In the news | PIM-DL: Expanding the Applicability of Commodity DRAM-PIMs for Deep Learning via Algorithm-System Co-Optimization
Microsoft at ASPLOS 2024: Advancing hardware and software for high-scale, secure, and efficient modern applications
PIM-DL is the first deep learning framework specifically designed for off-the-shelf processing-in-memory (PIM) systems, capable of offloading most computations in neural networks. Its goal is to surmount the computational limitations of PIM hardware by replacing traditional compute-heavy matrix multiplication operations…
In the news | Devex Book Club
Juan M. Lavista Ferres on the power of AI for good
In this episode of the Devex Book Club podcast, Juan M. Lavista Ferres discusses the many ways that AI can have a positive impact on the world, from identifying health trends to tracking beluga whales.
Ideas: Exploring AI frontiers with Rafah Hosn
| Rafah Hosn and Gretchen Huizinga
Energized by disruption, partner group product manager Rafah Hosn is helping to drive scientific advancement in AI for Microsoft. She talks about the mindset needed to work at the frontiers of AI and how the research-to-product pipeline is changing in…
统一化数据库：为大语言模型垂域应用奠定基础
编者按：检索增强生成（RAG）技术因在减少生成幻觉和虚构信息方面的显著效果，以及对知识及时更新能力的改善，正逐渐成为大语言模型系统的主流架构之一。随着 RAG 技术的广泛应用，其核心组件——向量数据库，也开始受到越来越多的关注，成为大模型中不可或缺的外挂知识库。 然而，向量数据库与传统关系型数据库有着显著区别，这给数据的统一管理、查询和更新带来了诸多不便。为此，微软亚洲研究院开发了 VBase 复...
编者按：欢迎阅读“科研上新”栏目！“科研上新”汇聚了微软亚洲研究院最新的创新成果与科研动态。在这里，你可以快速浏览研究院的亮点资讯，保持对前沿领域的敏锐嗅觉，同时也能找到先进实用的开源工具。 论文链接：https://www.microsoft.com/en-us/research/publication/an-empirical-study-on-low-gpu-utilization-of-d...
SAMMO: A general-purpose framework for prompt optimization
| Tobias Schnabel and Jennifer Neville
SAMMO optimizes prompts for LLMs by leveraging their structure to guide optimization. This minimizes the time and effort needed to find performant prompts on a variety of tasks.
Research Focus: Week of April 15, 2024
In this issue: New research on appropriate reliance on generative AI; Power management opportunities for LLMs in the cloud; LLMLingua-2 improves task-agnostic prompt compression; Enhancing COMET to embrace under-resourced African languages:
作者：系统与网络组 编者按：大模型的飞速发展给人们的生活带来了前所未有的便利。我们是否能够设想利用大模型的潜力，快速扫描整部百科全书、解析繁琐复杂的法律条款，甚至对文章进行精准引用呢？在未来，这些将统统可以实现。然而，目前传统的大模型的上下文窗口限制与昂贵的微调成本使得它们难以处理超长文本，从而限制了其应用潜力。为解决这一问题，微软亚洲研究院的研究员们提出了 LongRoPE。通过精细化非均匀位置...
Microsoft at NSDI 2024: Discoveries and implementations in networked systems
| Ranveer Chandra
Microsoft at NDSI 2024: Discoveries and implementations in networked systems Topics range from 5G, space, datacenters, and wide-area networking to applications in artificial intelligence, security, video conferencing, and gaming. Learn more about the discoveries and advances we're making with networked…