When devices integrate information and data into new channels in the channel, technology giants are committed to the convergence of business models and data ownership.
Every American family now has one thing in common: children and adults are playing Amazon Echo or Google Home. "Hey Google, tell me the time", "Hey, Alexa, tell me the weather", or "Hey, Siri, give me rock and roll." The questions are all the same, the wonders are the same - all through voice interaction Technical realization.
Any reader with traditional business thinking will think that these devices are "wasting time." But if you look closely, you will realize that the focus is not on the novelty of these devices, but on the data of all the communication channels of major businesses in the consumer life.
Soon, retailers, auto companies, insurance companies, and banks will work with technology companies such as Apple and Amazon to singularize consumers into creative business models and hybrid worlds that are voice-based assistants. It will understand a person's life. They will be able to extract information from all devices, including cars, home devices, cell phones, CRM systems and more. They will know what you do every day because you have already registered and shared your data.
The power of edge computingMuch of the information we see is static information. However, with the rise of edge computing, the technology to complete analysis at high speeds by capturing data from consumer devices, Google, Amazon, Apple, Netflix and Facebook seem to be at the center of the new revolution, with business models and data ownership. .
Amazon CTO Werner Vogels said: "We are working with partners to make voice a natural interface for communication, because we have seen the fatigue of the application."
He added that they want the developer community to be able to use AI and natural language processing and collect data to create new business services.
In India, we can clearly see that this is happening. Headquartered in Bangalore, India, 24/ 7.AI is a $400 million business process outsourcing company that has reinvented itself as a company that uses voice interfaces for fully automated services.
Professor S Sadagopan, Director of IIIT-B, said: "Perhaps this is the first time we have seen the convergence of data and interfaces in business models."
Share data with cloud service companiesOn the other side of the world, NASA has partnered with Amazon to provide open source geospatial data to the developer community and students around the world. Sharing data between retailers and manufacturers of cloud computing companies such as Google and Amazon is only a matter of time.
In this new field, the new religion is no longer about God in the cloud, but a technology company in the cloud that can accurately predict human behavior, said Homodeus author Yuval Noah Harari. "Dataism."
Humans have been trying to predict behavior – the price movements in the stock market are one of the reference standards for humans to decide to buy and sell stocks. In the future, these consumer data collected in home devices will power Amazon and Google's cloud as an important basis for predicting the future, such as buying and selling stocks.
Simply put, your driving data is collected by the car and it is stored in the cloud. The car is an edge device, and the analysis is performed by a human-trained artificial intelligence platform, and the training is continuously and repeatedly learned through the model. IoT devices in the garage and at home will act as a connection point. The following things must happen due to edge calculations:
1. IoT devices capture data
2. Connect these devices to the car and home
3. Data is captured in the cloud
4. Data is processed in the cloud
5. Voice Assistant makes suggestions on the go
6. A symbiotic relationship with all networks and businesses.
“The voice assistant knows when you are going home. You can tell it to keep the lights on or keep the garage open,†Werner said. He added that all these repetitive tasks will be replaced by AI that calculates and analyzes large amounts of data.
This argument is not limited to technology companies. Even Volkswagen, the automaker with $250 billion in revenue, has positioned itself as an operating system company.
ChrisTIan Senger, head of the electric vehicle department at Volkswagen, said: "The car is now an operating system. Volkswagen's operating system will provide a range of services from startups and partners." At present, Volkswagen is considering to build a focus in 2020. A secure data platform.
"This is an era of artificial intelligence, cars will generate a lot of data. In Volkswagen, we don't understand this, but we need to speed up these services. Security will be very important. We have achieved traffic safety, but information stealing It’s all of us worried,†said Christian.
“We can do a few things in the car of the future – music, banking, etc. Protecting information security is the responsibility of car manufacturers. We must overcome this difficulty. People now see the benefits of sharing data, and they Understand that this allows them to get services and make their daily lives better," ChrisTIan said.
The new "enemy" between Google and Amazon is the data boundaryIn India, several companies such as i2E1, WizGo and Bezirk are already doing edge calculations.
I2e1 collects data from more than 20 million devices per month. Satyam Darmora, founder of i2e1, said: "We analyze more than 50 million data points per day and estimate traffic across physical markets across the country, which helps our customers efficiently manage their current stores and plan new stores."
A good example of this is UIDAI, the world's largest biometric database—a big data infrastructure is used to implement a product that supports the digital identity of every Indian. The challenges facing big data projects are not technology itself for India, but more about poor planning and weak links to actual business problems.
Rolls Royce recently signed an agreement with Indian IT giant TCS to collect data on all machines and engines and provide analysis through voice or chat assistants.
AWS CEO Andy Jassy said: "Some system integrators from India are using Amazon's platform to provide new-era services like AI and machine learning.
It's no wonder that big companies like Google and Amazon are trying to grasp the pulse of the future. Data is no longer a proprietary technology for companies to stay ahead of the competition, and they will work with these cloud and data processing companies to create new business models. Now, with these edge devices and platforms, data is more social, but users ultimately want Amazon and Google to protect it, just as we let the country protect our borders from foreign aggression.
The new enemy is the digital boundary, but this is also the direction that Google and Amazon are putting more effort into.
