当前位置:网站首页>Anti counterfeiting AI ownerless light, who is playing real
Anti counterfeiting AI ownerless light, who is playing real
2022-07-22 12:23:00 【New observation of science and technology】
AI No owner lights !
Major smart home brands , Traditional lighting manufacturers , Both launched their own ownless lights this year , And all in step “AI” The label of .
This not only reminds people of the time when blockchain just became popular , Anyone who starts a company , And distributed 、 Decentralize and connect .
It seems that I don't stick this label , It's out of touch with the world .
And looking back AI No main lamp , Have to say , History is always similar .
You said smart home brand is close AI The label of , And a little credibility , After all, the Internet of things +AI, This combination is reasonable , It is also the general trend .
But those enterprises that have nothing to do with high technology , Whether it's real estate , It's still dry lighting , All claim to have launched AI No main lamp .
Give a feeling , Now? AI Is it so cheap , It can be achieved casually ?
today , Let's incarnate and fight the dummy , Give these a good dozen AI Fake without main lamp !
These traditional manufacturing industries , You can see through it at a glance , I won't occupy the page here , After all, fake can't be fake anymore .
This article focuses on smart home brands : Green rice Aqara、 Euribo ORVIBO And Ella IOT AylaHome.
Look at these all house smart brands that claim to master core technology , To launch the AI No main lamp , Is it true AI, It's still fake AI!
Before the formal comparison , We need to consult AI The real boss of the industry , What is real AI No main lamp .
“AI There must be big data to support data analysis . For smart homes , If someone did this , It must be based on the device in the user's home , And the overall environmental data of the home , Also be able to do data aggregation analysis .
in addition ,AI It has a feature , Be able to keep learning , This underlying technology is called machine learning , For smart homes , It should constantly learn the habits of users , Optimize the rule model or optimize the active trigger model , If you don't learn , Nor is it. AI.” well-known AI Yang Gong, product director of Unicorn company, concluded .
We can catch from this passage 2 Key words :“ big data + machine learning ”
Next , We target three companies , focusing AI What is the main lamp ? What big data does it need ? What is the difficulty and cost of implementation ? What is the difference between rule engine and rule engine ? Expose these AI It's really fun to have no master lamp , Or is it a pure gimmick .
Euribo ORVIBO: Empty products none AI Algorithm , Concept is greater than substance
Although oribo is in the smart home brand , Take the lead in testing the water without main lamp , But its main product is still the traditional ownerless lamp , Traditional panel control , Unable to voice 、 mobile phone APP, Sensor linkage control .
last year , Euribo has tasted the market sweetness of ownerless lamp , It began to officially turn to the research and development of intelligent ownerless lamp this year .
in other words , The time when euribo launched the smart home lamp is 2022 year . See clearly , This is just a smart home lamp , Not really AI No main lamp .
Although oribo builds its own private cloud , But it uses a rule engine , That is, by setting the scene and trigger conditions in the cloud in advance , Reach a thousand people “ Active ” Lighting services .
More generally speaking , Whether it's an old man or a child , Men or women , Whether it's family or strangers , Whether it's south or North , After entering this scene , The same services are triggered .
For example, the brand manufacturer has set it in advance , The living room at home is in the evening 7 Click to turn on the light automatically . Then whether your time zone is night or not 7 It's getting dark , Whether you get home on time or not , It will be at night 7 Turn on the light on time , This is rule engine driven “ Active ” service .
The most typical case , Xinjiang region , Everyone who has been there should know , Xinjiang in summer is night 11 It's getting dark , morning 5 Dawn begins at halfpastten ,14 Have lunch at 12:00 , that , If it is a rule engine driven lights will be what ?
Not in the cloud AI Algorithm and machine learning ability , It can be said that oribo's has been blocked AI The way ,
Even if I want to do it now , At least we need to invest nearly 100 people AI R & D team , It takes a year to really realize , And this is only access AI There is no access certificate in the main light field , More complex data and algorithm models , It's just AI The last mountain without a main lamp , Orebo wants to finish in a short time , It's almost impossible .
So-called AI No main lamp , It's just a gimmick chasing concepts .
Green rice Aqara: Backed by Jinshan cloud , But it's hard to make bricks without straw
The situation of green rice is different from that of oribo , According to his former employees , Green rice uses Jinshan cloud .
Although the external information of Jinshan cloud , Basically, the architecture can meet AI There is no realization condition of main lamp .
But as we mentioned earlier ,AI Without a master lamp, you need very large and complex data to constantly feed and learn .
These complex data , Both environmental data , Such as time 、 The weather 、 Luminous flux , Light intensity and other parameters ,
It also contains the spatial data , Such as the orientation of the room 、 The size of the window 、 Decoration surface material 、 The height of the house 、 Installation height of lamps .
It also contains user behavior data , Such as behavior and posture 、 Sound decibel 、 Heart rate 、 Respiratory rate, etc .
It is not enough for green rice to collect only from the product port of the user's family .
