Careful analysis of the company that released financial review article written by a robot "the August CPI rose 2% a 12-month high", you will find that it just lists and forecast data without additional analysis.
In fact, this is the most basic writing robot this robot automatic writing was born as early as a few years ago, and automatic writing is not the most terrible this algorithm based on data and lists of figures, but one day the robot really has the analytical skills, while the human machine is being developed this ability.
Following a comprehensive description and analysis of the top international news media examples of automated writing machines, these examples tell people "writing is both science and art, seems to be one of the least able to use automated things. But it was automated, and the program algorithm is improving rapidly. "Article excerpted from the book the age of robots, the author Mading·Fute
October 11, 2009, the Los Angeles Angels in the United States major league baseball playoff contest against the Boston Red Sox, and New York Yankees for the League Championship and the chance to enter the World Series Championship. This victory was particularly excited by angels, for just 6 months ago, most promising players and pitchers among them – Nike·yadenghate (Nick Adenhart) was killed by a drunk driver. A sports journalist at the beginning of the article describes the game:
Angel was 2 points behind the Nineth, the situation is very bad, but fuladimier·geleiluo (Vladimir Guerrero) key remember hitting Los Angeles with hope and eventually Yu fen Wei Park on Sunday with a 7:6 victory over the Boston Red Sox.
Guerrero as the angels scored 2 points, 4 stroke, hit 3 hits.
Guerrero said in an interview: "If you want to commemorate the Nike·yadenghate, as well as Anaheim took place in April, I may be used (in my career) the most beautiful blow, because I want to dedicate it to my former teammates, the guy died. "
Guerrero home run of the season as a whole performed very well, especially the day of the game. During the show on the day, Guerrero attacked index of 0.794. In 26 games during the day, he hit 5 home runs, scored 13 points.
The author of the article may not immediately get any writing awards, but the article is still a remarkable achievement: not because it is readable, grammatically correct, or baseball game has an accurate description, but because the author is a computer program.
Mention this software called "StatsMonkey" by students and researchers at Northwestern University's intelligent information laboratory created. StatsMonkey by objective data processing for a specific game, able to automatically write a sports article. The system does not simply list facts, it incorporates key elements of sports journalists will also join. StatsMonkey through statistical analysis, identify significant events that occurred during the game, and then it will generate a natural article, summed up the whole dynamic of the game, as well as attention to key points and key players of the game.
Top news media has used automatic writing technique
In 2010, the Northwestern University oversees the StatsMonkey computer science and journalism student researchers raised venture capital for the development team, and the establishment of a new company "automatic writing technology company", to commercialize the technology. Company hired a group of leading computer scientists and engineers, and then abandoned the StatsMonkey computer code, build a stronger, more comprehensive artificial intelligence engine, named "Quill" (Quill).
Automatic writing, including Forbes, technology has been used by top media, its automatically generated articles cover various fields, including sports, business and politics. The company's software approximately every 30 seconds to generate a news story, many of them published in don't want to admit to using the service on popular websites. In 2011 the industry Conference, the Wired magazine writer automatic writing technique for shidifen·liewei hope kelisidian·hamengde, co-founder of the company forecast over the next 15 years the proportion of news articles written by the program, and his answer is: 90%.
Automatic writing outside of technology companies to look far into the news business. "Quill" was designed to be a general analysis and narrative writing engine, to produce a range of quality report required for inside and outside the industry. "Quill" through various channels to collect data, including transactional databases, financial and sales reporting systems, Web sites, or even social media. And start analysis to tease out the most important and interesting facts and ideas. Finally, it will collect all the information into a coherent article, claims to be able to keep up with the best of human analysts.
"Quill" system configuration is successful, can almost instantaneously and generate business reports can provide uninterrupted, completely without human intervention. Company is one of the earliest supporters is In the Central Intelligence Agency's venture capital – Q-Tel. Company's software could be used by United States intelligence agencies collect raw data flow automatically into easily understandable language formats.
"Quill" technology proved to us, once only university education of skilled professional and technical personnel to control areas are vulnerable to automation. Of course, knowledge work typically need capacity. In addition, the analyst may also need to know how to get information from various systems, statistical or financial modeling, then people understand written reports and presentations. After all, the writing is both science and art, seems to be one of the least able to use automated things. But it was automated, and the program algorithm is improving rapidly. In fact, because knowledge work can be automated using software only, so in many cases, these jobs are lower than they are in need of physical skill is more likely to be affected.
Also, writing also happens to be the employers always complain about the college students ' ability of field. A recent survey of employer's, about half of the new hire's two-year college graduates and more than 1/4 of the four-year degree graduates of poor writing ability, some even reading skills are also very bad. If you really like automatic writing technique of intelligent software company said, comparable to the most capable human analysts, that all future university graduates to attain knowledge of employment growth is even more suspicious, and especially for those of you who didn't prepare.