It seems that the era of edge computing has arrived.
Extended reading: edge computing, AI chips, vertical applications... 2018 artificial intelligence how to vote?
In the past three years, smart applications have brought data mining to a new stage, and much of the credit goes to the rapid development of machine learning. Under such a huge change, where do VCs look for new investment opportunities?
A growing trend is that the construction of core machine learning tools and module services is maturing, and we are most interested in companies that target edge computing and vertical applications, specialized hardware (such as AI chips).
The following are the four product directions that AI Business Week considers worthy of attention and tracking, and we believe that these products will occupy a place in the future of smart applications:
1. Edge computing - from data to sceneWith the popularity of machine learning, cloud service providers have begun to provide the latest GPUs to train machine learning models. Nvidia's financial report shows that GPU demand has been growing steadily, and data-related business has almost tripled in the past two years. However, data-related business is only part of machine learning.
Nvidia earnings
Usually, after the enterprise trains the machine learning model on the GPU, it also hopes to achieve the same effect of edge calculation. For example, smart speakers need to process some audio recordings locally (such as "Alexa", "Hey Siri" or "OK ​​Google") to reduce power consumption, ensure privacy requirements, and reduce latency.
However, efficient and feasible edge computing has always been an intractable topic, and it is necessary for artificial intelligence to land, because our current machine learning methods have the following three flaws: expensive, online computing, and hardware development. influences. Rather than relying on obsolescence, it is better to stand behind and fortunately, one has already started a trial around the existing pain points.
2. AI chip - from the bottom of the boost industryUsing professional-grade chips to run software more efficiently, the hardware that supports edge computing is becoming more important in machine learning.
The most striking aspect of bitcoin mining is this trend. From the initial CPU to the GPU to the FPGA, it is now a dedicated chip, and now it is an ASIC that can only perform single operations. Such hardware evolution clearly shows that the performance of dedicated chips will be more promising than general-purpose hardware such as CPUs.
Bitcoin mining demanded chip demand
Fortunately, we see the same trend in machine learning – Nvidia continues to optimize GPUs for deep learning, Azure and AWS release FPGAs tailored to specific workloads, and Google releases machine learning optimizations TPU. Compared with traditional general-purpose chips (such as CPU), the key issue in this new field of AI chip is whether it can invest enough time and money for specific needs and early stage. These factors play a decisive role in balancing the edge calculation.
Edge computing based on low-cost, high-performance AI chips can be applied to high-volume, high-value cases such as emergency collision avoidance for autopilots, emergency systems for industrial equipment, and speech recognition for smart home devices. The combination of soft and hard will highlight a very high advantage in terms of cost performance, and for start-ups, this is a new and worthwhile market opportunity, because the focus of the head technology company will be more inclined Establish a broad platform or system.
3. Natural user interface - breakthrough from the level of human-computer interactionA key component of intelligent applications is the continuous improvement of the user interface, which includes a variety of human-computer interaction methods - text, voice, vision, gestures and other forms of body language.
The development of technology has made users more and more lazy. Human-computer interaction has become more comfortable and natural than before. Today, more than 35 million American consumers use voice to control their smart speakers. Even so, the current generation of digital assistants still has a hard time understanding the meaning of the user, such as a lot of humorous misunderstandings between Alexa and her users.
At the level of human-computer interaction, startups can improve user experience and machine comprehension from two perspectives:
1. Reduce the link between human-machine communication and expand the potential range of machine response;
2. Put the person in the cycle of human-computer interaction to enhance the machine learning system. ;
In fact, customer service robots have become the most widely used AI assistants in some areas. These robots can output specific commands based on the corresponding text, such as sending a meeting invitation, updating an invitation, canceling an invitation, or requesting more information, which enables them to better predict what the user is asking for and create a more user-friendly content for their users. Interactive mode. Adding a person to this cycle to audit machine learning predictions not only ensures the correct response of the robot, but also enhances the understanding of the AI ​​assistant, thus greatly improving the user experience.
4. Vertical field AI application - focus in focusFinally, AI applications that can deepen and solve problems in vertical areas will be the main focus of our company's investment.
The production costs of AI applications are getting lower and lower, and we have seen many teams cut into vertical application output solutions to address customer pain points, finance, supply chain, and healthcare. In a sense, the underlying technology is not important here. The most valuable idea comes from understanding the target market. The tools for mining data are rich enough, but the vertical AI application that can empower the existing industry is in the process of land reclamation. period.
There are many opportunities for artificial intelligence and machine learning. The focus of investment over the next decade will be on finding smart entrepreneurs who focus on edge computing, vertical applications, and specialized hardware.
We have been in the Lithium Battery industry for many years and always adhere to the full capacity of the product.Sufficient quantity is not false standard,ensure the power stability of lithium battery ,stable endurance. Our 3.2v 50ah LiFePo4 Battery battery have high diacharge, low battery resistance, stable discharge, longer life cycle. We have higher temperature resistance, higher power discharge.
3.2V Battery Cells,Prismatic Phosphate Lithium Batteries,3.2V 50Ah,3.2v lifepo4 battery,3.2V LifePO4 Lithium Battery Pack
Jiangsu Zhitai New Energy Technology Co.,Ltd , https://www.zhitainewenergy.com