IBM acquires Databand to improve machine learning models

2022-07-14 16:48:08
关注


IBM has acquired Israeli data observability start-up Databand for an undisclosed sum in a deal that it says will help to improve machine learning models by catching bad data early.


Databand is the fifth acquisition for IBM so far this year, including other AI and observability plays. (Photo by MIGUEL MEDINA/AFP via Getty Images)


What is Databand?

Founded in 2018, Databand produces software that is designed to help organisations improve the quality of their data. It says it does this by fixing errors, pipeline failures and poor-quality elements before they reach machine learning models.

It is the fifth acquisition for IBM in the AI and hybrid cloud space this year. These include the takeover of Envizi in January, a data and analytics software provider for environmental performance management and Microsoft Azure consultancy firm Neudesic in February.

                                                                                       

window.googletag = window.googletag || {cmd: []}; googletag.cmd.push(function() { googletag.defineSlot('/5269235/Tech_Monitor/Tech_Monitor_LB', [[320,50]], 'div-gpt-ad-8784900-1') .addService(googletag.pubads()); googletag.defineSlot('/5269235/Tech_Monitor/Tech_Monitor_BB', [[728,90], [970,250]], 'div-gpt-ad-8784900-2') .addService(googletag.pubads()); googletag.defineSlot('/5269235/Tech_Monitor/Tech_Monitor_Mobile_MPU', [[300,250]], 'div-gpt-ad-8784900-3') .addService(googletag.pubads()); googletag.pubads().enableSingleRequest(); googletag.enableServices(); }); googletag.cmd.push(function() { googletag.display('div-gpt-ad-8784900-3'); });


IBM says its goal with the acquisitions is to strengthen its software portfolio across data, AI and automation in an attempt to address the “full spectrum of observability”.


Poor data costs organisations about $12.9m every year, according to research from Gartner, and could lead to “poor decision making” in the long term if bad data feeds into predictive models used to drive the direction of a company or product, the analyst company says.

Data observability, like the platform developed by Databand, deals with understanding why a dataset or pipeline isn’t doing what you expect.

“You can’t protect what you can’t see, and when the data platform is ineffective, everyone is impacted –including customers,” said Josh Benamram, co-founder and CEO of Databand. “Joining IBM will help us scale our software and significantly accelerate our ability to meet the evolving needs of enterprise clients.”


Content from our partners


How clinical trials infrastructure is undergoing digital transformation


Webinar – Top 3 Ways to Build Security into DevOps


Tech sector is making progress on diversity, but advances must accelerate


IBM to integrate Databand with other recent acquisitions

Another relatively recent IBM acquisition, Instana, which was purchased in 2020, gave IBM software that provides businesses with the ability to manage the performance of complex cloud-native applications regardless of where they reside.


“By using Databand.ai with IBM Observability by Instana APM and IBM Watson Studio, IBM is well-positioned to address the full spectrum of observability across IT operations,” an IBM statement said.

This will include using Databand to alert engineers of missing data, then using Instana to explain where the missing data came from and produce a more complete view of the entire application infrastructure and data platform, which IBM says will “help organisations prevent lost revenue and reputation”.


   Data, insights and analysis delivered to you
   View all newsletters
   By The Tech Monitor team
   Sign up to our newsletters
   Sign up here

IBM says it will provide Databand with the resources to expand its observability capabilities and offer a broader range of options – across commercial and open source platforms.

“Our clients are data-driven enterprises who rely on high quality, trustworthy data to power their mission-critical processes. When they don’t have access to the data they need in any given moment, their business can grind to a halt,” said Daniel Hernandez, general manager for data and AI at IBM.

“With the addition of Databand.ai, IBM offers the most comprehensive set of observability capabilities for IT across applications, data and machine learning, and is continuing to provide our clients and partners with the technology they need to deliver trustworthy data and AI at scale.”

Read more: Post-Covid-19 flexibility driving IBM’s investments – UK CEO

参考译文
IBM收购Databand是为了改进机器学习模型



IBM 以未公开的价格收购了以色列数据可观测性初创公司 Databand,表示这笔交易将有助于通过及早识别劣质数据来改进机器学习模型。Databand 是 IBM 今年收购的第五家公司,其中包括其他几家人工智能和可观测性领域的公司。(图片来源:MIGUEL MEDINA/AFP via Getty Images)

什么是 Databand?Databand 成立于 2018 年,其软件旨在帮助组织提高数据质量。该公司表示,其软件通过在数据到达机器学习模型之前修复错误、流水线故障和低质量元素来实现这一点。Databand 是 IBM 今年在人工智能和混合云领域收购的第五家公司,其中包括今年 1 月收购的环境绩效管理数据和分析软件供应商 Envizi,以及 2 月收购的微软 Azure 咨询公司 Neudesic。

