Posts by fwjmath

1) Questions and Answers : LODA Language : What is the research goal of LODA miner (Message 143) Posted 15 May 2022 by fwjmath Post: I think that, at least for me, one of the key is how to know if the way a mined LODA program computes a sequence is new in some sense. It should be different from know formulas, it should be different from the definition, and it should have a proof. All these three parts are difficult and requires human input for the moment. So the best we may have for the moment is that people interested in a given sequence may take time (or not!) to look at such programs. This is not likely to happen for sequences with "nice" tag though, as many of them may already have nice formulas. I think that the best bet may be those with "hard" and/or "more" tags, and preferably without "less" tag and/or with "nice" tag. But not many such sequences have enough term for LODA program mining. Some tests may be needed. fwjmath.
2) Questions and Answers : LODA Language : What is the research goal of LODA miner (Message 72) Posted 14 May 2022 by fwjmath Post: As a researcher who use OEIS whenever meeting an integer sequence, I am quite interested in this project, but there are also some doubt. LODA seems to be a versatile language for computing integer sequences. It seems to encode primitive recursive functions, which is largely sufficient for all computations, but mind that Knuth's up arrow notation is also primitive recursive, so even though LODA programs will eventually halt, it may halt after a long time. Some of the sequences on OEIS are also out of reach for LODA, for instance A079365 of the Chaitin constant, but that is less important. But that's not my point. As a researcher, I am much more interested in the interpretation of these programs. In mathematics, not only do we want a program that computes a sequence, we also want to understand it, ususally by giving a proof. That is the only way that we can be sure that the program really computes the sequence. However, the LODA language seems to be quite difficult to interprete into meaningful mathematical content. Do you have concrete ideas on how to exploit the mined programs? fwjmath.