HADES: Hardware Accelerated Decoding for Efficient Speculation in Large Language Models

Large Language Models (LLMs) have revolutionized natural language processing by understanding and generating human-like text. However, the increasing demand for more sophisticated LLMs presents significant computational challenges due to their scale and complexity. This paper introduces Hardware Accelerated Decoding (HADES), a novel approach to enhance the performance and energy efficiency of LLMs. We address the design of an LLM accelerator with hardware-level speculative decoding support, a concept not previously explored in existing literature. Our work demonstrates how speculative decoding can significantly improve the efficiency of LLM operations, paving the way for more advanced and practical applications of these models.

Scam Detection for Ethereum Smart Contracts: Leveraging Graph Representation Learning for Secure Blockchain

The detection of scams within Ethereum smart contracts is a critical challenge due to their increasing exploitation for fraudulent activities, leading to significant financial and reputational damages. Existing detection methods often rely on contract code analysis or manually extracted features, which suffer from scalability and adaptability limitations. In this study, we introduce an innovative method that leverages graph representation learning to examine transaction patterns and identify fraudulent contracts. By transforming Ethereum transaction data into graph structures and employing advanced machine learning models, we achieve robust classification performance. Our method addresses label imbalance through SMOTE-ENN techniques and evaluates models like Multi-Layer Perceptron (MLP) and Graph Convolutional Networks (GCN). Experimental results indicate that the MLP model surpasses the GCN in this context, with real-world evaluations aligning closely with domain-specific analyses. This study provides a scalable and effective solution for enhancing trust and security in the Ethereum ecosystem.