Blog Archive

Thursday, August 31, 2017

小米、阿里智能音箱体验,它们到底是下一个“入口”,还是 Siri 一样的玩具?_智能_好奇心日报

小米、阿里智能音箱体验,它们到底是下一个“入口”,还是 Siri 一样的玩具?_智能_好奇心日报:



'via Blog this'

声纹识别在智能家居中很关键,但面临的问题也很棘手-钛媒体官方网站

声纹识别在智能家居中很关键,但面临的问题也很棘手-钛媒体官方网站: "声纹识别在智能家居中很关键,但面临的问题也很棘手"



http://www.tmtpost.com/2539771.html





'via Blog this'

GitHub - singaxiong/SignalGraph: Matlab-based deep learning toolkit that supports arbitrary directed acyclic graphs (DAG). Support DNN, LSTM, CNN layers and many signal processing layers. Include recipes/examples of using the tool for various tasks.

GitHub - singaxiong/SignalGraph: Matlab-based deep learning toolkit that supports arbitrary directed acyclic graphs (DAG). Support DNN, LSTM, CNN layers and many signal processing layers. Include recipes/examples of using the tool for various tasks.:





SignalGraph is a Matlab-based tool for building arbitrary directed acyclic graphs (DAG) for signal processing. The original purpose is to make it easy to apply deep learning techniques on speech signals on the Matlab platform. It should also be applied to other tasks, especially involving temporal trajectory data.



'via Blog this'

Tuesday, August 29, 2017

国内外智能音箱对比:国产数量不逊国外,但仍有国外产品的影子

国内外智能音箱对比:国产数量不逊国外,但仍有国外产品的影子:



阿里 | 天猫精灵X1:看亚马逊「不爽」
唤醒词:「天猫精灵」
天猫精灵发布之前就被传了很久,连同阿里的人工智能实验室也一同曝光。
天猫精灵X1内置的是阿里的第一款智能语音系统AliGenie,目前,能实现的功能包括智能家居中控、语音购物、手机充值、叫外卖、音频音乐播放等。
不过,AliGenie最特别的还是能和阿里的电商平台关联,购物、充值、查物流…
是不是又很熟悉……没错,Amazon Alexa也能做到这些。当然,这款产品是不是对标亚马逊,也只有阿里自己心里清楚。
除此之外,天猫精灵还加了指纹锁功能,据说2代的叮咚音箱也会在随后的更新中加入这一功能。
硬件上,天猫精灵用的是联发科的语音识别芯片,配了6个麦克风阵列,底部的发光带倒是能给用户不错的视觉效果。


'via Blog this'

Monday, August 28, 2017

Increase Your Performance by Changing Your Seat?

Increase Your Performance by Changing Your Seat?: "Increase Your Performance by Changing Your Seat?"



http://jobs.ieee.org/jobs/content/Increase-Your-Performance-by-Changing-Your-Seat-2017-08-24



By: Nancy Ordman
Employers often invest resources to improve employee performance. They tweak compensation and training programs and workspace design, among other factors, to accomplish this goal. They spend a lot of money to increase the value of their human capital; according to a study by Deloitte, in 2013 businesses spent an average of $1,169 per employee—$1,847 for tech workers—on human resource development. Plenty of virtual pages in the literature of management report on the relative effectiveness of various incentives and training programs. For the many employees who work in open-plan offices, the answer could be as simple as rearranging the seating plan.
Research recently reported in the Harvard Business Review indicates that an employee’s neighbors can have a big impact on work performance. The authors did not study the kinds of distracting neighbor behaviors—loud telephone conversations, space-poaching, chatting, general noise—that often evoke complaints about open-plan offices. Instead, they looked at the spillover effect that a highly productive employee can have on a less-productive colleague and concluded that spatial management can have a significant impact on performance.
Source: Share Alike 3.0 Unported license.
Authors Jason Corsello (Cornerstone) and Dylan Minor (Kellogg School of Business, Northwestern University) collaborated on the pathbreaking research that led to this conclusion. They analyzed two years’ worth of data on 2,000 employees at a large technology company, mining five human resources data sources. Among the data points included were employee seating location and distance between employees, as well as worker performance data. They broke performance data into productivity (amount of time to complete a task), effectiveness (the number of tasks a worker referred to someone else to solve), and quality (client satisfaction with outcome).
Measuring distance between co-workers enabled Corsello and Minor to model spillover, or impact of one worker on another, as a function of physical distance between workers. One of the company’s human resource practices, moving workers quasi-randomly to different teams based on company need, lends additional credence to the researchers’ conclusions.
Corsello and Minor classified workers into three groups:
  • Productive workers who worked quickly but whose work lacked quality (25 percent of the sample)
  • Quality workers who worked more slowly but who produced superior work (25 percent)
  • Generalists who were somewhere in the middle (50 percent)
The results uncovered strong symbiotic relationships between Productive and Quality workers, groups who are somewhat opposite. Quality workers might observe and adopt Productive workers’ good time-management skills, for example. Seating these types of workers in close proximity to each other and Generalists together increased productivity 13 percent and effectiveness 17 percent.
This research also touched on Toxic employees, defined as workers who were terminated for misconduct, workplace violence, drug or alcohol abuse, sexual harassment, fraud, and other violations of company policy. If Toxic employees sat near each other, the probability that one would be terminated increased 27 percent. Toxic workers influenced nontoxic workers, increasing the latter’s chances of turning Toxic.
What causes spillover?
The authors speculate that a combination of peer pressure and inspiration drove the results. Since employees moved regularly between teams, individuals likely rotated among some combination of Quality to Productive to Generalist to Toxic colleagues over the course of a year.
The takeaways from this research are equally important to employers and employees. If a worker sits near a toxic employee, they should ask to be relocated or for some other change to alleviate the negative spillover. First-line managers, who should know best what category best describes each of their staff, can rearrange seating to best achieve the benefits of positive spillover.
The researchers didn’t address other impacts that an open-plan office can have on productivity, or whether a particular configuration would work better than another. “Open plan” can mean anything from long straight tables with workstations every four feet to more private desk spaces with visual barriers between workers. Backlash against open-plan offices is growing, including the observation that some types of work are better-suited than others to wide-open spaces. Since the overwhelming majority of office spaces have some manner of openness, managers and workers should exploit whatever advantages they can find.

'via Blog this'