AWS makes a case for reputed scientist Biren cook “automated argument”.
Amazon AWS
The term “Reasoning” is a familiar metaphor in today’s artificial intelligence (AI) technology, which is often used to describe the so -called argument AI models such as OpenAI’s O1 or Dipsec AI’s R1 of AI’s R1.
Another kind of argument is quietly taking root in the most advanced applications, perhaps close to real argument.
Also: Will AI think like humans? We are not even close – and we are asking wrong questions
Recently, Amazon AWS is called “automatic argument” by Amazon AWS reputed scientist Bayran Cook, also known as “symbolic AI” or, more abstract, “formal verification”, for this.
It is an area of study as older as an artificial intelligence region, and, Cook said, it is faster merging with generous AI to create an exciting new hybrid, sometimes called “neuro-discourse AI”, which adds the best of automated logic and large language models.
Cook talked about automatic argument AWS Financial Services Seminar In New York in this May.
Whatever name you call it, the automatic argument refers to the algorithms that discover statements or claims about the world, which can be verified as truth using logic. The idea is that all knowledge is strictly supported which is capable of being logically capable.
Too: AWS says AI will promote the value of human creativity in financial services
As Cook said, “logic takes a model and lets us talk accurately about all possible data that it can produce.”
Cook gave a brief snipyt of the code as an example that shows how automatic arguments achieve that rigorous verification.
As Cook explained to his audience, a instruction loop in a piece of computer code can be predicted – with certainty – to stop walking at some point based on the conditions installed in its statements. So, the question, “Can this loop run forever?” Can be answered with logical analysis.
In the example of cook, two variables, x and y are integer; Y is positive, and more than XY. Y is repeatedly reduced by X, to reduce the value of X. Finally, the y will reduce the X from X to reduce the X from X. At that point, the conditions of the code loop have been violated, and the loop will end.
Simple facts – that will eventually be smaller than XY – the code loop can be logically estimated without running itself. This is probably the most important element of automated argument, a theory that returned to cook repeatedly: automated logic can answer fundamental questions about something with complete tests and logic rather than logic.
“This is a symbolic AI,” said Cook. “We find arguments, step by step, and we can conduct mechanically tests using the foundation of mathematical logic to ensure that each statement is true. And then automatic logic is the algorithm discovery for the arguments of that form.”
Cook said that such step-by-step solutions went back in the late 1950s in the morning of AI. In fact, in 1959, a top-of-the-line IBM machine, 704, ran a form of automatic argument to prove all theorems. Whitehead and Russell’s famous Principia Mathematica,
But there has been a lot of progress since then, Cook told the audience. Through the new algorithm, “equipment keeps on remarkably better”.
Too: What is Deepsek AI? is it safe? Here is everything you need to know
AWS has now been using automated arguments for a decade, Cook said, to achieve real -world functions such as guaranteeing delivery of AWS services according to Slas, or confirming network safety.
Translating a problem can be logically evaluated in those words, such as the code loop, it is all necessary.
For example, network safety often involves statements that are either right or absolutely wrong, which means they can be tested in the same way as the code loops automatically to determine whether the conditions are met or violated.
“When you see questions (AWS) customers ask, they use a lot of words, ‘for all,’ and ‘always’ and ‘never’,” like “is my data encrypted in comfort and transit?”
Cook said, “These are universal statements; they are highest than the big, if not immoral, if there are no infinite sets,” Cook said. Cook said, “It is not possible to test any policy to know such absolutism.” “The number of lifespan of the Sun will take all possible authority requests to fully test, it will take 92,686 points to write” – not practical, in other words.
Automatic argument, identity of AWS and using access management equipment IAM analystCook said, which is available for free for four years, “can solve the same problem in seconds.” “This is a proposal of the value of logic and mathematical argument contrary to the entire test.”
Cook argued that the power of automated logic means that it would be rapidly “a form of artificial super-intelligence”.
Too: Openai’s O1 is more than any major AI model. Why does it matter
“For some time, we have a form of artificial super-intelligence, if you do, it just speaks JSON,” Cook said. He said that automated arguments have been used to solve “open mathematics estimates,” the goods that “are in the headlines,” he said.
“We are solving in milliseconds or seconds or hours that humans can never solve, such as a hundred lifetime.”
Other uses in AWS include proving the purity of the open-source code developed by AWS and even “to prove the purity of the front door of AWS”, meaning that requests are to allow or reject requests for access to AWS that come two billion times from customers in a second.
