture Probability Hypothesis Density (GM-PHD) filter with a collaborative fusion algorithm. Measurements are preprocessed in a detect-before-track fashio ...

2015

Safety Issues in Human-Robot Interactions

Aude Billard, Milos Vasic

Safety is an important consideration in human-robot interactions. Robots can perform powerful movements that can cause hazards to humans surrounding them. To prevent accidents, it is important to identify sources of potential harm, to determine which of th ...

2013