On May 8th, the Google I/O 2018 Developer Conference was staged in Mountain View, California. In the second year of the comprehensive promotion of the "AI first" strategy, this year's artificial intelligence can be said to be at the top of the audience at Google's conference.
There is no doubt that the tech circle has once again been screened by Google's wonderful demo. This kind of look is all about AI black technology, bursting, it is likely that the media and public opinion can not think about the meaning behind these AI functions in a short time. But after a day or two of calming down, the discussion around the conference seems to have some new directions.
For example, some media said that most of the AI ​​functions displayed by Google this time can be seen in the Baidu system.
In fact, if you think about it, you will find that at this stage, Baidu and Google's core are doing one thing: using the natural interactions brought by AI and the ability to combine multiple data, the original functions of the war are glued together.
Let's compare the wonderful "similarity" of this Google conference and see what mystery lies behind it.
From sound to everything: voice interaction takes over the mobileUsing voice to drive mobile products is a major proposition of this year's Google Conference.
Not counting the Google assistants who are still in the research and development stage, Google’s main new feature is the use of voice assistants to drive maps. Similar functions have also appeared on mobile phones Baidu, Baidu map, such as Baidu map app can use DuerOS to achieve voice operations.
It can be seen that Google is strengthening the voice assistant's connection with its mobile products. For example, voice control of take-away, map, search and other capabilities. And this is obviously Baidu has long noticed, and constantly open the field of the situation. Open the mobile phone Baidu today, you will find that the voice interaction provided by DuerOS can operate the search, find content and find various services. In comparison, Google's voice assistant is still in the process of exploring mobile products, and Baidu has already loaded DuerOS into the autopilot scene and introduced the in-vehicle system (DuerOS for Apollo).
(FAW Pentium SENIA R9 equipped with DuerOS for Apollo vehicle system)
The software capabilities of Google I/O, but at the meeting, Google also demonstrated the voice assistant + screen speaker solution with Lenovo. Interestingly, in the first quarter, Baidu just launched a small-screen smart speaker at home, and has begun to introduce vertical market solutions and strengthen skills development.
There is no doubt that Google and Baidu have a high degree of overlap in the commercialization of voice interaction. The fast implementation of sound-driven mobile software can be seen as a strategic dividend for Baidu and Google in the field of voice interaction and NLP.
Know you better than you know: the flow of information embraces the correct posture of AIIn addition to voice, this time Google also showed off the new Google News + AI program. Combined with AI technology, Google News can accurately determine what users are interested in and push.
For Chinese users, this ability is obviously no stranger. However, it should be noted that Google is adopting a knowledge map with a new deep learning algorithm to recommend content for users. According to the information presented at the conference, this feature pays more attention to the integrity and internal relevance of the news, rather than simple keyword recommendations.
There is a hidden change that is often confused in the country: based on data recommendation content, and the opportunity data + depth AI recommended content is not the same. One of the reasons why today's headlines are criticized is that the algorithms behind them often only recommend based on keywords, ignoring the content and logic of information reading. The result of indulging in keyword recommendation must also be the proliferation of the title party and the picture party. Content producers can only rely on the eyeball to get out. This has also caused a number of disturbances in the headlines of today.
Baidu has also noticed this problem. The current Baidu information flow product is an AI recommendation engine that combines the manual editing and reviewing of media attributes, AI content understanding ability and knowledge map correlation ability.
In the experience of information flow products, compared with simple and rudimentary keyword recommendation, the combination of deep AI technology and information flow will bring about obvious high-quality content and deep reading experience, thus stimulating how the user experience and goodwill increase The combination of AI depth and information flow provides a better value experience and may be the main exploration path for the next news product.
Obviously both Google and Baidu have chosen to go deeper into the combination of deep AI technology and media value.
Exploring the world begins with exploring AI: Autopilot racingFrom travel to driving, it is also one of the main themes of this Google conference. It highlights the new developments in Google Maps and Waymo driverless, and apparently has hoped to overweight the AI ​​experience in the field of travel, ready to proceed to the commercialization process. This rhythm is similar to Baidu's strategic direction today.