"Quill" is just one of the many new applications
"Quill" writing engine is one of the many new applications are being developed and used by global corporations, institutions and Government acquisition and storage of the data. It is estimated that the total amount of data stored in the world now is measured in tens of thousands of exabytes (1 exabytes = 1,000,001,000 gigabytes), and the data also has its own Moore's law to accelerate growth, approximately doubling once every three years. Almost all data are now stored in digital format, so you can access directly from the computer. Google servers to handle every day only about 24 petabytes (1 Petabyte = 1.001 million megabytes) of information, chiefly millions of users every day in search of information.
All these data have a wide variety of sources. As far as the Internet, sources, including access to the site, search queries, email, social media interaction and ad clicks, and so on. Business sources, including transactions, customer contact, communication, and financial data, accounting and sales system. In the real world, the sensor will continue to capture factories, hospitals, cars, aircraft, as well as many other consumer electronics and industrial equipment of real-time operational data.
The vast majority of these data is what computer scientists call the "unstructured" data. In other words, the capture of data in a variety of formats, often have difficulty matching or comparison. With traditional relational database systems are very different, the traditional information systems, lined up together, fast, reliable and accurate search and retrieval. The unstructured nature of data has led to specialized for development of new tools to understand various channels to collect information.
The rapid development in this field, at least in the limited sense, is only an example computers began to encroach on the uniquely human abilities. After all, continually dealing with our environmental resources in the amount of information is one of the particularly good at something. Of course, the difference is that in large areas of the computer to be able to scale processing of information, and it is impossible for a person. Big data is to include business, politics, medicine and almost every field of natural and Social Sciences has had a revolutionary impact.
Behind all this is big data
Both now and in the future, growing mountains of data are increasingly being seen as exploitation of valuable resources. As oil and gas and other extractive industries continue to benefit from technological advancements, we can be sure that computing power to accelerate development and continuous improvement in software technology will enable the company to explore new strategies, increased profits directly. In fact, probably investors expected to make data-intensive companies like Facebook have made a huge market valuations.
Machine learning is a computer data, and statistical relationship and found it written in their program of a technology, it is one of the most effective means to get the value of the data. Machine learning in General consists of two steps: first algorithm known data, and then use the new information to solve similar problems. A common application in machine learning spam filter. Google online translation tool can prove most machine learning, when it came out, is also the most exciting. It is through the analysis and comparison of millions of pages of text has been translated into many languages, and one possibility is called "Rosetta" (Rosetta Stone) way to work. Machine learning in many ways, but one of the most powerful and one of the best methods is the use of artificial neural network technology, or with one of basic operational principle of the brain in the same system. When the beer Street machine not indeed
Intelligent algorithm for large data and accompanying it are the workplace and their bosses have a direct impact, especially large ones, more and more employees work lots of assessments and their social interactions and statistics. More companies than ever to rely on so-called "specialists" to hiring, firing, appraisal and promotion staff, and collected and the amount of work data is amazing. Some companies capture the keystrokes for each employee for each word. In the case of employee knowledge, may also collect their email, phone records, Web searches, database queries, file access, device access, and many other types of data. Although all of these data collection and analysis is General at first in order to achieve more efficient management and staff performance appraisal, but it could end up being used for other purposes, including the development of software for automation of work being performed.
Data revolution might of knowledge occupations has two especially important influence. First, in many cases, capturing data directly with specific tasks and automation. Just as a person might learn history and through exercise of specific tasks to the new work, intelligent algorithms through the same methods can do. For example, in November 2013, for which Google can automatically generate personalized email and social media response systems applied for a patent. Analysis of the working principle of this system is the first one past email and social media interaction. On the basis of this analysis, the system will have on the future of mail, Twitter or blog automatically write back, and with that person's writing style and tone. It's easy to imagine, the system will eventually be used to realize a lot of automation of the Exchange.
2nd impact on knowledge work, may also be a significant impact is that large data will change the way the company and its management. Data and forecast algorithm has the potential to change all the organizations and industry knowledge of nature and amount of work. Predictions based on data will increasingly replace human experience and judgment. As senior managers rely more on automated tools to generate data to make decisions, so the demand for analysis and management staff will continue to shrink. Although today there are a number of knowledge workers for a number of leadership collect information, analyze, but may end up with only one manager and a strong program to complete. Each organization may be compressed to streamline. Middle management would evaporate, and civilian personnel and technical analysts, and many jobs will disappear directly.
No comments:
Post a Comment