Search AI Products and News

A recent study from Microsoft and Carnegie Mellon University indicates that excessive reliance on artificial intelligence (AI) tools may be eroding people's critical thinking abilities. The research team surveyed 319 knowledge workers and collected 936 real-world examples of generative AI usage across various industries, including IT, design, administration, and finance. The study analyzed six dimensions of critical thinking: knowledge, understanding, application, analysis, synthesis, and evaluation. The findings reveal that people's thinking patterns in problem-solving change significantly after using AI tools.

Recently, researchers from Microsoft Research and Carnegie Mellon University released a new study revealing potential issues for knowledge workers using generative AI (such as Copilot and ChatGPT). The research team surveyed 319 knowledge workers who use generative AI weekly to explore their application of critical thinking while using these tools. Image source note: Image generated by AI, courtesy of image licensing provider Midjourney. The research findings indicate that those...

The T-MAC (Table-Lookup-based MAC) technology aims to address the memory and computational constraints of deploying large language models (LLMs) on edge devices. By quantizing model weights into low-bit representations, T-MAC replaces traditional multiplication operations with look-up tables (LUT), significantly enhancing running efficiency on CPUs. This method greatly reduces the memory required for calculations, enabling billion-parameter LLMs to run efficiently on resource-constrained devices for smart upgrades. Compared to existing implementations, T-

Recently, a research team from Microsoft Research and Beihang University has jointly introduced a novel framework called E5-V, aimed at providing a more efficient solution for multi-modal embeddings. With the continuous advancement of artificial intelligence, multi-modal large language models (MLLMs) have become a focal point of research, as they are capable of understanding both textual and visual information simultaneously, thereby better handling complex data relationships. However, effective

Microsoft researchers have recently unveiled an innovative research called SpreadsheetLLM, aimed at addressing the challenges faced by large language models (LLMs) when parsing electronic spreadsheets. According to a paper published on Arxiv on July 12th, SpreadsheetLLM enables LLMs to "read" spreadsheet content through an encoding framework. This research is expected to significantly enhance the efficiency of data management and analysis in spreadsheets, allowing users to ask questions of AI in

Researcher Ellie Pavlick states that training mini language models is akin to sequencing the fruit fly genome rather than the human genome. Microsoft researchers have achieved significant results using children's stories to train mini language models. Compared to large language models, mini language models train faster and have mechanisms that are easier to understand. This approach helps in training larger models and also aids in analyzing their behavior. Research findings suggest that after being trained with children's stories, mini language models can narrate coherent and grammatically correct stories.

{Point 1: Microsoft Research Team believes GPT-4 can accelerate medical processes and improve efficiency. Point 2: GPT-4 can handle large-scale unstructured patient data and demonstrates impressive capabilities in medical tasks. Point 3: Microsoft is developing the LLaVA-Med model to integrate medical data and research more closely, accelerating healthcare and research.}

{1: Microsoft Research Asia has proposed a method for industrial control using pre-trained large language models (LLM). 2: The research utilizes the GPT-4 model to control air conditioning systems, achieving results superior to traditional reinforcement learning methods with only a few example samples. 3: The goal of this research is to explore the potential of directly using pre-trained LLMs for industrial control tasks and to design methods for addressing industrial control tasks with lower technical debt.}