Fakhri Karray

AAAI Conference 2025 Conference Paper

Internal Activation Revision: Safeguarding Vision Language Models Without Parameter Update

• Qing Li
• Jiahui Geng
• Derui Zhu
• Zongxiong Chen
• Kun Song
• Lei Ma
• Fakhri Karray

Warning: This paper contains offensive content that may disturb some readers. Vision-language models (VLMs) demonstrate strong multimodal capabilities but have been found to be more susceptible to generating harmful content compared to their backbone large language models (LLMs). Our investigation reveals that the integration of images significantly shifts the model's internal activations during the forward pass, diverging from those triggered by textual input. Moreover, the safety alignments of LLMs embedded within VLMs are not sufficiently robust to handle the activations discrepancies, making the models vulnerable to even the simplest jailbreaking attacks. To address this issue, we propose an internal activation revision approach that efficiently revises activations during generation, steering the model toward safer outputs. Our framework incorporates revisions at both the layer and head levels, offering control over the model's generation at varying levels of granularity. In addition, we explore three strategies for constructing positive and negative samples and two approaches for extracting revision vectors, resulting in different variants of our method. Comprehensive experiments demonstrate that the internal activation revision method significantly improves the safety of widely used VLMs, reducing attack success rates by an average of 48.94%, 34.34%, 43.92%, and 52.98% on SafeBench, Safe-Unsafe, Unsafe, and MM-SafetyBench, respectively, while minimally impacting model helpfulness.

AAAI Conference 2024 Conference Paper

Robustly Train Normalizing Flows via KL Divergence Regularization

• Kun Song
• Ruben Solozabal
• Hao Li
• Martin Takáč
• Lu Ren
• Fakhri Karray

In this paper, we find that the training of Normalizing Flows (NFs) are easily affected by the outliers and a small number (or high dimensionality) of training samples. To solve this problem, we propose a Kullback–Leibler (KL) divergence regularization on the Jacobian matrix of NFs. We prove that such regularization is equivalent to adding a set of samples whose covariance matrix is the identity matrix to the training set. Thus, it reduces the negative influence of the outliers and the small sample number on the estimation of the covariance matrix, simultaneously. Therefore, our regularization makes the training of NFs robust. Ultimately, we evaluate the performance of NFs on out-of-distribution (OoD) detection tasks. The excellent results obtained demonstrate the effectiveness of the proposed regularization term. For example, with the help of the proposed regularization, the OoD detection score increases at most 30% compared with the one without the regularization.

EAAI Journal 2024 Journal Article

Yield estimation and health assessment of temperate fruits: A modular framework

• Jamil Ahmad
• Wail Gueaieb
• Abdulmotaleb El Saddik
• Giulia De Masi
• Fakhri Karray

IJCAI Conference 2016 Conference Paper

EBEK: Exemplar-Based Kernel Preserving Embedding

• Ahmed Elbagoury
• Rania Ibrahim
• Mohamed S. Kamel
• Fakhri Karray

With the rapid increase in the available data, it becomes computationally harder to extract useful information. Thus, several techniques like PCA were proposed to embed high-dimensional data into low-dimensional latent space. However, these techniques don't take the data relations into account. This motivated the development of other techniques like MDS and LLE which preserve the relations between the data instances. Nonetheless, all these techniques still use latent features, which are difficult for data analysts to understand and grasp the information encoded in them. In this work, a new embedding technique is proposed to mitigate the previous problems by projecting the data to a space described by few points (i. e, exemplars) which preserves the relations between the data points. The proposed method Exemplar-based Kernel Preserving (EBEK) embedding is shown theoretically to achieve the lowest reconstruction error of the kernel matrix. Using EBEK in approximate nearest neighbor task shows its ability to outperform related work by up to 60% in the recall while maintaining a good running time. In addition, our interpretability experiments show that EBEK's selected basis are more understandable than the latent basis in images datasets.

ICRA Conference 2011 Conference Paper

Integrating visual exploration and visual search in robotic visual attention: The role of human-robot interaction

• Momotaz Begum
• Fakhri Karray

A common characteristics of the computational models of visual attention is they execute the two modes of visual attention (visual exploration and visual search) separately. This makes a visual attention model unsuitable for real-world robotic applications. This paper focuses on integrating visual exploration and visual search in a common framework of visual attention and the challenges resulting from such integration. It proposes a visual attention-oriented speech-based human robot interaction framework which helps a robot to switch back-and-forth between the two modes of visual attention. A set of experiments are presented to demonstrate the performance of the proposed framework.