IBM 表示,其收购目标是通过数据、人工智能和自动化领域来增强其软件组合,以应对“完整的可观测性光谱”。

相关人工智能与自动化内容:  
- Meta 的新 Sphere AI 工具将检查维基百科条目的准确性  
- 人工智能与自动化:沃达丰与谷歌云签署“高风险”人工智能协议  
- 情绪识别技术大多效果不佳,为何企业仍对其投资?  
- 英国将保护商业人工智能数据挖掘免受版权主张的影响  

根据 Gartner 的研究,劣质数据每年给企业带来大约 1290 万美元的损失,如果劣质数据被输入用于引导公司或产品方向的预测模型,可能会导致“决策失误”。

数据可观测性,就像 Databand 开发的平台一样,涉及理解数据集或流水线为何没有按预期运作。Databand 联合创始人兼首席执行官 Josh Benamram 表示:“你无法保护你看不到的东西。当数据平台失效时,所有人都会受到影响,包括客户。”“加入 IBM 将帮助我们扩大软件规模,并显著加快满足企业客户不断变化的需求的能力。”

合作伙伴内容:  
- 临床试验基础设施正在经历数字化转型  
- 网络研讨会:构建 DevOps 安全性的三种最佳方法  
- 科技行业在多样性方面取得进展,但提升步伐必须加快  
- IBM 将整合 Databand 与其他最近的收购  

另一项相对较新的 IBM 收购是 Instana,该公司于 2020 年被收购,为 IBM 提供了软件,使企业能够管理复杂云原生应用程序的性能,无论这些应用程序部署在哪里。

IBM 在一份声明中表示:“通过将 Databand.ai 与 IBM Observability by Instana APM 和 IBM Watson Studio 结合使用,IBM 做到了在 IT 操作中全面实现可观测性的最佳位置。”这将包括使用 Databand 向工程师发出数据缺失的警报,然后通过 Instana 解释数据缺失的来源,并生成整个应用程序基础设施和数据平台的更完整视图。IBM 表示,这将“帮助企业避免收入和声誉的损失”。

我们将数据、见解和分析直接送达您  
查看所有通讯  
由 Tech Monitor 团队提供  
注册我们的通讯  
在此注册  

IBM 表示,它将为 Databand 提供资源以扩展其可观测性功能,并提供更广泛的选项——涵盖商业和开源平台。IBM 的数据与 AI 业务总经理 Daniel Hernandez 表示:“我们的客户是数据驱动型企业,他们依赖高质量、可信的数据来推动其关键业务流程。当他们在某一时刻无法访问所需数据时,其业务可能就会陷入停滞。”“通过 Databand.ai 的加入,IBM 将为 IT 领域的应用程序、数据和机器学习提供最全面的可观测性功能,并继续为我们的客户和合作伙伴提供他们需要的技术,以实现大规模可信数据和人工智能。”  

更多阅读:后疫情时代的灵活性正在驱动 IBM 的投资——英国 CEO  

本文涉及的主题:IBM

您觉得本篇内容如何
评分

相关产品

宇鸿敏芯 YHD75-200-5 多维力传感器

适用范围: 应用于工业机器人,实现精密打磨、精密装配、示教学习等功能,还适用于协作机器人、医疗机器人、空间机器人等。

LuminWave 洛微科技 LW-FS8864-SME ToF飞行时间光学传感器

LW-FS8864系列产品是高精度近场4线/8线MicroLiDAR传感器模组,具备最远可达4m的多像素精确测距能力,同时该产品在室外使用拥有良好的抗日光、抗环境光的优势,可配合MCU内嵌的深度学习等AI该产品提供最高可达8线的垂直分辨率,具有良好的抗日光、抗环境光的能力,可应用于户外场景,并针对机器人和无人车近场避障方案进行了优化。

ST LSM6DSOXTR 加速度传感器

意法半导体 (STMicroelectronics) 推出的LSM6DSOXTR正是这一趋势的典型代表,它在传统六轴 IMU 的基础上,创新性地集成了机器学习核心 (MLC) 和有限状态机 (FSM),

Spectro Scientific / AMETEK 斯派超 TurnkeyLab (ITL) 油传感器和分析仪

使用工业摩擦学实验室进行的测量确定润滑剂和机器状况,这是任何有效的预测性维护系统的核心。由于工业摩擦学实验室是一个交钥匙系统,由一家供应商提供和安装,因此在启动过程中,与新方法和新设备相关的担忧和学习曲线误差最小化。设备的状态监测可在仪器安装后和人员初始培训期间立即开始。

评论

您需要登录才可以回复|注册

提交评论

techmonitor

这家伙很懒,什么描述也没留下

关注

点击进入下一篇

欧洲新人工智能法案将加强伦理道德审核

提取码
复制提取码
点击跳转至百度网盘