1. Bitcoin tumbled more than 10% back below $ 11,000;
Bitcoin prices plunged more than 10% on Wednesday to fall below $ 11,000, while prices of other major cryptocurrencies plunged as well, with Litecoin down 13.82% and Rebo Bank declining 9.58%, Ethereum fell 8.31%.
The following is a summary of the daily currency collected by Tencent Technology (Note: except bitcoin futures, the following data are from the encrypted digital currency data analysis website CoinMarketCap):
Bitcoin
In the past 24 hours, the spot price of Bitcoin fell 10.47%. As of 6:30 am today, the latest transaction price of Bitcoin was 10,522.50 U.S. dollars.
Bitcoin exchange for February delivery on the Chicago Board of Trade (CME) fell $ 1,425 to settle at $ 10,310, down 12.14%, touching as high as $ 11,340 for the day and as low as $ 10220.
Ethereum
In the past 24 hours, Ethereum fell 8.31%. As of 6:30 am today, the latest price of Ethereum was $ 840.38.
Litecoin
The past 24 hours, Litecoin prices fell 13.82%, as of 6:30 today, Litecoin latest transaction price of 210.57 US dollars.
Rui Bo coins
In the past 24 hours, Ruibo currency fell 9.58%, as of 6:30 this morning, the latest turnover of Ruibo currency was 1.00 US dollars.
Industry information
1. Some media said that the exchange had stopped trading and its actual controller Chen Jun ran .Therefore, Advisor Qu Jia Wei admitted that there are some related party transactions in Germany, fraud, and disappear, and accused Chen Army wanted to abolish the CEO without authorization, tge tge retreat, modify tge total Qu Jiawei said at this time the team is trying to solve the problem through legal channels.
2. Recently, the financial blogger @ Block Coin microblogging said that the frog dog official Micro Micro send it privately, broke the blockchain game frog dog dogger roll up the road, the shareholders money all swept away. Dog official microblogging forward the microblogging, respond 'yes', admit that there is no matter how follow-up, frog dog official micro did not disclose.
3. The Office of Foreign Assets Control (OFAC) of the U.S. Department of the Treasury issued a public response on the website saying that the digital currency issued by Venezuela should be considered a bond based on government credit and pointed out that pursuant to Executive Decree No. 13808, U.S. citizen transactions were prohibited for longer than 30-day Venezuelan national debt; citizens who trade Venezuelan digital currencies are at risk of being punished.
4. The U.S. Immigration and Customs Enforcement (ICE) arrested Morgan Rockcoons, the chief executive of Bitcoin, Inc., on the 9th as it traded privately for about 10 bitcoins with Undercover Agents. Court records show that on December 30, 2016, Rockcoons exchanged about 10 bitcoins (about $ 9,200 at the time) with law enforcement officials for about $ 14,500 in cash on January 8, 2017, knowing the cash source was "not clean." But Rockcoons' push Special and his later communications with Bitcoin Magazine seem to suggest that at first he did not know he was selling bitcoins to a law enforcement official.
5. Vice-Chairman of the European Commission, Valdis Dombrovskis, said the committee will hold a roundtable on February 26 to discuss the issue of cryptocurrency regulation. Delegates include representatives of central banks, representatives of regulatory authorities and some market participants.
6. According to Bitcoin.com, South Korean Prime Minister Lee Lo-yuan ordered a new Code of Conduct to prevent digital currency insider trading.Korean Anti-Corruption and Civil Rights Commission released last week's Guidelines for Code of Conduct on Digital Currency to governments and public agencies, There is an added requirement for digital currency that public officials should not use the information available in their job to assist in the investment or trading of property related to securities, real estate, etc. or to provide such information to others in order to facilitate their transactions or investments, FSS is not a government agency and FSS staff are not covered by the code of conduct FSS officials said they will develop internal guidelines prohibiting FSS staff from using internal information to trade or invest in digital currency.
7. According to the team of Ruipo Coins, Itaú Unibanco, the largest bank in Brazil and Latin America, will use xCurrent payment technology from RuiBo to handle cross-border payment and remittance business.
2. Bitcoin bounced two weeks 95% Agency views mixed;
Yin Yi
Bitcoin, once dropped to $ 6,000 on February 6, has recovered most of its lost ground after a nearly two-week call of the Spring Festival. On February 21, the bitcoin price continued its rally, rising to $ 11,720 straight Force 12,000 US dollars mark, up half a month since up to 95%.
The second-largest digital currency, Ethereum, also rose gradually after falling to a low of 565 on February 6, up 56% in two weeks and hovering around $ 880 at the current price. Reybo prices are similar to the previous two, with two weeks Up 77%.