For example, the AR navigation function displayed in Google Maps has actually been deployed in Baidu Maps and Apollo application platforms. Last year, Baidu Maps launched the first AR navigation capability in the world. With the road adsorption image algorithm and ability, this function has several core capabilities such as route adsorption, sensor data smoothing, road matching, AR route planning, etc., and can realize new experiences such as voice interaction and support for inertial navigation.
The most direct comparison, embodied in Baidu's AR navigation experience, now has a mass production program, which is clearly Google's need to continue to catch up.
On the top of the map, what is the unmanned race of the two major technology giants. In the second half of this year, it is likely to become the "drift period" of unmanned landing. Commercial landing and real car on the road will be the focus of the giant competition.
At the conference, Google released the next unmanned taxi trial plan for Waymo in Arizona.
On the Pacific side, Baidu’s landing plan seems to be smoother. From the upcoming Apolon, to the unmanned landing plan of more regions, with the support of national policies and industrial ecology, the race seems to be more favorable to Baidu.
From the mobile era to the AI ​​era: the ignored artificial intelligence spider web effectIt is found that Google’s cool demo at the press conference hides the high consistency of the AI ​​giants and may provide a new perspective for us to observe the commercialization of AI technology.
We know that in the B-end market, AI is more implemented as a means of blessing and empowerment. It will not subvert anything, but it will strengthen the company's capabilities, comprehensively improve business efficiency and reduce costs. In the Internet field, such capabilities often seem to be ignored by the public.
From the extent of Google's press conference and Baidu's AI business system, we will be a little surprised to discover the fact that AI is not just about creating new things, but also about the so-called "old things", such as the mobile Internet era. The fields of mailbox, information flow, mobile map, O2O service have become full of surprises. For example, the AR navigation capability of Baidu map reduces the use cost in many scenarios, which can be said to make the map experience have a qualitative improvement.
Through the combination of AI and information, the combination of natural interaction and service, and the combination of geographic time and space and recognition capabilities, the mobile service system is emitting new vitality due to AI. We often say that the mobile era is how the second half is, but in fact it may not be the end of the first half.
Just like blessing traditional industries, AI can also bless the mobile Internet. Today's Baidu and Google are doing something like this. What's interesting is that through the combination of Baidu AI technology and mobile services, the market and user increments obtained far exceed market expectations.
Then, in the next step, will the pattern of the mobile industry change more?
The products can provide various specifications of rectifier bridge devices according to customer requirements. The electrical properties, appearance, reliability, safety indicators and environmental protection indicators of the products all meet the relevant standards.
Planar Die Construction Sealed Glass Case Ideally Suited for Automated Insertion - 75V Nominal Zener Voltages
Case: MiniMELF, Glass Terminals: Solderable per MIL-STD-202, Method 208 Polarity: Cathode Band Approx. Weight: 0.05 grams= 25°C unless otherwise specified Symbol Pd VF RqJA Tj, TSTG Value to +175 Unit mW V K/W °C
Characteristic Forward Voltage = 200mA Thermal Resistance, Junction to Ambient Air (Note 2) Operating and Storage Temperature Range Notes:
1. Tested with Pulses, 20ms. 2. Valid provided that Electrodes are kept at Ambient Temperature.
1. Tested with pulses = 20 ms. 2. Valid provided that electrodes are kept at ambient temperature.
VZ, ZENER VOLTAGE (V) Fig. 1, Zener Current vs Zener Voltage
25 20 VZ, ZENER VOLTAGE (V) Fig. 8, Zener Current vs Zener Voltage
TA, AMBIENT TEMPERATURE (°C) Fig. 3, Power Dissipation vs Ambient Temperature
15 20 VZ, ZENER VOLTAGE (V) Fig. 4, Differential Zener Impedance
VZ, ZENER VOLTAGE (V) Fig. 5, Junction Capacitance vs Zener Voltage
MINI MELF,Smd mini melf,zener mini melf,mini melf package,mini melf resistor,mini melf diode,mini melf resistor datasheet
Changzhou Changyuan Electronic Co., Ltd. , https://www.cydiode.com