IROS Conference 2008 Conference Paper

Object- and space-based visual attention: An integrated framework for autonomous robots

• Momotaz Begum
• George K. I. Mann
• Raymond G. Gosine
• Fakhri Karray

This paper argues that the object- and space-based modes of visual attention can be naturally integrated in a common mathematical framework. In an earlier work [1] we have proposed a mathematical model of visual attention for robotic system exploiting the knowledge of visual attention mechanism of the primates. This paper investigates on the validity of the proposed model for robotic systems through experimentation on a real robot. The paper sheds light on a number of real world issues involved with the design of visual attention system for physically embodied robots and explains how the proposed Bayesian model of visual attention addresses these issues. The object- and space-based modes of visual attention are naturally integrated in the model and is reflected in the sequential Monte Carlo implementation of the model on a real robot.

IROS Conference 2003 Conference Paper

Real world implementation of fuzzy anti-swing control for behavior-based intelligent crane system

• Jiaming Wang
• Hao Li
• Fakhri Karray
• Otman A. Basir

There exist several industrial applications for large crane systems. Most of them experience serious problems with load swing. This paper presents a fuzzy based control scheme to minimize load swing for crane systems while maintaining continuous payload transportation. The control system of the crane is built using behavior-based approaches. In the control system developed, each module generates behaviors, and improvement in the performance of the system proceeds by adding new modules to the system. In order to develop the anti-swing module, fuzzy logic controller is applied using information extracted from potentiometers. The fuzzy controller provides a mechanism for dealing with imprecise sensor data. The anti-swing behaviors are successfully implemented by formulating a set of fuzzy rules. The performance of the developed system is illustrated by both simulations and experiments. The simulation and experimental results of the system show that the system remains stable under several operating situations.

ICRA Conference 2002 Conference Paper

A Hybrid Adaptive Fuzzy Approach for the Control of Cooperative Manipulators

• Wail Gueaieb
• Fakhri Karray
• Salah Al-Sharhan
• Otman A. Basir

We examine in this article the complex problem of simultaneous position and internal force control in multiple cooperative manipulator systems. This is done in the presence of unwanted parametric and modeling uncertainties as well as external disturbances. A decentralized adaptive hybrid intelligent control scheme is proposed here. The controller makes use of a multi-input multi-output fuzzy logic engine and a systematic online adaptation mechanism. Unlike conventional adaptive controllers, the proposed one does not require a precise model of the system's dynamics. The performance of the proposed controller is compared to that of a well known conventional adaptive controller.

IROS Conference 1998 Conference Paper

A software-based procedure for robotic end effector error correction

• Medhat Moussa
• Martin Hill
• James Fernandes
• Fakhri Karray

Presents a procedure for modeling and eliminating an end effector configuration error as a result of a faulty joint or a damaged link. This procedure provides an inexpensive software alternative to hardware replacement. A neural network model was developed and tested an a 6 DOF PUMA robot. The network approximates the error based on data obtained through observing the robot while executing a set of MOVE commands. The results show that, regardless of the error source, the robot's accuracy could be highly improved even when a small number of data points are used.

IROS Conference 1998 Conference Paper

Robust joint trajectory tracking of a flexible lightweight manipulator

• Fakhri Karray
• S. Tafazolli
• Wail Gueaieb

A robust control design for high performance joint trajectory tracking of a flexible lightweight manipulator system is proposed. The design is based on a combined controller-observer scheme involving the sliding manifold approach and the optimal interpolation technique. This controller provides the designer with an enhanced joint tracking performance when the system is subject to parametric variations due to structural disturbances caused by link flexibility and load uncertainties. The parametric variations are handled by sliding control and the estimation of the nonlinearly excited elastic dynamics by an optimal interpolator of the structure's dynamic responses. The design procedure is progressive, i. e. , we start with a basic controller and then modify it in order to improve the performance. Closed loop simulations with the various designed controllers are used to validate the analytical results and to help choosing the most suitable one.