Lui Sha
YOU?
Author Swipe
View article: Real-DRL: Teach and Learn in Reality
Real-DRL: Teach and Learn in Reality Open
This paper introduces the Real-DRL framework for safety-critical autonomous systems, enabling runtime learning of a deep reinforcement learning (DRL) agent to develop safe and high-performance action policies in real plants (i.e., real phy…
View article: VISAT: Benchmarking Adversarial and Distribution Shift Robustness in Traffic Sign Recognition with Visual Attributes
VISAT: Benchmarking Adversarial and Distribution Shift Robustness in Traffic Sign Recognition with Visual Attributes Open
We present VISAT, a novel open dataset and benchmarking suite for evaluating model robustness in the task of traffic sign recognition with the presence of visual attributes. Built upon the Mapillary Traffic Sign Dataset (MTSD), our dataset…
View article: GapTracer: Unraveling RPC Obfuscations in Provenance Graphs for Attack Source Tracing
GapTracer: Unraveling RPC Obfuscations in Provenance Graphs for Attack Source Tracing Open
View article: AISafetyLab: A Comprehensive Framework for AI Safety Evaluation and Improvement
AISafetyLab: A Comprehensive Framework for AI Safety Evaluation and Improvement Open
As AI models are increasingly deployed across diverse real-world scenarios, ensuring their safety remains a critical yet underexplored challenge. While substantial efforts have been made to evaluate and enhance AI safety, the lack of a sta…
View article: Be a Multitude to Itself: A Prompt Evolution Framework for Red Teaming
Be a Multitude to Itself: A Prompt Evolution Framework for Red Teaming Open
Large Language Models (LLMs) have gained increasing attention for their remarkable capacity, alongside concerns about safety arising from their potential to produce harmful content. Red teaming aims to find prompts that could elicit harmfu…
View article: How Far are LLMs from Being Our Digital Twins? A Benchmark for Persona-Based Behavior Chain Simulation
How Far are LLMs from Being Our Digital Twins? A Benchmark for Persona-Based Behavior Chain Simulation Open
Recently, LLMs have garnered increasing attention across academic disciplines for their potential as human digital twins, virtual proxies designed to replicate individuals and autonomously perform tasks such as decision-making, problem-sol…
View article: Neural Probabilistic Circuits: Enabling Compositional and Interpretable Predictions through Logical Reasoning
Neural Probabilistic Circuits: Enabling Compositional and Interpretable Predictions through Logical Reasoning Open
End-to-end deep neural networks have achieved remarkable success across various domains but are often criticized for their lack of interpretability. While post hoc explanation methods attempt to address this issue, they often fail to accur…
View article: How Far are LLMs from Being Our Digital Twins? A Benchmark for Persona-Based Behavior Chain Simulation
How Far are LLMs from Being Our Digital Twins? A Benchmark for Persona-Based Behavior Chain Simulation Open
View article: Resource Allocation Algorithm for Sensing Video Transmission Over Cell-Free Radio Access Networks
Resource Allocation Algorithm for Sensing Video Transmission Over Cell-Free Radio Access Networks Open
This paper proposes a communication-aware, game-theoretic bandwidth allocation strategy for multi-user visual data transmission over cell-free networks, addressing the joint optimization of resource efficiency and perceptual quality from t…
View article: Risk Assessment for Low-Voltage Distribution Network Based on Vmd-Lstm Ultra-Short-Term Load Forecasting Model
Risk Assessment for Low-Voltage Distribution Network Based on Vmd-Lstm Ultra-Short-Term Load Forecasting Model Open
View article: Plug-and-Play Training Framework for Preference Optimization
Plug-and-Play Training Framework for Preference Optimization Open
Recently, preference optimization methods such as DPO have significantly enhanced large language models (LLMs) in wide tasks including dialogue and question-answering. However, current methods fail to account for the varying difficulty lev…
View article: Physics-model-guided Worst-case Sampling for Safe Reinforcement Learning
Physics-model-guided Worst-case Sampling for Safe Reinforcement Learning Open
Real-world accidents in learning-enabled CPS frequently occur in challenging corner cases. During the training of deep reinforcement learning (DRL) policy, the standard setup for training conditions is either fixed at a single initial cond…
View article: Verification and Validation of a Vision-Based Landing System for Autonomous VTOL Air Taxis
Verification and Validation of a Vision-Based Landing System for Autonomous VTOL Air Taxis Open
Autonomous air taxis are poised to revolutionize urban mass transportation, however, ensuring their safety and reliability remains an open challenge. Validating autonomy solutions on air taxis in the real world presents complexities, risks…
View article: Bayesian Data Augmentation and Training for Perception DNN in Autonomous Aerial Vehicles
Bayesian Data Augmentation and Training for Perception DNN in Autonomous Aerial Vehicles Open
Learning-based solutions have enabled incredible capabilities for autonomous systems. Autonomous vehicles, both aerial and ground, rely on DNN for various integral tasks, including perception. The efficacy of supervised learning solutions …
View article: LLMs know their vulnerabilities: Uncover Safety Gaps through Natural Distribution Shifts
LLMs know their vulnerabilities: Uncover Safety Gaps through Natural Distribution Shifts Open
Safety concerns in large language models (LLMs) have gained significant attention due to their exposure to potentially harmful data during pre-training. In this paper, we identify a new safety vulnerability in LLMs: their susceptibility to…
View article: Simplex-enabled Safe Continual Learning Machine
Simplex-enabled Safe Continual Learning Machine Open