On the other hand , With big data is the algorithm model ,AI What kind of algorithm will be used for no master lamp , I'm not sure , But light is related 、 Environment related 、 User related 、 Designing relevant algorithms should be essential .
at present , There is hardly a data company on the market , These data can be provided for green rice , There is no one AI Enterprises in the field of algorithms , It can help green rice establish these algorithm models , Even if it is ByteDance 、BAT In this way AI A leading manufacturer .
so to speak , Data and models of ordinary industries cannot be reused in AI No main light , Even though Jinshan cloud has a very mature AI Machine training architecture and algorithm model , It's also for nothing .
So , Although green rice has got into AI Tickets without main lights , But really want to achieve AI No main lamp , Still out of reach .
Eira IOT AylaHome: A long run AIoT many years , It's still too early for the melon to ripen
Ella IOT as PaaS Veteran in the cloud , stay 2019 Started to do in the cloud AI Partial upgrade .
2021 In, it launched based on AI Engine wide intelligent operating system , This system has three key points ,“ people 、 Space 、 Environmental Science ”, It fits the above AI Data source without main lamp “ User data 、 Design data 、 Environmental data ”.
Ella IOT began to think very early AI What constitutes the data of the non main lamp .
Many people think that , Capture the historical data of user operation lights , It is enough to establish the data model related to the light source . Others think that , Environment affects light ,AI The data source without main light should be time zone 、 climate .
But actually , The user is only the user of the light source , And the design of light source , All are designed according to the user's home space , Such as the illumination in the time zone 、 The climate ; The floor height of the room 、 toward 、 Luminous flux ; The position of the window 、 size ; Installation position of lamps 、 Angle and a series of special conditions , Special design .
A lot of times , Users' use of lights is in “ The small white ” Stage , Most people only use basic lighting . In addition, the operation of traditional lighting fixtures is complex , Even if you know that you can adjust the ambient lighting , There will be no high-frequency use , That means , Only capture the data of users using lights in the home space , Not enough to be AI Reference basis for active recommendation .
Big data warehouse established by Ella IOT , It's not just about grabbing the operation data of the client , Illumination and climate in the time zone , It has also established cooperation with tens of thousands of lighting designers across the country , Constantly with implied diversification 、 The design data of personalized spatial parameters are fed to the cloud Algorithm .
meanwhile , Based on these data , Build on people 、 Space 、 Environment centered algorithm model , Constantly simulate different spaces , In different circumstances , Changes in natural light sources , And human behavior logic in different spaces and environments . This is what Ella IOT builds based on AI Space smart base .
but AI Real landing without main lamp , It takes long training and trial and error , Ella IOT, which is just starting , Just deal with all kinds of complex situations in the cloud ,SaaS The increasing lighting scene at the end , Low frequency of use of various light scenes on the client , It will take most effort , Want to really AI No master lamp is mature , It will take a long time to verify .
边栏推荐
- 循环结构之for()循环和二次for嵌套
- Research on the evaluation and certification system of IOT terminals in smart city scenarios in China
- Ansible project best practices
- Do you dare to use BigDecimal without mastering these pits?
- Application of 3D simulation and visual display in industrial plant
- Flutter实战-自定义键盘(三)
- Virtual machine dual network card configuration
- Actual combat of flutter - customized keyboard (III)
- 网站上线功能及兼容性测试小工具集…
- Legendary mobile game service tutorial: legendary mobile game ares engine installation tutorial
猜你喜欢
Machine learning Basics (4) filters
The wave of the meta universe is shocking. Seize the opportunity and work together
Morphology of machine learning foundation (5)
The evolution history of the background platform transformation of flybook management after the introduction of cloudwego
JMeter test script learning (login script)
HCIP第十二天
Part 02: distributed load balancing
Day04 禅道-从安装到卸载
缺陷報告作業
Chrome 开发者工具 network 显示 Provisional headers are shown 的几种原因
随机推荐
Cdh5, cdh6 Deployment Guide (stable)
Effect of database index on write operation
The wave of the meta universe is shocking. Seize the opportunity and work together
Chang'an's new strongest SUV is coming. The interior is very three-dimensional, and the sense of science and technology is bursting
SQL bypass case
《如何打一场数据挖掘赛事》进阶版
从基础到进阶,100道测试开发面试题,进大厂涨薪必备
网络空间资产测绘
数据库索引的缺点
Free download! Databricks data insights: from introduction to practice
Day010 循环结构
LSA type, OSPF optimization and topology configuration
6.paddlepaddle之自定义模型构建使用
[quick start Tutorial 4] crazy shell · open source formation UAV -openmv script burning
打假AI无主灯,欧瑞博、艾拉物联、绿米Aqara谁是玩真的
【oops-framework】本地存储
Interface document evolution atlas, some ancient interface document tools, you may not have used them
一文读懂Elephant Swap的LaaS方案的优势之处
history.go()和history.back(),history.go()不刷新
Chrome 开发者工具 network 显示 Provisional headers are shown 的几种原因