Market believes that Bitcoin broke the $ 10,000 mark, allowing all walks of life to regain the confidence of digital currencies such as bitcoin.South Korean regulators began to loose the position for cryptocurrencies this week.According to South Korean media reports, and a Contrary to the previous month's tough sentiment, Choe Hyung-sik, head of the South Korean Financial Supervisory Authority (FSS), said the South Korean government will support the 'normal transaction' of digital currencies.
The report mentioned that since the beginning of this year, the South Korean government made a series of crackdowns on digital currency transactions, leaving many South Korean banks unwilling to open accounts for digital currency transactions. Choe said the South Korean government will encourage banks and digital currency exchanges Trading and cooperation.
In January of this year, the South Korean government once wished to ban the encryption of currency exchanges, causing all digital currencies to slump across the board, with the bitcoin rate plunging by more than 30%. Since then, Vice Prime Minister and Minister of Finance and Governor Kim Dong-yuan has said there is no intention of prohibiting or suppressing the digital currency market , The most urgent task at the moment is to supervise the exchange.
Research firm CanaccordGenuity said in a newly released research report that the historical price of bitcoin shows bitcoin prices will rise rapidly in a few weeks after a few months of consolidation.The report also said that from the potential bitcoin price The safest way to benefit from a rise is to buy and hold over the long term, and trying to measure hold time can be difficult.
Indeed, there have been times when huge amounts of money have flowed into the bitcoin market in the recent past, with a buyer spending nearly $ 400 million more on behalf of Bitcoin from February 9 to 12. Recently, Tim Draper, founder of venture capital firm Draper Fisher Jurvetson, Media interview said that the total market capitalization of the world currency market is 86 trillion US dollars in the future, a large part of which will belong to the digital currency. Three years ago, Tim Draper had predicted bitcoin price will reach 10,000 US dollars / BTC.
Equally optimistic is LlewClaasen, executive director of the Bitcoin Foundation, which reported that the bitcoin price could fluctuate within three to six months this year, according to U.S. media reports, but will eventually exceed 40,000 U.S. dollars turn off.
While Deutsche Bank and Goldman Sachs have all issued relatively negative reports on Bitcoin, JPMorgan Chase & Co.'s Outlook for Digital Currency is even more optimistic.
The report mentions that it is unlikely that encrypted digital currencies will disappear completely, and that Bitcoin can easily survive in many different forms and forms because many players are desperate to decentralize, peer-to-peer networks and anonymity.
However, the larger volatility of bitcoin, while appealing to traders, is not a welcome surge for those assets that are expected to be widely adopted.Based on bitcoin's historical performance, cryptocurrencies Of the volatility may be higher than the stock, commodities portfolio more than 10 times.
At the same time, the current regulatory issues of bitcoin have not reached consensus on a global scale and only the United States, South Korea, China and other countries have introduced corresponding policies on virtual currency transactions JP Morgan believes that for the digital cryptocurrencies to replace the government Issuing currencies and competing with them are all extremely difficult, for example the dollar, the euro and the renminbi are actually in a natural monopoly in their respective regions and will not easily give up their mint yield, which means that although the cryptocurrency will continue There is still a long way to go to challenge the gold market or become a global currency.
3. For the first time, the Bitcoin transaction fee crisis temporarily dropped to less than $ 1.
Last December, bitcoin trading was once as high as $ 34, or $ 34 per user on average. With such a high transaction fee it was almost fatal to crack down on small transactions, Valve announced in the same month that it would stop Bitcoin is being supported for payments, and many businesses switch to other alternative digital currencies with lower transaction costs, but bitcoin transaction fees dropped below $ 1 for the first time since September last year.
Why is the transaction fee going down? One reason is thought to be the gradual cooling of the bitcoin bubble, when Bitcoin's currency dropped to $ 6,000 at the beginning of this month and has now risen to $ 11,000, but at least it has slowed its fanatical demand for speculative purposes Decline, so transaction fees also fell.
The question is whether this will continue, and the Bitcoin community is currently developing a new payment network called Lightning that significantly expands the capacity of Bitcoin networks by shifting the most common transactions to the blockchain.
4. Google AI can interpret retinal images to predict the risk of heart disease;
Google, with its deep learning framework such as TensorFlow, is a researcher who uses AI (Artificial Intelligence) technologies to benefit all mankind in a cross-cutting way. According to the latest research report released by Google, the tools they are currently developing have been able to read through Retina images are used to predict the risk of heart disease, and the accuracy is similar to that commonly used in medical science. In addition, there is a "non-invasive" advantage.
Recently, Google and its subsidiary focused on health technology Verily co-published the development of AI algorithm (using machine learning technology) through the interpretation of the patient's retinal images and speculate on patient age, blood pressure, whether there is a history of smoking and their risk of heart disease The accuracy of a dissertation is comparable to current predictions (paper published in Nature Biomedical Engineering).