Cook said that all these applications – AIM analysts, codes, AWS access authority, and many other equipment and services – attract an internal automatic argument to infrastructure in AWS called Zelkova, which can translate policies into mathematical formulas.
Cook said that a lot of speed for automated logic and Zelkova has come from the financial services industry.
Citing Investment Bank and hedge funds, Cook said, “We have really shared a good partnership with people like Goldman, Brijwatter.” Technology has helped those customers to “deploy rapidly, and, in fact, have a lot of money to save.”
Too: Google’s Deepmind unit says AI is beyond human knowledge
(John Kane, who heads the market development efforts in financial services for AWS, recently talked to ZDNET about the use of automated arguments for financial customers.)
The future of automated logic is melting it with generative AI, a synthesis refers to as a neuro-appearance.
At the most basic level, it is possible to translate from the terms of the natural language into formulas that can be strictly analyzed by Zelkova using logic.
In this way, General AI can be a non-technical person to frame its goal in informal, natural language terms, and then took automatic argument and applied it hardly. Two subjects can be added to non-logic to provide access to formal evidence, in other words.
Also: Apple’s controversial research paper actually tells us about LLMS
“You are an expert in financial services, in immigration law, with automated logic investigation, we give a person the ability to encoded, and the rules are derived here.”
Another cause of a hybrid is to deal with the limits of liberal AIs that have become clear, especially called hallucinations or confusion, the trend of large language models (LLM) to produce false claims, sometimes it happens wildly.
Cook said, “People became super excited about him (LLMS), and now he is beginning to realize that Oh, wait, some of these things are the limitations.” “You cannot force infinite data in these things, and they will always be better.”
Critics of scholars, especially current generic AI approaches have long discussed the idea of a hybrid neuro-delicious approach. Renowned General AI Scriptic Gary Marcus has suggested that General Ai really needs something like a formal argument for the land.
Also: Clobing every benchmark with AI model, this is the time for human evaluation
Even an enterprise-supported startup Sign The mission statement of which means what it sees as the LLM boundaries, will cross it.
Cook offered a practical example of the hybrid approach: check the veracity of chat bots.
“In a chat bot, you have questions and answers, and you want to know, is it true?” Cook said. Automatic argument allows you to evaluate statements according to formal arguments.
An example is currently an offering from AWS to preview, AWS Re: Invent, called Automated argument checkThe program can take the output of the natural language of a chatbot and can turn it into a formal argument that can then be verified.
Cook used a chat with a bank loan chatbot as an example. A person asks how much time it should take to get approval for his loan application. The chatbot reacts with a series of statements, such as “1 business day of approval.”
Automatic argument works to verify whether those answers from bots are true or not.
Cook explained to Cook, “In the background, what we are doing, we are taking natural language lessons, we are mapping it in mathematical logic, we are proving or dislike the purity of statements, and then we are providing witnesses, so you can pull on it, as a customer, the log of logic, that is true, but a way it can be audited.”
Cook said that the automatic argument will become even more important in the era of agent AI. He said, “Where things are going on, we are listening more and more about agents; on the promotional curve, it is like a new, new entry,” he said.
Also: Google’s new AI tool indicates in Opal apps, no coding is required.
“If you are going to allow the natural language to be converted into action, which makes one-way-darwaja decision on your behalf, then with your reputation, with your career, with your code, this purity is going to be exactly paramount. With agent AI, we are allowing the mortal to write and perform the system essentially.”
Agentic AI consists of many AI systems working parallel, and it should be solved in the way the automatic argument has solved the other distributed system in AWS, they argued.
For example, in the case of AWS’s S3 storage system, internal equipment, zelkova, was used to “prove the accuracy of the distributed system design,” he said.
“S3 (Amazon’s object storage) has hundreds of protocols under the hood,” Cook explained. “Assuming that all machines are speaking the protocol correctly, you will get strong stability – collectively, we will get the right result.”
He said that the same group can be exploited to verify the work of the polling approach, a type of knowledge of the crowd, the actions of the agents.
Also: Hacker Amazon’s Q AI Coding Assistant slips the malicious ‘wipe’ command – and the gods are worried
“This is the kind of thing that we can show with automatic arguments very quickly and very easily.”
Cook expressed optimism that automated logic and merger of General AI would continue to progress.
“I’m glad I am alive and I am happy to be a businessman in the field,” he said. “Because these branches are really really coming back together now.”
People wishing to find out this topic want to start further Cook’s introductory blog post on automatic argument since 2021,
Want more stories about AI? Sign up for innovationOur weekly newspapers.
