Large language models (LLMs) have achieved remarkable success across various domains, but training them centrally requires massive data collection and annotation efforts, making it costly for individual parties. Federated learning (FL) has emerged as a promising solution, enabling collaborative training of LLMs on decentralized data while preserving privacy (FedLLM). Although frameworks like OpenFedLLM, FederatedScope-LLM, and…
Human-computer interaction (HCI) focuses on designing and using computer technology, particularly the interfaces between people (users) and computers. Researchers in this field observe how humans interact with computers & design technologies that let humans interact with computers in novel ways. HCI encompasses various areas, such as user experience design, ergonomics, and cognitive psychology, aiming to…
Retrieval Augmented Generation (RAG) is a method that enhances the capabilities of Large Language Models (LLMs) by integrating a document retrieval system. This integration allows LLMs to fetch relevant information from external sources, thereby improving the accuracy and relevance of the responses generated. This approach addresses the limitations of traditional LLMs, such as the need…
Large language models (LLMs) have the potential to lead users to make poor decisions, especially when these models provide incorrect information with high confidence, which is called hallucination. This confident misinformation has the potential to be very dangerous since it might persuade people to act based on erroneous assumptions, which could have negative consequences. A…
The training of Large Language Models (LLMs) like GPT-3 and Llama on a large scale faces significant inefficiencies due to hardware failures and network congestion. These issues lead to substantial GPU resource waste and extended training durations. Specifically, hardware malfunctions cause interruptions in training, and network congestions force GPUs to wait for parameter synchronization, further…
Apple made a significant announcement, strongly advocating for on-device AI through its newly introduced Apple Intelligence. This innovative approach emphasizes the integration of a ~3 billion parameter language model (LLM) on devices like Mac, iPhone, and iPad, leveraging fine-tuned LoRA adapters to perform specialized tasks. This model claims to outperform larger models, such as the…
The study of evolution by natural selection at the molecular level has advanced significantly with the advent of genomic technologies. Traditionally, researchers have focused on observable traits like flowering time or growth. However, gene expression provides an intermediate phenotype that connects genomic data to these macroscopic traits, offering a deeper understanding of selection pressures. In…
The vulnerability of AI systems, particularly large language models (LLMs) and multimodal models, to adversarial attacks can lead to harmful outputs. These models are designed to assist and provide helpful responses, but adversaries can manipulate them to produce undesirable or even dangerous outputs. The attacks exploit inherent weaknesses in the models, raising concerns about their…
Large Language Models (LLMs) have become an essential tool in artificial intelligence, primarily due to their generative capabilities and ability to follow user instructions effectively. These features make LLMs ideal for developing chatbots that interact seamlessly with users. However, the text-based nature of LLMs has limited chatbots to text-only interactions. In recent years, significant efforts…
China’s Kuaishou Technology has created a buzz in text-to-video generation with its groundbreaking Kling AI video model. This advanced text-to-video generation model is revolutionizing the industry by producing highly realistic videos from simple text prompts, setting a new benchmark in AI-driven video creation. High-Quality Video Generation Kling AI stands out for its ability to create…