This paper proposes the SeC-Learning Machine: Simplex-enabled safe continual learning for safety-critical autonomous systems. The SeC-learning machine is built on Simplex logic (that is, ``using simplicity to control complexity'') and phys…
View article: Towards a Unified View of Preference Learning for Large Language Models: A Survey
Towards a Unified View of Preference Learning for Large Language Models: A Survey Open
Large Language Models (LLMs) exhibit remarkably powerful capabilities. One of the crucial factors to achieve success is aligning the LLM's output with human preferences. This alignment process often requires only a small amount of data to …
View article: HSF: Defending against Jailbreak Attacks with Hidden State Filtering
HSF: Defending against Jailbreak Attacks with Hidden State Filtering Open
With the growing deployment of LLMs in daily applications like chatbots and content generation, efforts to ensure outputs align with human values and avoid harmful content have intensified. However, increasingly sophisticated jailbreak att…
View article: Perception simplex: Verifiable collision avoidance in autonomous vehicles amidst obstacle detection faults
Perception simplex: Verifiable collision avoidance in autonomous vehicles amidst obstacle detection faults Open
Advances in deep learning have revolutionized cyber‐physical applications, including the development of autonomous vehicles. However, real‐world collisions involving autonomous control of vehicles have raised significant safety concerns re…
View article: Taming Algorithmic Priority Inversion in Mission-Critical Perception Pipelines
Taming Algorithmic Priority Inversion in Mission-Critical Perception Pipelines Open
The paper discusses algorithmic priority inversion in mission-critical machine inference pipelines used in modern neural-network-based perception subsystems and describes a solution to mitigate its effect. In general, priority inversion oc…
View article: Synergistic Perception and Control Simplex for Verifiable Safe Vertical Landing
Synergistic Perception and Control Simplex for Verifiable Safe Vertical Landing Open
Perception, Planning, and Control form the essential components of autonomy\nin advanced air mobility. This work advances the holistic integration of these\ncomponents to enhance the performance and robustness of the complete\ncyber-physic…
View article: Backup Plan Constrained Model Predictive Control with Guaranteed Stability
Backup Plan Constrained Model Predictive Control with Guaranteed Stability Open
This article proposes and evaluates a new safety concept called backup plan safety for path planning of autonomous vehicles under mission uncertainty using model predictive control (MPC). Backup plan safety is defined as the ability to com…
View article: Physics-Regulated Deep Reinforcement Learning: Invariant Embeddings
Physics-Regulated Deep Reinforcement Learning: Invariant Embeddings Open
This paper proposes the Phy-DRL: a physics-regulated deep reinforcement learning (DRL) framework for safety-critical autonomous systems. The Phy-DRL has three distinguished invariant-embedding designs: i) residual action policy (i.e., inte…
View article: Physical Deep Reinforcement Learning Towards Safety Guarantee
Physical Deep Reinforcement Learning Towards Safety Guarantee Open
Deep reinforcement learning (DRL) has achieved tremendous success in many complex decision-making tasks of autonomous systems with high-dimensional state and/or action spaces. However, the safety and stability still remain major concerns t…
View article: Phy-Taylor: Physics-Model-Based Deep Neural Networks
Phy-Taylor: Physics-Model-Based Deep Neural Networks Open
Purely data-driven deep neural networks (DNNs) applied to physical engineering systems can infer relations that violate physics laws, thus leading to unexpected consequences. To address this challenge, we propose a physics-model-based DNN …
View article: Perception Simplex: Verifiable Collision Avoidance in Autonomous Vehicles Amidst Obstacle Detection Faults
Perception Simplex: Verifiable Collision Avoidance in Autonomous Vehicles Amidst Obstacle Detection Faults Open
Advances in deep learning have revolutionized cyber-physical applications, including the development of Autonomous Vehicles. However, real-world collisions involving autonomous control of vehicles have raised significant safety concerns re…
View article: Verifiable Obstacle Detection
Verifiable Obstacle Detection Open
Perception of obstacles remains a critical safety concern for autonomous vehicles. Real-world collisions have shown that the autonomy faults leading to fatal collisions originate from obstacle existence detection. Open source autonomous dr…
View article: LiDAR Cluster First and Camera Inference Later: A New Perspective Towards Autonomous Driving
LiDAR Cluster First and Camera Inference Later: A New Perspective Towards Autonomous Driving Open
Object detection in state-of-the-art Autonomous Vehicles (AV) framework relies heavily on deep neural networks. Typically, these networks perform object detection uniformly on the entire camera LiDAR frames. However, this uniformity jeopar…
View article: SchedGuard: Protecting against Schedule Leaks Using Linux Containers
SchedGuard: Protecting against Schedule Leaks Using Linux Containers Open
Real-time systems have recently been shown to be vulnerable to timing inference attacks, mainly due to their predictable behavioral patterns. Existing solutions such as schedule randomization lack the ability to protect against such attack…
View article: Backup Plan Constrained Model Predictive Control
Backup Plan Constrained Model Predictive Control Open
This article proposes a new safety concept: backup plan safety. The backup plan safety is defined as the ability to complete one of the alternative missions in the case of primary mission abortion. To incorporate this new safety concept in…