It is reported that, in order to train the algorithm, Google and Verily scientists analyzed the medical data of nearly 300,000 patients (including eye scan data and general medical data) .According to the algorithm developed by Google, specializing in machine learning Adre Luke Oakden-Rayner, a medical researcher at the University of Germany, one of Australia's eight elite universities, told The Verge that Google's algorithms provide more useful information than current methods by using clinical data to further assist physicians in diagnostics , But not to replace the doctor.
Google training algorithm to read the retina (retinal) imaging method has its own, as early as the medical community has long been developed .Fold eyes covered with blood vessels to reflect the degree of physical health, the past, doctors can use the microscope and retinal photography tools to Judging the patient's age, smoking history and blood pressure, and these are important indicators used to determine cardiovascular health.
It is understood, Google AI algorithm through the retinal image to predict the probability of patients suffering from heart disease, the accuracy rate of 70% accuracy than previously used to predict the risk of cardiovascular disease risk of SCORE slightly worse SCORE method requires blood tests (In other words, invasive therapies) with an accuracy of about 72%.
Both the left and right are retinal photographic images, with the left as usual and the right hand side showing how Google's algorithm calculates intraocular pressure through a specific blood vessel labeled as a green blood vessel. )
Alun Hughes, a professor of cardiovascular physiology and pharmacy at the University of London School of Medicine, said Google's approach appears to be reliable because of the long-established potential of predicting the risk of cardiovascular disease by observing the retina, and he believes it is possible to use AI Accelerate the speed of the current medical analysis, but he also believes that Google's algorithm needs further verification to obtain higher credibility.
In the past, Google also published the use of AI (machine learning technology) to interpret retina fundus images and further prevent the risk of blindness due to retinopathy in diabetic patients, which can further assist doctors in regions where the relative lack of medical resources is available in India diagnosis.
From the above results, AI (artificial intelligence) related technology to further replace the doctor to the patient's diagnosis of the day away from us still quite far.But if we can properly use the AI tool for medical research, disease diagnosis and prevention are all Can bring a certain degree of positive help and may also have the opportunity to improve the quality of medical care while reducing the workload of front-line medical staff.
5. The key to breaking the AI limit lies in the "human brain" ...;
Now it's time to take a serious look at human brain-inspired technology! Barbara De Salvo, chief scientist at CEA-Leti, pointed out that without the new paradigms and algorithms, it would eventually be difficult for AI devices to achieve higher processing performance and lower power requirements ...
It is not new that scientists are returning to the human brain to find new alternatives to today's computing technologies.
Barbara De Salvo, CEA-Leti chief scientist, pointed out that since the early dawn of Artificial Intelligence (AI) in the mid-1950s, there have been many twists and turns with regard to AI, despite the fact that many scientists and engineers are always interested in the concept.
However, it is time to take a look at brain-inspired technologies, at least, "We should recognize that there are other technology examples that will help us put forward good questions." De Salvo at this year's International Solid-State Circuits Conference (ISSCC), after receiving a keynote address from EE Times, said: "Without the new paradigms and algorithms, the industry will eventually find it harder to meet tighter power requirements."
While the industry has been pushing embedded AI platforms such as Movidius' Myriad 2, Mobileye's EyeQ5, and Nvidia's Xavier, in order to meet the demand for end-devices that handle large amounts of analysis with less power consumption, she points out: " We are far from the need to reach this step. "
(Source: CEA-Leti)
De Salvo compares the computational efficiency (GOPS / W) of the corollary phase and the GOPS of several AI chips during a keynote speech.
(Source: CEA-Leti)
She pointed out that there is still a huge gap between the needs of the industry and existing solutions. "No chip - whether commercial, prototyping, or chip design or development in academia - is capable of delivering less than 100μW However, this is exactly the kind of performance that edge devices need to achieve because they have to rely on energy harvesting or miniature batteries for years to come.
Why return to human brain research?
De Salvo said that we know that the human brain weighs about 2% of the body, yet uses 20% of the body's metabolism.
At 20W, the human brain can handle 1,011 GOPS. De Salvo stressed that so far, there is no processor in the world that "has the same performance and power as the human brain."
De Salvo explained that this energy efficiency class has evolved over the course of human evolution for a long time, maximizing brain function while minimizing energy use.
That's where the semiconductor industry borrows from biology, she said, and traditional computing architectures are working to meet the power requirements, mainly because "energy is consumed whenever processors and memory are in communication." By contrast, The synapses of the brain include memory and arithmetic in a single architecture, and she explains that this ingenious technique provides the foundation for a brain-inspired non-von Neumann computer architecture.
