1. Can we have a generic AI Agent 'A' monitoring ad-hoc agents 'Bi' (each Bi handles syntax of one language) which will map Bi's syntax information into an agreeable Universal Syntax?
2. What problems can we face if we try to build 'A' that will work for whatever syntax model Bi use? (Here my notion is to capture feature information as much as possible. And also that many MTs are based on different models.)
3. How successful is MT so far?
Deductive reasoning and problem solving are still sheer unsurmountable tasks. Early AI researchers developed algorithms that imitated the step-by-step reasoning that humans use when they solve puzzles or make logical deductions. For difficult problems, most of these algorithms can require enormous computational resources – most experience a "combinatorial explosion": the amount of memory or computer time required becomes astronomical when the problem goes beyond a certain size. Human beings solve most of their problems using fast, intuitive judgments rather than the conscious, step-by-step deduction that early AI research was able to model.
Reasoning can be reduced to performing a search. Simple exhaustive searches however are rarely sufficient for most real world problems. The solution, for many problems, is to use "heuristics" or "rules of thumb" that eliminate choices that are unlikely to lead to the goal. A very different kind of search came to prominence in the 1990s, based on the mathematical theory of optimization. For many problems, it is possible to begin the search with some form of a guess and then refine the guess incrementally until no more refinements can be made.
Many problems in AI (in reasoning, planning, learning, perception and robotics) require the agent to operate with incomplete or uncertain information. Probabilistic algorithms can be used for filtering, prediction, smoothing and finding explanations for streams of data, helping perception systems to analyze processes that occur over time (e.g., hidden Markov models or Kalman filters).
Artificial intelligence, by claiming to be able to recreate the capabilities of the human mind, is both a challenge and an inspiration for philosophy. Are there limits to how intelligent machines can be? Is there an essential difference between human intelligence and artificial intelligence? Can a machine have a mind and consciousness?
Mary Shelley's Frankenstein considers a key issue in the ethics of artificial intelligence: if a machine can be created that has intelligence, could it also feel? If it can feel, does it have the same rights as a human? The idea also appears in modern science fiction, including the films I Robot, Blade Runner and A.I.: Artificial Intelligence, in which humanoid machines have the ability to feel human emotions. This issue, now known as "robot rights", is currently being considered by, for example, California's Institute for the Future, although many critics believe that the discussion is premature.
You might want to check Grammatical Framework, which provides a syntax based translation between about 20 languages, for which resource grammars are developed:
http://www.grammaticalframework.org/
Hi
you may want to check UNL (Universal Networking Language).
As for you question on how good an MT is?
An ideal MT system is required to be Fully Automatic, High Quality and General Purpose. But unfortunately, we can only provide any of the two features. The systems can be either
1. General Purpose Fully Automatic - Here we are not looking at the quality so we can have some dis-fluent translations. An EBMT/SMT can serve the purpose.
2. High Quality Fully Automatic - Here we need to implement a Transfer based system which requires some hundreds of man years.
3. General Purpose High Quality - Here MT's outputs is post edited with humans so although it loses automatic feature but can provide good translations across domains, so will be general purpose.
Deductive reasoning and problem solving are still sheer unsurmountable tasks. Early AI researchers developed algorithms that imitated the step-by-step reasoning that humans use when they solve puzzles or make logical deductions. For difficult problems, most of these algorithms can require enormous computational resources – most experience a "combinatorial explosion": the amount of memory or computer time required becomes astronomical when the problem goes beyond a certain size. Human beings solve most of their problems using fast, intuitive judgments rather than the conscious, step-by-step deduction that early AI research was able to model.
Reasoning can be reduced to performing a search. Simple exhaustive searches however are rarely sufficient for most real world problems. The solution, for many problems, is to use "heuristics" or "rules of thumb" that eliminate choices that are unlikely to lead to the goal. A very different kind of search came to prominence in the 1990s, based on the mathematical theory of optimization. For many problems, it is possible to begin the search with some form of a guess and then refine the guess incrementally until no more refinements can be made.
Many problems in AI (in reasoning, planning, learning, perception and robotics) require the agent to operate with incomplete or uncertain information. Probabilistic algorithms can be used for filtering, prediction, smoothing and finding explanations for streams of data, helping perception systems to analyze processes that occur over time (e.g., hidden Markov models or Kalman filters).
Artificial intelligence, by claiming to be able to recreate the capabilities of the human mind, is both a challenge and an inspiration for philosophy. Are there limits to how intelligent machines can be? Is there an essential difference between human intelligence and artificial intelligence? Can a machine have a mind and consciousness?
Mary Shelley's Frankenstein considers a key issue in the ethics of artificial intelligence: if a machine can be created that has intelligence, could it also feel? If it can feel, does it have the same rights as a human? The idea also appears in modern science fiction, including the films I Robot, Blade Runner and A.I.: Artificial Intelligence, in which humanoid machines have the ability to feel human emotions. This issue, now known as "robot rights", is currently being considered by, for example, California's Institute for the Future, although many critics believe that the discussion is premature.
It would be interesting for you to read: "The Language Instinct" by Steven Pinker. He argues that humans are born with an innate capacity for language. You will find nice clues for your job!
Thank you all for your responses. I am going through your ideas one by one and will reply to each of you with comments very soon.
The only solution I can think of is via the ETS representation, by translating a sentence in any language into the universal (ETS) representation.
http://csjarchive.cogsci.rpi.edu/proceedings/2008/pdfs/p1888.pdf
http://www.cs.unb.ca/~goldfarb/BOOK.pdf
Hi, Lev. Yours is a completely new thing for me. I was thinking of a representational account where i can give equal weigh to both syntactic and semantic values. I am going through your pointers. Thank you for your contribution.
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Artificial Intelligence