6 papers
TOGGLE: Temporal Logic-Guided Large Language Model Compression for Edge
Khurram Khalil, Khaza Anuarul Hoque
Large Language Models (LLMs) deliver exceptional performance across natural language tasks but demand substantial computational resources, limiting their deployment on resource-con…
FlipLLM: Efficient Bit-Flip Attacks on Multimodal LLMs using Reinforcement Learning
Khurram Khalil, Khaza Anuarul Hoque
Generative Artificial Intelligence models, such as Large Language Models (LLMs) and Large Vision Models (VLMs), exhibit state-of-the-art performance but remain vulnerable to hardwa…
RIFT: A Scalable Methodology for LLM Accelerator Fault Assessment using Reinforcement Learning
Khurram Khalil, Muhammad Mahad Khaliq, Khaza Anuarul Hoque
The massive scale of modern AI accelerators presents critical challenges to traditional fault assessment methodologies, which face prohibitive computational costs and provide poor…
ApproXAI: Energy-Efficient Hardware Acceleration of Explainable AI using Approximate Computing
Ayesha Siddique, Khurram Khalil, Khaza Anuarul Hoque
Explainable artificial intelligence (XAI) enhances AI system transparency by framing interpretability as an optimization problem. However, this approach often necessitates numerous…
EPSILON: Adaptive Fault Mitigation in Approximate Deep Neural Network using Statistical Signatures
Khurram Khalil, Khaza Anuarul Hoque
The increasing adoption of approximate computing in deep neural network accelerators (AxDNNs) promises significant energy efficiency gains. However, permanent faults in AxDNNs can…
Explainable AI-Guided Efficient Approximate DNN Generation for Multi-Pod Systolic Arrays
Ayesha Siddique, Khurram Khalil, Khaza Anuarul Hoque
Approximate deep neural networks (AxDNNs) are promising for enhancing energy efficiency in real-world devices. One of the key contributors behind this enhanced energy efficiency in…