The essence of brain-inspired principles of operation lies in such things as spike coding and spike-timing-dependent plasticity (STDP), etc. She points out that in terms of the state of the neurons encoded in the system : In the past, neurons were coded using simulations or numerical values; the most recent trend in neuromorphic computing, however, is to encode the value of a neuron as a pulse or spike. She explains: "Neurons have no frequency and are purely event-oriented . "
The scientific community believes that spike coding and STDP have promise.If the input / output signal is represented by a pulse (spike), the multiplication between the input signal and the synaptic weight will fall to gating at the synapse level. The goal here is to reduce power consumption by setting up spikes or event-based signaling using heterogeneous solutions.
(Source: CEA-Leti)
Does this look like event-driven image processing technology developed by Chronocam? She replied: "Yes, but Chronocam's solution is aimed at" vision "applied to artificial retina, and our industry mission is to apply neuromorphic principles Extend beyond the entire computing area beyond vision. "
De Salvo said IBM's TrueNorth is an ideal example of a neuromorphic CMOS chip.Nowadays, DynapSEL, a dynamic neuromorphic processor built with 28nm FD-SOI, is a large-scale multi-core neuromorphic processor As another example, DynapSEL is currently one of the development programs under the NeuRAM 3 European Cooperative Research Program.
Neurons in TrueNorth are digitally processed, while DynapSEL is modeled, however, neither of these chips effectively performs the full function of the neuromorphic system because memory is not based on the functioning of neuromorphology.
The scientific community is trying to get memory closer to the processing unit, revolutionizing the traditional memory hierarchy and implementing "in-memory computing." So far, neuromorphic hardware has not yet reached its full potential, De Salvo said. "We Need to use ultra-high-density 3D architecture in neuromorphic computing "to maximize connectivity and reconfigurability between neurons and synapses.
Insect brain
We now know that there are 100 billion neurons in our brains, and smart insects like bees have about 950,000 neurons in about 1mm cuboid brain De Salvo pointed out that in all kinds of insects, the bees' brains are the largest She said: "We have mastered the neuronal image in the bee's brain. However, we know very little about neurons in the human brain."
Scientists are keen to profile the bees' brains because "bees are so smart." The bees have many sensors that can navigate through their foraging ways and know how to communicate in one group. To live together in a group, the bees must Know the function of other bees. "
Honey bees have excellent sensing systems and the ability to learn and remember, and they know how to integrate sensory data - making complex decisions in the form of "sensor fusion."
"Of course, the bee's brain is small, but its simplified form provides a good template for the AI system," De Salvo said. "It underscores the need to think in a systematic way because the organism does not separate the sensor from the signal deal with. "
AI latest progress?
Ten years ago, no one could imagine the breakthrough that we are seeing now in AI.
Barbara De Salvo, Chief Scientist, CEA-Leti
However, according to De Salvo, the industry may shift its focus too fast to AI applications due to the prospect of AI. In fact, she said: "We still lack a comprehensive understanding of the AI system ... just like deep learning . "
She explains that hidden hierarchies in deep learning and the resulting system complexity make it hard to say which attributes are the ones most responsible for improving performance. "Compared to general intelligence, learning, abstraction and reasoning The universality achieved is still very limited. "
Deep learning is excellent in terms of classification, but "prediction remains a fundamental issue in neuronal computing," De Salvo said. "Recently, neural networks have failed to perform simple tasks and humans have never missed anything in this area." De Salvo And cites A. Nguyen et al., "Deep Neural Networks are Easily Fooled: High Confidence Predictions for Unrecognized Images," she worries that this will eventually Hinder the development of the market and even "betray the trust of users, thus causing serious moral questions."
Compile: Susan Hong
(Reference text: AI 's Limits Send Scientists Back to the Brain, by Junko Yoshida) eettaiwan
6.Yole: 2023 machine vision industry will reach 4 billion US dollars
According to Yole Développement survey, the overall size of the market for machine vision cameras will grow from 2 billion U.S. dollars in 2017 to 4 billion U.S. dollars in 2023 with an overall compound annual growth rate of 12%. Machine vision is the wave industry The heart of automation. Industrial and automated machine vision is not limited to robotics, but it involves almost any work related to manufacturing.
According to Alexis Debray, a technology and market analyst at Yole, the quality requirements drive machine vision in the automotive, electronics, semiconductor, food and packaging industries as shown in the diagram, which is the main application of machine vision to grow from $ 1.4 billion in 2017 to 2023 In fact, machine vision has reached 140% CAGR, especially in automotive applications, not only in industrial automation but also in smart agriculture, license plate recognition, and more recently in self-driving applications. In 2014 Swatch Group announced the launch of Sistem51, the world's first fully automatic mechanical watch, and Canon announced the deployment of one million robots in the camera manufacturing and Foxconn manufacturing consumer electronics products. As a result, sales of automation products surpassed the growth of industrial production in the future This trend may accelerate.
